Multi-pair multi-base interpretation based on artificial intelligence

Abstract

The technology disclosed herein relates to base interpretation based on artificial intelligence. The disclosed technology relates to accessing the progress of a set of per-cycle analyte channels generated for a sequencing cycle of a sequencing run; a window of a set of analyte channels per cycle in progress is processed by a neural network based base interpreter (NNBC) for a window of a sequencing cycle of a sequencing run such that the NNBC processes a subject window of the set of analyte channels per cycle in progress for a subject window of a sequencing cycle of the sequencing run, and using the NNBC to generate a temporary base interpretation prediction for three or more sequencing cycles in a target window of a sequencing cycle from a plurality of windows at which a particular sequencing cycle appears at different positions, to generate a temporary base interpretation prediction for the particular sequencing cycle; and determining a base interpretation for the particular sequencing cycle based on the plurality of base interpretation predictions.

Description

[0001] This application is a divisional application of the application filed on February 19, 2021, with application number 202180015480.7 and invention title "Multi-to-Multi-Base Interpretation Based on Artificial Intelligence". Technical Field

[0002] The technologies disclosed in this invention relate to artificial intelligence-type computers and digital data processing systems, as well as corresponding data processing methods and products for simulating intelligence (i.e., knowledge-based systems, inference systems, and knowledge acquisition systems); and include systems for uncertainty inference (e.g., fuzzy logic systems), adaptive systems, machine learning systems, and artificial neural networks. Specifically, the disclosed technologies relate to using deep neural networks, such as deep convolutional neural networks, for data analysis. Priority application

[0003] This application claims priority and benefit to U.S. Provisional Patent Application No. 62 / 979,414 (Attorney's File No. ILLM1016-1 / IP-1858-PRV), filed February 20, 2020, entitled "ARTIFICIAL INTELLIGENCE-BASEDMANY-TO-MANY BASE CALLING," and U.S. Patent Application No. 17 / 180,542 (Attorney's File No. ILLM 1016-2 / IP-1858-US), filed February 19, 2021, also entitled "ARTIFICIAL INTELLIGENCE-BASEDMANY-TO-MANY BASE CALLING." These priority applications are incorporated herein by reference as if fully illustrated herein for all purposes. Literature merged

[0004] The following references are incorporated herein by reference as if they were shown in their entirety in this article:

[0005] U.S. Provisional Patent Application No. 62 / 979,384, entitled “ARTIFICIAL INTELLIGENCE-BASED BASE CALLINGOF INDEX SEQUENCES”, filed on February 20, 2020 (Attorney’s File No. ILLM 1015-1 / IP-1857-PRV).

[0006] U.S. Provisional Patent Application No. 62 / 979,385, entitled “KNOWLEDGE DISTILLATION-BASED COMPRESSION OFARTIFICIAL INTELLIGENCE-BASED BASE CALLER”, filed on February 20, 2020 (Attorney’s File No. ILLM 1017-1 / IP-1859-PRV).

[0007] U.S. Provisional Patent Application No. 63 / 072,032, entitled “DETECTING AND FILTERING CLUSTERS BASED ON ARTIFICIAL INTELLIGENCE-PREDICTED BASE CALLS”, filed on August 28, 2020 (Attorney’s File No. ILLM 1018-1 / IP-1860-PRV).

[0008] U.S. Provisional Patent Application No. 62 / 979,412, entitled “MULTI-CYCLE CLUSTER BASED REAL TIMEANALYSIS SYSTEM”, filed on February 20, 2020 (Attorney’s File No. ILLM 1020-1 / IP-1866-PRV).

[0009] U.S. Provisional Patent Application No. 62 / 979,411, entitled “DATA COMPRESSION FOR ARTIFICIALINTELLIGENCE-BASED BASE CALLING”, filed on February 20, 2020 (Attorney’s File No. ILLM 1029-1 / IP-1964-PRV).

[0010] U.S. Provisional Patent Application No. 62 / 979,399, entitled “SQUEEZING LAYER FOR ARTIFICIALINTELLIGENCE-BASED BASE CALLING”, filed on February 20, 2020 (Attorney’s File No. ILLM 1030-1 / IP-1982-PRV).

[0011] U.S. non-provisional patent application No. 16 / 825,987, entitled “TRAINING DATA GENERATION FOR ARTIFICIALINTELLIGENCE-BASED SEQUENCING”, filed on March 20, 2020 (Attorney’s File No. ILLM 1008-16 / IP-1693-US).

[0012] U.S. non-provisional patent application number 16 / 825,991, entitled “ARTIFICIAL INTELLIGENCE-BASED GENERATION OF SEQUENCING METADATA”, filed on March 20, 2020 (Attorney’s file number ILLM 1008-17 / IP-1741-US).

[0013] U.S. non-provisional patent application number 16 / 826,126, entitled “ARTIFICIAL INTELLIGENCE-BASED BASE CALLING”, filed on March 20, 2020 (Attorney’s file number ILLM 1008-18 / IP-1744-US).

[0014] U.S. non-provisional patent application No. 16 / 826,134 (Attorney's File No. ILLM 1008-19 / IP-1747-US), filed on March 20, 2020, entitled "ARTIFICIAL INTELLIGENCE-BASED QUALITYSCORING"; and

[0015] The U.S. non-provisional patent application No. 16 / 826,168 entitled “ARTIFICIAL INTELLIGENCE-BASED SEQUENCING” was filed on March 21, 2020 (Attorney’s File No. ILLM 1008-20 / IP-1752-PRV-US). Background Technology

[0016] The topics discussed in this section should not be considered prior art simply because they are mentioned here. Similarly, problems mentioned in this section or related to the topics provided as background art should not be assumed to have been previously recognized in the prior art. The topics in this section merely represent different methods, which themselves may correspond to specific implementations of the technology protected by the claims.

[0017] Deep neural networks are artificial neural networks that use multiple nonlinear and complex transformation layers to continuously model high-level features. Deep neural networks provide feedback via backpropagation, which carries the difference between the observed output and the predicted output to adjust the parameters. Deep neural networks have evolved with the availability of large training datasets, the ability to perform parallel and distributed computing, and sophisticated training algorithms. Deep neural networks have facilitated significant advancements in many fields such as computer vision, speech recognition, and natural language processing.

[0018] Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) are components of deep neural networks. CNNs have been particularly successful in image recognition with architectures that include convolutional layers, nonlinear layers, and pooling layers. RNNs are designed to utilize sequential information from the input data and feature recurrent connections between building blocks such as perceptrons, long short-term memory units, and gated recurrent units. Furthermore, many other emerging deep neural networks for finite contexts have been proposed, such as deep spatiotemporal neural networks, multidimensional recurrent neural networks, and convolutional autoencoders.

[0019] The goal of training a deep neural network is to optimize the weight parameters in each layer, gradually combining simpler features into more complex ones, allowing the learning of the most suitable hierarchical representation from the data. A single loop of the optimization process proceeds as follows: First, given the training dataset, the forward pass sequentially computes the output in each layer and propagates the function signal forward through the network. In the final output layer, the target loss function measures the error between the inferred output and the given label. To minimize the training error, the backward pass uses the chain rule to backpropagate the error signal and computes the gradient with respect to all weights throughout the neural network. Finally, the weight parameters are updated using an optimization algorithm based on stochastic gradient descent. While batch gradient descent performs parameter updates for each complete dataset, stochastic gradient descent provides a stochastic approximation by performing updates for each small set of data examples. Several optimization algorithms are derived from stochastic gradient descent. For example, the Adagrad and Adam training algorithms perform stochastic gradient descent while adaptively modifying the learning rate based on the update frequency of the gradient for each parameter and momentum, respectively.

[0020] Another core element in deep neural network training is regularization, which refers to strategies designed to avoid overfitting and thus achieve good generalization performance. For example, weight decay adds a penalty factor to the target loss function, causing the weight parameters to converge to a smaller absolute value. Dropout randomly removes hidden units from the neural network during training and can be considered as an ensemble of possible subnetworks. To enhance the power of dropout, novel activation functions, maximum output, and dropout variants for recurrent neural networks (known as rnnDrop) have been proposed. Furthermore, batch normalization provides a novel regularization method by normalizing the scalar features of each activation within a microbatch and learning the mean and variance of each as parameters.

[0021] Given the multidimensional and high-dimensional nature of sequence data, deep neural networks hold immense promise for bioinformatics research due to their wide applicability and enhanced predictive power. Convolutional neural networks (CNNs) have been used to solve sequence-based problems in genomics, such as motif discovery, pathogenic variant identification, and gene expression inference. CNNs employ a weight-sharing strategy, which is particularly useful for studying deoxyribonucleic acid (DNA) because it captures sequence motifs—short, recurring local patterns in DNA presumed to have significant biological functions. A hallmark of CNNs is the use of convolutional filters.

[0022] Unlike traditional classification methods based on meticulously designed and handcrafted features, convolutional filters perform adaptive learning of features, analogous to the process of mapping raw input data to an informational representation of knowledge. In this sense, convolutional filters act as a series of motif scanners, as a set of such filters is able to identify relevant patterns in the input and update itself during the training process. Recurrent neural networks can capture long-range dependencies in sequence data of varying lengths, such as protein or DNA sequences.

[0023] Therefore, there is an opportunity to use a deep learning-based framework for template generation and base calling.

[0024] In the era of high-throughput technologies, accumulating the most interpretable data at the lowest cost per job remains a significant challenge. Cluster-based nucleic acid sequencing methods, such as those using bridging amplification to form clusters, have made important contributions to increasing nucleic acid sequencing throughput. These cluster-based methods rely on sequencing dense populations of nucleic acids immobilized on solid-state supports and typically involve using image analysis software to deconvolve the optical signals generated during the simultaneous sequencing of multiple clusters located at different locations on the solid-state support.

[0025] However, such solid-phase nucleic acid cluster-based sequencing technologies still face considerable obstacles that limit achievable throughput. For example, in cluster-based sequencing methods, there can be barriers to identifying the nucleic acid sequences of two or more clusters that are too physically close to each other to be spatially distinguishable or that actually physically overlap on a solid carrier. For instance, current image analysis software may require valuable time and computational resources to determine which of two overlapping clusters emitted the light signal. Therefore, trade-offs regarding the quantity and / or quality of available nucleic acid sequence information are unavoidable for various detection platforms.

[0026] High-density nucleic acid cluster-based genomics methods have also been extended to other areas of genome analysis. For example, nucleic acid cluster-based genomics can be used for sequencing applications, diagnostics and screening, gene expression analysis, epigenetic analysis, polymorphic genetic analysis, and so on. Each of these nucleic acid cluster-based genomic technologies is also limited when data generated by closely spaced or spatially overlapping nucleic acid clusters cannot be resolved.

[0027] Clearly, there remains a need to increase the quality and quantity of rapidly and cost-effectively available nucleic acid sequencing data for a wide range of applications, including genomics (e.g., for genomic characterization of any and all animal, plant, microbial, or other biological species or populations), pharmacogenetics, transcriptomics, diagnostics, prognosis, biomedical risk assessment, clinical and research genetics, personalized medicine, drug efficacy and drug interaction assessment, veterinary medicine, agriculture, evolution and biodiversity research, aquaculture, forestry, oceanography, ecological and environmental management, and other purposes.

[0028] The disclosed technology provides neural network-based methods and systems that address these and similar needs, including increasing throughput in high-throughput nucleic acid sequencing technologies, and offers other related advantages. Attached Figure Description

[0029] In the accompanying drawings, similar reference numerals generally refer to similar parts in all different views. Furthermore, the drawings are not necessarily drawn to scale, but rather emphasize the principles of the disclosed technology. In the following description, various specific embodiments of the disclosed technology are described with reference to the following drawings, wherein:

[0030] Figure 1A , Figure 1B and Figure 1C The disclosed multi-base pair interpretation is shown.

[0031] Figure 1D and Figure 1E Different examples of the disclosed multiple-base pair interpretations are shown.

[0032] Figure 2 , Figure 3 and Figure 4 Different specific implementations of the base determination generator are shown.

[0033] Figure 5 A specific implementation of the disclosed multi-cyclic gradient backpropagation is shown.

[0034] Figure 6 It is a flowchart of a specific implementation of the disclosed technology.

[0035] Figure 7The technical effects and advantages of the disclosed technology are shown.

[0036] Figure 8A and Figure 8B A specific implementation of a sequencing system is described. This sequencing system includes a configurable processor.

[0037] Figure 9 This is a simplified block diagram of a system for analyzing sensor data (such as base reading sensor output) from a sequencing system.

[0038] Figure 10 This is a simplified diagram illustrating an aspect of the base-reading operation, which includes the functionality of a runtime program executed by the host processor.

[0039] Figure 11 It is a configurable processor (such as, Figure 9 A simplified diagram of the configuration of the configurable processor.

[0040] Figure 12 It is a computer system that can be used by the disclosed sequencing system to implement the base interpretation technology disclosed herein. Detailed Implementation

[0041] The following discussion is presented to enable any person skilled in the art to implement and use the disclosed technology, and is provided in the context of a particular application and its requirements. Various modifications to the specific embodiments disclosed will be apparent to those skilled in the art, and the general principles defined herein can be applied to other specific embodiments and applications without departing from the spirit and scope of the disclosed technology. Therefore, the disclosed technology is not intended to be limited to the specific embodiments shown, but is to be accorded the broadest scope consistent with the principles and features disclosed herein.

[0042] Sequencing images

[0043] Base interpretation is the process of determining the nucleotide composition of a sequence. Base interpretation involves analyzing image data, or sequencing images, which are generated during sequencing runs (or sequencing reactions) performed by sequencing instruments such as Illumina's iSeq, HiSeqX, HiSeq 3000, HiSeq 4000, HiSeq 2500, NovaSeq 6000, NextSeq 550, NextSeq 1000, NextSeq 2000, NextSeqDx, MiSeq, and MiSeqDx.

[0044] The following discussion outlines a method for generating sequencing images based on a specific implementation and describes its contents.

[0045] Base interpretation decodes intensity data encoded in the sequencing image into the nucleotide sequence. In one specific implementation, the Illumina sequencing platform employs cycle reversible termination (CRT) chemistry for base interpretation. This process relies on growing a nascent strand complementary to a template strand containing fluorescently labeled nucleotides, while tracking the emission signal of each newly added nucleotide. The fluorescently labeled nucleotides have a 3' removable block of fluorophore signal anchored to the nucleotide type.

[0046] Sequencing is performed in repeated cycles, each cycle consisting of three steps: (a) elongating the nascent strand by adding fluorescently labeled nucleotides; (b) exciting the fluorophore using one or more lasers of the sequencing instrument's optical system and imaging it through different filters of the optical system to produce a sequencing image; and (c) lysing the fluorophore and removing the 3' block to prepare for the next sequencing cycle. The incorporation and imaging cycles are repeated until a specified number of sequencing cycles are reached, thereby defining the read length. Using this method, each cycle queries a new location along the template strand.

[0047] The immense power of Illumina sequencers stems from their ability to simultaneously execute and sense millions, or even billions, of clusters (also known as “analytes”) undergoing a CRT reaction. A cluster comprises approximately a thousand identical copies of the template strand, but differs in size and shape. Before sequencing, these clusters from the template strand are grown by bridging or exclusion amplification of the input library. The purpose of amplification and cluster growth is to increase the intensity of the emission signal, as imaging devices cannot reliably sense the fluorophore signal of a single strand. However, the strands within a cluster are physically close together, so the imaging device perceives the cluster of strands as a single point.

[0048] Sequencing occurs in a flow cell (or biosensor), a small glass slide that holds the input chain. The flow cell is connected to an optical system that includes microscopic imaging, an excitation laser, and a fluorescence filter. The flow cell comprises multiple chambers called channels. These channels are physically separated from each other and can contain different labeled sequencing libraries that can be distinguished in the absence of sample cross-contamination. In some implementations, the flow cell includes a patterned surface. A “patterned surface” refers to the arrangement of different regions within or on the exposed layer of a solid support.

[0049] Imaging devices in sequencing instruments (e.g., solid-state imaging devices such as charge-coupled devices (CCDs) or complementary metal-oxide-semiconductor (CMOS) sensors) take snapshots at multiple locations along the channel in a series of non-overlapping regions (called blocks). For example, each channel may contain sixty-four or ninety-six blocks. A block can hold hundreds of thousands to millions of clusters.

[0050] The output of a sequencing run is a sequencing image. The sequencing image depicts the intensity emission of clusters and their surrounding background using a grid (or array) of pixelated units (e.g., pixels, superpixels, subpixels). The intensity emission is stored as the intensity values ​​of the pixelated units. The sequencing image has the grid dimensions w, x, and h of the pixelated units, where w (width) and h (height) are any numbers in the range of 1 to 100,000 (e.g., 115×115, 200×200, 1800×2000, 2200×25000, 2800×3600, 4000×400). In some implementations, w and h are the same. In other implementations, w and h are different. The sequencing image depicts the intensity emission generated due to nucleotide incorporation into the nucleotide sequence during the sequencing run. The intensity emission originates from the associated clusters and their surrounding background.

[0051] Neural Network-Based Base Determination

[0052] The following discussion focuses on the neural network-based base reader 102 described herein. First, according to a specific embodiment, the input to the neural network-based base reader 102 is described. Then, examples of the structure and form of the neural network-based base reader 102 are provided. Finally, according to a specific embodiment, the output of the neural network-based base reader 102 is described.

[0053] Data flow logic provides sequencing images to a neural network-based base reader 102 for base interpretation. The neural network-based base reader 102 accesses the sequencing images on a block-by-block (or region-by-block) basis. Each block is a sub-grid (or sub-array) of pixelated units within a grid of pixelated units that forms the sequencing image. A block has a size q × r for its sub-grid of pixelated units, where q (width) and r (height) are any numbers in the range of 1 to 10000 (e.g., 3×3, 5×5, 7×7, 10×10, 15×15, 25×25, 64×64, 78×78, 115×115). In some embodiments, q and r are the same. In other embodiments, q and r are different. In some embodiments, the blocks extracted from the sequencing image have the same size. In other embodiments, the blocks have different sizes. In some embodiments, blocks may have overlapping pixelated units (e.g., at edges).

[0054] Sequencing produces m sequencing images per sequencing cycle for corresponding m image channels. That is, each sequencing image has one or more image (or intensity) channels (similar to the red, green, and blue (RGB) channels in a color image). In one embodiment, each image channel corresponds to one of a plurality of filter wavelength bands. In another embodiment, each image channel corresponds to one of a plurality of imaging events in a sequencing cycle. In yet another embodiment, each image channel corresponds to a combination of illumination using a specific laser and imaging through a specific optical filter. Image patches are tiled (or accessed) from each of the m image channels for a specific sequencing cycle. In different embodiments such as four-channel chemistry, two-channel chemistry, and single-channel chemistry, m is 4 or 2. In other embodiments, m is 1, 3, or greater than 4.

[0055] For example, consider a sequencing run implemented using two different image channels (blue and green channels). Then, in each sequencing cycle, the sequencing run produces blue and green images. Thus, for a series of k sequencing cycles of the sequencing run, k pairs of blue and green image sequences are produced as output and stored as sequencing images. Therefore, a neural network-based base reader 102 generates sequences of k pairs of blue and green image blocks for block-level processing.

[0056] The input image data (or a single instance of forward pass or a single forward traversal) for a single iteration of the neural network-based base reader 102 used for base interpretation includes data for a sliding window used for multiple sequencing cycles. The sliding window may include, for example, the current sequencing cycle, one or more previous sequencing cycles, and one or more subsequent sequencing cycles.

[0057] In one specific implementation, the input image data includes data for three sequencing cycles, such that the data to be interpreted for the current (time t) sequencing cycle is accompanied by: (i) data for the left wing / context / previous / next / before (time t-1) sequencing cycle and (ii) data for the right wing / context / next / follower / after (time t+1) sequencing cycle.

[0058] In another specific implementation, the input image data includes data for five sequencing cycles, such that the data for the current (time t) sequencing cycle to be base-read is accompanied by: (i) data for the first left-wing / context / previous / next / before (time t-1) sequencing cycle; (ii) data for the second left-wing / context / previous / next / before (time t-2) sequencing cycle; (iii) data for the first right-wing / context / next / follow / after (time t+1) sequencing cycle; and (iv) data for the second right-wing / context / next / follow / after (time t+2) sequencing cycle.

[0059] In another embodiment, the input image data includes data for seven sequencing cycles, such that the data for the current (time t) sequencing cycle for base interpretation is accompanied by: (i) data for the first left-wing / context / previous / previous / before (time t-1) sequencing cycle; (ii) data for the second left-wing / context / previous / previous / before (time t-2) sequencing cycle; (iii) data for the third left-wing / context / previous / previous / before (time t-3) sequencing cycle; (iv) data for the first right-wing / context / next / follow / after (time t+1) sequencing cycle; (v) data for the second right-wing / context / next / follow / after (time t+2) sequencing cycle; and (vi) data for the third right-wing / context / next / follow / after (time t+3) sequencing cycle. In other embodiments, the input image data includes data for a single sequencing cycle. In other specific implementations, the input image data includes data for 10, 15, 20, 30, 58, 75, 92, 130, 168, 175, 209, 225, 230, 275, 318, 325, 330, 525, or 625 sequencing cycles.

[0060] According to one specific implementation, a neural network-based base reader 102 processes image patches through its convolutional layers and generates alternative representations. The alternative representations are then used by the output layer (e.g., a softmax layer) to generate base reads only for the current (time t) sequencing cycle or each sequencing cycle within a sequencing cycle, i.e., the current (time t) sequencing cycle, the first and second previous (time t-1, time t-2) sequencing cycles, and the first and second subsequent (time t+1, time t+2) sequencing cycles. The resulting base reads form sequencing reads.

[0061] In one embodiment, the neural network-based base reader 102 outputs a base read for a single target cluster used in a specific sequencing cycle. In another embodiment, the neural network-based base reader 102 outputs a base read for each of a plurality of target clusters used in a specific sequencing cycle. In yet another embodiment, the neural network-based base reader 102 outputs a base read for each of a plurality of target clusters used in each of a plurality of sequencing cycles, thereby generating a base read sequence for each target cluster.

[0062] In one embodiment, the neural network-based base discriminator 102 is a multilayer perceptron (MLP). In another embodiment, the neural network-based base discriminator 102 is a feedforward neural network. In yet another embodiment, the neural network-based base discriminator 102 is a fully connected neural network. In a further embodiment, the neural network-based base discriminator 102 is a fully convolutional neural network. In an even further embodiment, the neural network-based base discriminator 102 is a semantic segmentation neural network. In yet another further embodiment, the neural network-based base discriminator 102 is a generative adversarial network (GAN).

[0063] In one embodiment, the neural network-based base reader 102 is a convolutional neural network (CNN) with multiple convolutional layers. In another embodiment, the neural network-based base reader 102 is a recurrent neural network (RNN), such as a long short-term memory network (LSTM), a bidirectional LSTM (Bi-LSTM), or a gated recurrent unit (GRU). In yet another embodiment, the neural network-based base reader 102 includes both CNN and RNN.

[0064] In other specific implementations, the neural network-based base discriminator 102 can use 1D convolution, 2D convolution, 3D convolution, 4D convolution, 5D convolution, dilated or dilated convolution, transposed convolution, depthwise separable convolution, pointwise convolution, 1×1 convolution, grouped convolution, flattened convolution, spatial and cross-channel convolution, shuffled grouped convolution, spatially separable convolution, and deconvolution. The neural network-based base discriminator 102 can use one or more loss functions, such as logistic regression / logarithmic loss, multi-class cross-entropy / softmax loss, binary cross-entropy loss, mean squared error loss, L1 loss, L2 loss, smoothed L1 loss, and Huber loss. The neural network-based base discriminator 102 can use any parallelism, efficiency, and compression scheme, such as TFRecords, compressed encoding (e.g., PNG), sharding, parallel discrimination with map transformation, batch processing, prefetching, model parallelism, data parallelism, and synchronous / asynchronous stochastic gradient descent (SGD). The neural network-based base reader 102 may include upsampling layers, downsampling layers, recursive connections, gates and gate memory cells (such as LSTM or GRU), residual blocks, residual connections, high-speed connections, skip connections, peephole connections, activation functions (e.g., nonlinear transformation functions such as modified linear unit (ReLU), leaky ReLU, exponential lined unit (ELU), sigmoid, and hyperbolic tangent (tanh)), batch normalization layers, regularization layers, dropout layers, pooling layers (e.g., max pooling or average pooling), global average pooling layers, and attention mechanisms.

[0065] The neural network-based base reader 102 is trained using a gradient update technique based on backpropagation. Exemplary gradient descent techniques that can be used to train the neural network-based base reader 102 include stochastic gradient descent, batch gradient descent, and microbatch gradient descent. Some examples of gradient descent optimization algorithms that can be used to train the neural network-based base reader 102 are Momentum, Nesterov accelerated gradient, Adagrad, Adadelta, RMSprop, Adam, AdaMax, Nadam, and AMSGrad.

[0066] In one specific implementation, the neural network-based base reader 102 uses a specialized architecture to separate the processing of data for different sequencing cycles. The motivation for using this specialized architecture is first described. As mentioned above, the neural network-based base reader 102 processes image patches for the current sequencing cycle, one or more previous sequencing cycles, and one or more subsequent sequencing cycles. Data from additional sequencing cycles provides sequence-specific context. The neural network-based base reader 102 learns the sequence-specific context during training and performs base interpretation on that context. Furthermore, data from previous and subsequent sequencing cycles provide second-order contributions of pre-phase and phasing signals to the current sequencing cycle.

[0067] However, images captured at different sequencing cycles and in different image channels are misaligned relative to each other and have residual registration errors. To address this misalignment, a specialized architecture includes spatial convolutional layers that do not mix information between sequencing cycles and only mix information within a sequencing cycle.

[0068] Spatial convolutional layers (or spatial logic) use so-called “isolated convolutions,” which achieve isolation by processing data from each of multiple sequencing cycles independently via “dedicated, non-shared” convolutional sequences. This isolated convolution convolves only the data and resulting feature maps from a given sequencing cycle (i.e., within the cycle), without convolving the data and resulting feature maps from any other sequencing cycles.

[0069] For example, consider input image data including: (i) the current image patch for the current (time t) sequencing cycle to be base-determined; (ii) the previous image patch for the previous (time t-1) sequencing cycle; and (iii) the next image patch for the next (time t+1) sequencing cycle. The specialized architecture then initiates three separate convolutional pipelines: the current convolutional pipeline, the previous convolutional pipeline, and the next convolutional pipeline. The current data processing pipeline receives the current image patch for the current (time t) sequencing cycle as input and processes it independently through multiple spatial convolutional layers to produce a so-called "current spatial convolutional representation" as the output of the final spatial convolutional layer. The previous convolutional pipeline receives the previous image patch for the previous (time t-1) sequencing cycle as input and processes it independently through multiple spatial convolutional layers to produce a so-called "previous spatial convolutional representation" as the output of the final spatial convolutional layer. The next convolutional pipeline receives the next image patch as input for the next (time t+1) sequencing cycle and processes the next image patch independently through multiple spatial convolutional layers to produce the so-called "next spatial convolutional representation" as the output of the final spatial convolutional layer.

[0070] In some implementations, the current convolutional pipeline, the previous convolutional pipeline, and the next convolutional pipeline are executed in parallel. In some implementations, spatial convolutional layers are part of a spatial convolutional network (or subnetwork) within a specialized architecture.

[0071] The neural network-based base reader 102 further includes a temporal convolutional layer (or temporal logic) that mixes information between sequencing cycles (i.e., between cycles). The temporal convolutional layer receives its input from the spatial convolutional network and operates on the spatial convolutional representation produced by the final spatial convolutional layer of the corresponding data processing pipeline.

[0072] The inter-cycle operability freedom of temporal convolutional layers stems from the fact that misalignment attributes are cleared from the spatial convolutional representation by stacking or cascading isolated convolutions performed by a sequence of spatial convolutional layers, which exist in the image data fed into the spatial convolutional network as input.

[0073] Temporal convolutional layers use so-called "combined convolutions," which convolve input channels in subsequent inputs group by group on a sliding window basis. In one specific implementation, these subsequent inputs are the subsequent outputs produced by previous spatial or temporal convolutional layers.

[0074] In some implementations, temporal convolutional layers are part of a temporal convolutional network (or subnetwork) within a specialized architecture. The temporal convolutional network receives its input from a spatial convolutional network. In one implementation, the first temporal convolutional layer of the temporal convolutional network combines spatial convolutional representations between sequencing cycles group by group. In another implementation, subsequent temporal convolutional layers of the temporal convolutional network combine subsequent outputs from previous temporal convolutional layers. The output of the final temporal convolutional layer is fed into an output layer that produces the final output. The output is used for base derivation of one or more clusters at one or more sequencing cycles.

[0075] The data flow logic provides per-cycle cluster data to the neural network-based base reader 102. The per-cycle cluster data is used for a first subset of the sequencing cycles of multiple clusters and sequencing runs. For example, consider a sequencing run with 150 sequencing cycles. The first subset of sequencing cycles can then include any subset of the 150 sequencing cycles, such as the first 5, 10, 15, 25, 35, 40, 50, or 100 sequencing cycles of the 150-cycle sequencing run. Furthermore, each sequencing cycle produces a sequencing image depicting the intensity emission of clusters among the multiple clusters. Thus, the per-cycle cluster data for the first subset of sequencing cycles for multiple clusters and sequencing runs includes sequencing images only for the first 5, 10, 15, 25, 35, 40, 50, or 100 sequencing cycles of the 150-cycle sequencing run, and does not include sequencing images for the remaining sequencing cycles of the 150-cycle sequencing run.

[0076] The neural network-based base reader 102 performs base interpretation on each of the multiple clusters in each sequencing cycle within the first subset of the sequencing cycles. To do this, the neural network-based base reader 102 processes the per-cycle cluster data and generates an intermediate representation of the per-cycle cluster data. Then, the neural network-based base reader 102 processes the intermediate representation through an output layer and generates a per-cluster, per-cycle probability quadruples for each cluster and each sequencing cycle. Examples of output layers include softmax functions, log-softmax functions, ensemble output averaging functions, multilayer perceptron uncertainty functions, Bayesian Gaussian distribution functions, and cluster strength functions. The per-cluster, per-cycle probability quadruples are stored as probability quadruples and are referred to herein as “base-wise probabilities” because there are four nucleotide bases: A, C, T, and G.

[0077] The softmax function is a preferred function for multi-level classification. The softmax function calculates the probability of each target class across all possible target classes. The output of the softmax function ranges from zero to one, and the sum of all probabilities equals one. The softmax function calculates the exponent of a given input value and the sum of the exponents of all input values. The ratio of the exponent of the input values ​​to the sum of the exponents is the output of the softmax function, referred to in this paper as "exponential normalization".

[0078] Formally, training a so-called softmax classifier is regressing to class probabilities, not to the true classifier, because it doesn't return the class itself, but rather the confidence predictions for each class probability. The softmax function takes a class value and converts them into probabilities that sum to 1. The softmax function converts any real-valued... Dimensional vectors are compressed to real values ​​in the range of 0 to 1. Dimensional vector. Therefore, the softmax function is used to ensure that the output is a valid, exponentially normalized probability mass function (non-negative and summing to 1).

[0079] Intuitively, the softmax function is a "soft" version of the maximization function. The term "soft" comes from the fact that the softmax function is continuous and differentiable. Instead of choosing a maximum element, it breaks the vector into parts where the largest input element receives a larger proportion of the value, and the others receive smaller proportions. This property of the output probability distribution makes the softmax function suitable for probabilistic interpretation in classification tasks.

[0080] Let us This is considered as the input vector to the softmax layer. The number of units in the softmax layer is the number of nodes in the softmax layer, and therefore... The length of the vector is the number of units in the softmax layer (if there are ten output units, then there are ten). element).

[0081] for Dimensional vector The softmax function uses exponential normalization (exp) to produce another normalized value that is in the range [0, 1] and sums to one. Dimensional vector .

[0082]

[0083]

[0084] For example, applying the softmax function to three categories as It should be noted that the three outputs always sum to 1. Therefore, they define a discrete probability mass function.

[0085] The four-fold probability per cluster and per cycle represents the probability of identifying A, C, T, and G as bases incorporated into a specific cluster at a specific sequencing cycle. When the output layer of the neural network-based base reader 102 uses the softmax function, the probability in the four-fold probability per cluster and per cycle is an exponentially normalized classification score, and the sum of the classification scores is one.

[0086] In one implementation, the method includes processing the convolutional representation through an output layer to generate the probability of incorporating A, C, T, and G bases in the target analyte at the current sequencing cycle, and classifying the base as A, C, T, or G based on this probability. In one implementation, the probability is an exponentially normalized fraction generated by a softmax layer.

[0087] In one embodiment, the method includes deriving output pairs of target analytes from the output, the output pairs identifying class markers for bases of A, C, T, or G incorporated into the target analyte at the current sequencing cycle, and performing base interpretation of the target analyte based on these class markers. In one embodiment, class marker 1, 0 identifies an A base; class marker 0, 1 identifies a C base; class marker 1, 1 identifies a T base; and class marker 0, 0 identifies a G base. In another embodiment, class marker 1, 1 identifies an A base; class marker 0, 1 identifies a C base; class marker 0.5, 0.5 identifies a T base; and class marker 0, 0 identifies a G base. In yet another embodiment, class marker 1, 0 identifies an A base; class marker 0, 1 identifies a C base; class marker 0.5, 0.5 identifies a T base; and class marker 0, 0 identifies a G base. In a further embodiment, class markers 1 and 2 identify A bases; class markers 0 and 1 identify C bases; class markers 1 and 1 identify T bases; and class markers 0 and 0 identify G bases. In one embodiment, the method includes deriving class markers of the target analyte from the output, which identify A, C, T, or G bases incorporated into the target analyte at the current sequencing cycle, and performing base interpretation of the target analyte based on these class markers. In one embodiment, class marker 0.33 identifies A bases; class marker 0.66 identifies C bases; class marker 1 identifies T bases; and class marker 0 identifies G bases. In another embodiment, class marker 0.50 identifies A bases; class marker 0.75 identifies C bases; class marker 1 identifies T bases; and class marker 0.25 identifies G bases. In one embodiment, the method includes: deriving a single output value from the output; comparing the single output value with class value ranges corresponding to bases A, C, T, and G; assigning the single output value to a specific class value range; and performing base interpretation on the target analyte based on the assignment. In one embodiment, a sigmoid function is used to derive the single output value, and the single output value is in the range of 0 to 1. In another embodiment, a class value range of 0 to 0.25 represents A bases, a class value range of 0.25 to 0.50 represents C bases, a class value range of 0.50 to 0.75 represents T bases, and a class value range of 0.75 to 1 represents G bases.

[0088] Further details about the neural network-based base reader 102 can be found in U.S. Provisional Patent Application No. 62 / 821,766 entitled “ARTIFICIAL INTELLIGENCE-BASED SEQUENCING”, filed March 21, 2019 (Attorney’s File No. ILLM 1008-9 / IP-1752-PRV), which is incorporated herein by reference.

[0089] Multiple-base pair interpretation

[0090] According to one embodiment, the disclosed technique enables a neural network-based base reader 102 to generate base reads for a given input window that are not only for the central sequencing cycle but also for the flanking sequencing cycles. Specifically, in one embodiment, the disclosed technique simultaneously generates base reads for cycles N, N+1, N-1, N+2, N-2, etc., for a given input window. In other words, a single forward propagation / traversal / base read iteration of the neural network-based base reader 102 generates base reads for multiple sequencing cycles within the input window of the sequencing cycle; this base read is referred to herein as a "many-to-many base read".

[0091] The disclosed technique then uses the disclosed many-to-many base readouts to generate multiple base readouts for the same target sequencing cycle, which appear across multiple sliding input windows. For example, the target sequencing cycle may appear at different positions in the multiple sliding input windows (e.g., starting at position N+2 in the first sliding window, proceeding to position N+1 in the second sliding window, and ending at position N in the third sliding window).

[0092] Multiple base reads are performed on the target sequencing cycle to generate multiple candidate correct base reads for that cycle. The disclosed technique then evaluates these multiple candidate correct base reads as aggregates and determines the final base read for the target sequencing cycle. Aggregate analysis techniques such as averaging, consensus, and weighted consensus can be used to select the final base read for the target sequencing cycle.

[0093] Figure 1A , Figure 1B and Figure 1C The disclosed many-to-many base reading 100 is shown. According to a specific implementation of the disclosed technique, a neural network-based base reader 102 (i.e., base reader 102) processes at least the right-wing input, the center input, and the left-wing input, and produces at least the right-wing output, the center output, and the left-wing output.

[0094] Many-to-many base reading 100 is configured to provide data for n sequencing cycles as input to base reader 102 and generate base readings for any number of cycles out of the n cycles in one iteration of base reading (i.e., one forward pass instance). The target sequencing cycle 108 can be read n times and can appear / occur / fall at various positions in the n base reading iterations.

[0095] Target sequencing cycle 108 can be the central sequencing cycle in some base reading iterations ( Figure 1BIn other iterations, the target sequencing cycle 108 can be a right-wing / context sequencing cycle adjacent to the central sequencing cycle. Figure 1A ), or it could be a left-wing / context sequencing cycle adjacent to the central sequencing cycle ( Figure 1C The offset to the right or left from the center sequencing cycle can also be variable. That is, the target sequencing cycle 108 in n base read iterations can fall at the center position, immediately to the right of the center position, immediately to the left of the center position, any offset to the right of the center position, any offset to the left of the center position, or any other position in the n base read iterations. The base read iterations for the target sequencing cycle can have inputs of sequencing cycles of varying lengths within a given input window of the sequencing cycle, and can also have multiple base read outputs for sequencing cycles of various lengths.

[0096] In one embodiment, the disclosed technique includes a process for accessing a per-cycle analyte channel set generated for a sequencing cycle of a sequencing run; a window of the per-cycle analyte channel set in the process for the sequencing cycle of the sequencing run is processed by a neural network-based base reader 102, such that the neural network-based base reader 102 processes an object window of the per-cycle analyte channel set in the process for an object window of the sequencing cycle of the sequencing run, and the neural network-based base reader 102 generates provisional base read predictions for three or more sequencing cycles in the object window of the sequencing cycle from multiple windows appearing at different positions for a particular sequencing cycle, to generate provisional base read predictions for that particular sequencing cycle; and a base read for that particular sequencing cycle is determined based on the provisional base read predictions.

[0097] In one embodiment, the disclosed technique includes accessing a series of per-cycle analyte channel sets generated for a sequencing cycle of a sequencing run; processing windows of the per-cycle analyte channel sets in the series via a window of the sequencing cycle of the sequencing run using a neural network-based base reader 102, such that the neural network-based base reader 102 processes object windows of the per-cycle analyte channel sets in the series within object windows of the sequencing cycle of the sequencing run, and generates base reading predictions for two or more sequencing cycles within the object windows of the sequencing cycles; and processing multiple windows of the per-cycle analyte channel sets in the series via multiple windows of the sequencing cycle of the sequencing run using the neural network-based base reader 102, and generating outputs for each window within the multiple windows.

[0098] Each of the multiple windows may include a specific set of per-cycle analyte channels for a specific sequencing cycle of the sequencing run. The output for each window of the multiple windows includes (i) a base readout prediction for the specific sequencing cycle and (ii) one or more additional base readout predictions for one or more additional sequencing cycles of the sequencing run, thereby generating multiple base readout predictions for a specific sequencing cycle across the multiple windows (e.g., generated in parallel or simultaneously by the output layer). Finally, the disclosed technique includes determining the base readout for a specific sequencing cycle based on these multiple base readout predictions.

[0099] The right-wing input 132 includes current image data 108 for the current sequencing cycle (e.g., cycle 4) for the sequencing run, supplemented by previous image data 104 and 106 for one or more previous sequencing cycles preceding the current sequencing cycle (e.g., cycles 2 and 3). The right-wing output 142 includes right-wing base readout prediction 114 for the current sequencing cycle and base readout predictions 110 and 112 for previous sequencing cycles.

[0100] The central input 134 includes current image data 108, supplemented by previous image data 106 (e.g., cycle 3) and subsequent image data 116 for one or more subsequent sequencing cycles after the current sequencing cycle (e.g., cycle 5). The central output 144 includes central base readout prediction 120 for the current sequencing cycle and base readout predictions 118 and 122 for previous and subsequent sequencing cycles.

[0101] The left-wing input 136 includes current image data 108, supplemented by subsequent image data 116 and subsequent image data 124. The left-wing output 146 includes left-wing base interpretation prediction 126 for the current sequencing cycle and base interpretation prediction 128 and base interpretation prediction 130 for subsequent sequencing cycles (e.g., cycle 5 and cycle 6).

[0102] Figure 1D and Figure 1E Different examples of the disclosed multi-base pair interpretation are shown. Figure 1D and Figure 1EIn the diagram, blue boxes represent specific or target sequencing cycles (or their data). Specific sequencing cycles are also considered, and the current sequencing cycle is a specific implementation of the disclosed technology. Orange boxes represent sequencing cycles (or their data) that differ from the specific sequencing cycle. Green circles represent one or more base reads generated for a specific sequencing cycle. Base reads can be generated by any base reader (such as Illumina's Real-Time Analysis (RTA) software or the disclosed neural network-based base reader 102). The data used for the sequencing cycle can be images or some other type of input data, such as current readings, voltage changes, pH scale data, etc.

[0103] Go to Figure 1D The first example of many-to-many base interpretation 180 illustrates three base interpretation iterations 180a, 180b, and 180c, and three input windows / groups for the corresponding sequencing cycles w1, w2, and w3 (or their data). In one embodiment, the base interpretation iterations generate base interpretations for each sequencing cycle within the corresponding input windows of the sequencing cycles. In another embodiment, the base interpretation iterations generate base interpretations for only some sequencing cycles (e.g., only specific sequencing cycles) within the corresponding input windows of the sequencing cycles. Moreover, specific sequencing cycles may appear at different positions within the input windows / groups of sequencing cycles w1, w2, and w3. In other embodiments (not shown), two or more input windows / groups of sequencing cycles may have specific sequencing cycles at the same position. Furthermore, the input windows / groups of sequencing cycles w1, w2, and w3 have specific sequencing cycles as at least one overlapping cycle, and also have one or more non-overlapping cycles. That is, the orange boxes at different positions within the different input windows / groups of sequencing cycles represent different non-overlapping cycles. Finally, three base read iterations 180a, 180b, and 180c generate three base reads (e.g., three green circles) for a specific sequencing cycle. These three base reads can be considered provisional base reads and are subsequently analyzed as aggregates for the final base read for that specific sequencing cycle. Figure 2 , Figure 3 and Figure 4 The following describes different examples of the analysis.

[0104] The second and third examples of many-to-many base reads 181 and 182 illustrate that a particular sequencing cycle can be located anywhere in the input window / group of the sequencing cycle and has any number of right-wing and left-wing cycles, or no wing cycles at all (e.g., the third window (w3) in the third base many-to-many base read example 182). Three-base read iterations 181a, 181b, and 181c generate three base reads (e.g., three green circles) for a particular sequencing cycle. These three base reads can be considered provisional base reads and are subsequently analyzed as aggregates for the final base read for that particular sequencing cycle. Figure 2 , Figure 3 and Figure 4 The following describes different examples of the analysis. Three base read iterations 182a, 182b, and 182c generate three base reads (e.g., three green circles) for a specific sequencing cycle. These three base reads can be considered provisional base reads and are subsequently analyzed as aggregates for the final base read for that specific sequencing cycle. Figure 2 , Figure 3 and Figure 4 The following describes different examples of the analysis.

[0105] Figure 1E An example 183 of many-to-many base reads is shown, comprising five base read iterations 183a-183e. Each base read iteration generates a base read prediction for a specific sequencing cycle by processing five corresponding windows / sets / groups of input data, where data for that specific sequencing cycle appears in different locations. The five base read iterations 183a-183e generate five base reads (e.g., five green circles) for a specific sequencing cycle. These five base reads can be considered provisional base reads and are subsequently analyzed as aggregates for a final base read for that specific sequencing cycle. Figure 2 , Figure 3 and Figure 4 The following describes different examples of the analysis.

[0106] Figure 2 , Figure 3 and Figure 4 Different specific implementations of a base readout generator are illustrated. Base readout generator 202 (e.g., running on a host processor) (e.g., via a PCIe bus, Ethernet, or InfiniBand (IB)) is coupled to a neural network-based base readout generator 102 (e.g., running on a chip) and is configured to generate base readouts for the current sequencing cycle (e.g., cycle 4) based on right-wing base readout predictions, center base readout predictions, and left-wing base readout predictions for the current sequencing cycle.

[0107] The current image data used for the current sequencing cycle depicts the intensity emission of the analyte captured in the current sequencing cycle and its surrounding background. Right-wing base readout prediction 114, center base readout prediction 120, and left-wing base readout prediction 126 for the current sequencing cycle (e.g., cycle 4) identify the probability that one or more analyte bases incorporated into the analyte are A, C, T, and G at the current sequencing cycle. In one embodiment, this probability is an exponentially normalized fraction generated by a softmax layer used as the output layer of base reader 102.

[0108] In one embodiment, right-wing base readout prediction 114 for the current sequencing cycle considers the predetermined phase effect between the current sequencing cycle (e.g., cycle 4) and previous sequencing cycles. In one embodiment, center-wing base readout prediction 120 for the current sequencing cycle (e.g., cycle 4) considers the predetermined phase effect between the current sequencing cycle and previous sequencing cycles, as well as the phase-fixing effect between the current sequencing cycle and subsequent sequencing cycles. In one embodiment, left-wing base readout prediction 126 for the current sequencing cycle (e.g., cycle 4) considers the phase-fixing effect between the current sequencing cycle and subsequent sequencing cycles.

[0109] like Figure 2 As shown, the base interpretation generator is further configured to include an averager 204 that sums base-by-base the probabilities of right-wing base interpretation prediction 114, center base interpretation prediction 120, and left-wing base interpretation prediction 126 for the current sequencing cycle (e.g., cycle 4); determines a base-by-base average 212 based on the base-by-base summation; and generates a base interpretation 214 for the current sequencing cycle (e.g., cycle 4) based on the highest average of the base-by-base averages (e.g., 0.38).

[0110] like Figure 3 As shown, the base readout generator is further configured to include a cosenter 304 that determines a preliminary base readout for each of the right-wing base readout predictions 114, the center base readout prediction 120, and the left-wing base readout prediction 126 for the current sequencing cycle (e.g., cycle 4) based on the highest probability among the possibilities, thereby generating a preliminary base readout sequence 306, and generating a base readout for the current sequencing cycle based on the most common base readout 308 in the preliminary base readout sequence.

[0111] like Figure 4As shown, the base readout generator is further configured to include a weighted sensor 404 that determines preliminary base readouts for each of the right-wing, center, and left-wing base readout predictions for the current sequencing cycle based on the highest probability among the possibilities, thereby generating a preliminary base readout sequence 406; applying a base-by-base weight 408 to a corresponding preliminary base readout in the preliminary base readout sequence, and generating a weighted preliminary base readout sequence 410; and generating a base readout for the current sequencing cycle (e.g., cycle 4) based on the most weighted base readout 412 in the weighted preliminary base readout sequence. In some embodiments, the base-by-base weight 408 is preset, for example, on a cycle-by-cycle basis. In other embodiments, for example, the base-by-base weight 408 is learned using a least squares method.

[0112] exist Figure 6 In one specific embodiment shown, the disclosed technique includes accessing current image data for the current sequencing cycle used in the sequencing run, previous image data for one or more previous sequencing cycles preceding the current sequencing cycle, and subsequent image data for one or more subsequent sequencing cycles following the current sequencing cycle (Action 602); processing different groups of the current image data, previous image data, and subsequent image data by a neural network-based base reader, and generating a first base readout prediction, a second base readout prediction, and a third base readout prediction for the current sequencing cycle (Action 612); and generating a base readout for the current sequencing cycle based on the first base readout prediction, the second base readout prediction, and the third base readout prediction (Action 622).

[0113] In one specific implementation, the different groups include a first group, a second group, and a third group. The first group includes current image data and previous image data; the second group includes current image data, previous image data, and subsequent image data; and the third group includes current image data and subsequent image data.

[0114] In one specific implementation, the disclosed technique includes processing a first group of bases using a neural network-based base reader to generate a first base read prediction; processing a second group of bases using a neural network-based base reader to generate a second base read prediction; and processing a third group of bases using a neural network-based base reader to generate a third base read prediction.

[0115] In one specific implementation, the first base readout prediction, the second base readout prediction, and the third base readout prediction for the current sequencing cycle identify the probability that one or more bases in the analyte incorporated into the analyte are A, C, T, and G at the current sequencing cycle.

[0116] In one specific implementation, the disclosed technique includes generating a base readout for the current sequencing cycle by summing base-by-base a first base readout prediction, a second base readout prediction, and a third base readout prediction for the current sequencing cycle; determining a base-by-base average value based on the base-by-base summation; and generating a base readout for the current sequencing cycle based on the highest of the base-by-base average values.

[0117] In one specific implementation, the disclosed technique includes generating a base read for the current sequencing cycle by determining a preliminary base read based on the highest probability among the first, second, and third base read predictions, thereby generating a preliminary base read sequence; and generating a base read for the current sequencing cycle based on the most common base read in the preliminary base read sequence.

[0118] In one specific implementation, the disclosed technique includes generating a base read for the current sequencing cycle by determining a preliminary base read for each of the first, second, and third base read predictions based on the highest probability among the probabilities, thereby generating a preliminary base read sequence; applying a base-by-base weight to a corresponding preliminary base read in the preliminary base read of the preliminary base read sequence to generate a weighted preliminary base read sequence; and generating a base read for the current sequencing cycle based on the most weighted base read in the weighted preliminary base read sequence.

[0119] In one specific implementation known as “multi-cycle training, single-cycle inference”, base reader 102 is trained to generate two or more base read predictions for two or more sequencing cycles during training using a base reader generator, but only generates base read predictions for a single sequencing cycle during inference.

[0120] In one specific implementation known as “multi-cycle training, multi-cycle inference”, base reader 102 is trained to generate two or more base reading predictions for two or more sequencing cycles during training, and the same operation is performed during inference using base reader generator 202.

[0121] Multi-cycle gradient backpropagation

[0122] Figure 5 A specific implementation of the disclosed "Multi-cyclic Gradient Backpropagation 500" is shown. For example... Figure 5As shown, the many-to-many base interpretation 100 is further configured to include a trainer that calculates errors 512, 532, and 552 between base interpretations generated by the base interpretation generator 202 for the current sequencing cycle (e.g., cycle 3), the previous sequencing cycle (e.g., cycle 2), and the subsequent sequencing cycle (e.g., cycle 4) based on the right wing output 506, the center output 504, and the left wing output 502 of the neural network-based base interpreter 102 and the corresponding benchmark true base interpretations 554, 534, and 514. Based on the errors, the trainer determines corresponding gradients 542, 522, and 562 for the current sequencing cycle, the previous sequencing cycle, and the subsequent sequencing cycle, and updates the parameters of the neural network-based base interpreter by backpropagating the gradients.

[0123] Technical effects / advantages

[0124] Figure 7 The technical effects and advantages of the disclosed technology are shown.

[0125] "Multi-loop training and single-loop inference" are specifically implemented in Figure 7 It is known as the "DL 3C intermediate value" and improves the base interpretation error rate by 8% through real-time base interpretation software based on traditional non-neural network analysis.

[0126] "Multi-loop training and multi-loop inference" are specifically implemented in Figure 7 The value is referred to as the "DL 3C average value" and the base interpretation error rate was improved by another 8% in the specific implementation of the DL 3C intermediate value.

[0127] The base interpretation sequencing cycle is repeated multiple times to improve the accuracy of base interpretation, and to detect and analyze base interpretation differences and fuzzy base interpretations.

[0128] Multi-cycle gradient backpropagation also improves the gradient of base reader 102 and its base reading accuracy throughout the training task of base reading.

[0129] sequencing system

[0130] Figure 8A and Figure 8B A specific implementation of a sequencing system 800A is described. The sequencing system 800A includes a configurable processor 846. The configurable processor 846 implements the base interpretation technology disclosed herein. The sequencing system is also referred to as a "sequencing instrument".

[0131] The sequencing system 800A can be operated to obtain any information or data relating to at least one of biological or chemical substances. In some specific implementations, the sequencing system 800A is a workstation that may resemble a desktop device or desktop computer. For example, most (or all) of the systems and components used to perform the desired reaction may be located within a common housing 802.

[0132] In certain specific implementations, the Sequencing System 800A is a nucleic acid sequencing system configured for a variety of applications, including but not limited to de novo sequencing, resequencing of whole or target genome regions, and metagenomics. The sequencer can also be used for DNA or RNA analysis. In some implementations, the Sequencing System 800A can also be configured to generate reaction sites in biosensors. For example, the Sequencing System 800A can be configured to receive a sample and generate surface-attached clusters of clonal amplified nucleic acids derived from the sample. Each cluster can constitute a reaction site in a biosensor or be part of it.

[0133] An exemplary sequencing system 800A may include a system socket or interface 810 configured to interact with a biosensor 812 to perform a desired reaction within the biosensor 812. (The following is about...) Figure 8A In the description, the biosensor 812 is loaded into the system socket 810. However, it should be understood that a cartridge including the biosensor 812 can be inserted into the system socket 810, and in some states, the cartridge can be temporarily or permanently removed. As described above, among other things, the cartridge may also include fluid control components and fluid storage components.

[0134] In a particular implementation, the sequencing system 800A is configured to perform a large number of parallel reactions within a biosensor 812. The biosensor 812 includes one or more reaction sites where the desired reaction can occur. These reaction sites may be, for example, fixed to a solid surface of the biosensor or to a bead (or other movable substrate) located within a corresponding reaction chamber of the biosensor. The reaction sites may include, for example, clusters of cloned and amplified nucleic acids. The biosensor 812 may include a solid-state imaging device (e.g., a CCD or CMOS imaging device) and a flow cell mounted thereon. The flow cell may include one or more flow channels that receive a solution from the sequencing system 800A and direct the solution to the reaction sites. Optionally, the biosensor 812 may be configured to incorporate a thermal element for transferring heat energy into or out of the flow channels.

[0135] The sequencing system 800A may include various components, parts, and systems (or subsystems) that interact with each other to perform predetermined methods or assay protocols for biological or chemical analysis. For example, the sequencing system 800A includes a system controller 806 that can communicate with the various components, parts, and subsystems of the sequencing system 800A, as well as the biosensor 812. For example, in addition to the system socket 810, the sequencing system 800A may also include a fluid control system 808 to control the flow of fluids throughout the fluid network of the sequencing system 800A and the biosensor 812; a fluid storage system 814 configured to contain all fluids (e.g., gases or liquids) that can be used by the bioassay system; a temperature control system 804 that can regulate the temperature of the fluids in the fluid network, the fluid storage system 814, and / or the biosensor 812; and an illumination system 816 configured to illuminate the biosensor 812. As described above, if a cartridge having the biosensor 812 is loaded into the system socket 810, the cartridge may also include fluid control components and fluid storage components.

[0136] As also shown in the figure, the sequencing system 800A may include a user interface 818 for user interaction. For example, the user interface 818 may include a display 820 for displaying or requesting information from the user and a user input device 822 for receiving user input. In some specific implementations, the display 820 and the user input device 822 are the same device. For example, the user interface 818 may include a touch-sensitive display configured to detect the presence of an individual touch and also identify the location of the touch on the display. However, other user input devices 822 may be used, such as a mouse, touchpad, keyboard, keypad, handheld scanner, voice recognition system, motion recognition system, etc. As will be discussed in more detail below, the sequencing system 800A may communicate with various components, including biosensors 812 (e.g., in the form of cartridges), to perform desired responses. The sequencing system 800A may also be configured to analyze data obtained from the biosensors to provide the user with the required information.

[0137] System controller 806 may include any processor-based or microprocessor-based system, including those using microcontrollers, reduced instruction set computers (RISCs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), coarse-grained reconfigurable architectures (CGRAs), logic circuits, and any other circuitry or processor capable of performing the functions described herein. The examples above are merely illustrative and are therefore not intended to limit the definition and / or meaning of the term "system controller" in any way. In an exemplary specific implementation, system controller 806 executes a set of instructions stored in one or more storage elements, memories, or modules to perform at least one of acquiring detection data and analyzing the detection data. Detection data may include multiple pixel signal sequences, such that pixel signal sequences from each of millions of sensors (or pixels) can be detected within a plurality of base interpretation cycles. Storage elements may be in the form of information sources or physical memory elements within the sequencing system 800A.

[0138] The instruction set may include various commands instructing the sequencing system 800A or biosensor 812 to perform specific operations (such as the methods and procedures described herein in various specific embodiments). The instruction set may be in the form of a software program that forms part of one or more tangible, non-transitory computer-readable media. As used herein, the terms “software” and “firmware” are interchangeable and include any computer program stored in memory for execution by a computer, including RAM memory, ROM memory, EPROM memory, EEPROM memory, and non-volatile RAM (NVRAM) memory. The above memory types are merely exemplary and therefore do not limit the types of memory that can be used to store computer programs.

[0139] The software can take various forms, such as system software or application software. Furthermore, the software can be a collection of independent programs, or a program module or part of a program module within a larger program. The software may also include modular programming in the form of object-oriented programming. After the detection data is obtained, the detection data can be processed automatically by the sequencing system 800A in response to user input or in response to a request from another processor (e.g., a remote request via a communication link). In the illustrated embodiment, the system controller 806 includes an analysis module 844. In other embodiments, the system controller 806 does not include the analysis module 844, but the analysis module 844 can be accessed (e.g., the analysis module 844 may be hosted separately in the cloud).

[0140] The system controller 806 can be connected to the biosensor 812 and other components of the sequencing system 800A via a communication link. The system controller 806 can also be communicatively connected to off-site systems or servers. The communication link can be hardwired, wired, or wireless. The system controller 806 can receive user input or commands from the user interface 818 and user input device 822.

[0141] The fluid control system 808 includes a fluid network and is configured to guide and regulate the flow of one or more fluids through the fluid network. The fluid network may be in fluid communication with a biosensor 812 and a fluid storage system 814. For example, selected fluid may be drawn from the fluid storage system 814 and guided in a controlled manner to the biosensor 812, or fluid may be drawn from the biosensor 812 and directed toward, for example, a waste reservoir in the fluid storage system 814. Although not shown, the fluid control system 808 may include a flow sensor that detects the velocity or pressure of the fluid within the fluid network. The sensor may communicate with a system controller 806.

[0142] Temperature control system 804 is configured to regulate the temperature of fluids at different regions of the fluid network, fluid storage system 814, and / or biosensor 812. For example, temperature control system 804 may include a thermal cycler that interfaces with biosensor 812 and controls the temperature of fluid flowing along reaction sites in biosensor 812. Temperature control system 804 may also regulate the temperature of solid-state elements or components of sequencing system 800A or biosensor 812. Although not shown, temperature control system 804 may include sensors for detecting the temperature of fluids or other components. The sensors may communicate with system controller 806.

[0143] The fluid storage system 814 is in fluid communication with the biosensor 812 and can store various reaction components or reactants for carrying out the desired reaction therein. The fluid storage system 814 can also store fluids for washing or cleaning the fluid network and biosensor 812, as well as for diluting reactants. For example, the fluid storage system 814 may include various reservoirs for storing samples, reagents, enzymes, other biomolecules, buffer solutions, aqueous solutions, and nonpolar solutions, etc. Furthermore, the fluid storage system 814 may include a waste reservoir for receiving waste from the biosensor 812. In embodiments including a cartridge, the cartridge may include one or more of a fluid storage system, a fluid control system, or a temperature control system. Therefore, one or more components relating to those systems described herein may be housed within a cartridge housing. For example, the cartridge may have various reservoirs for storing samples, reagents, enzymes, other biomolecules, buffer solutions, aqueous solutions, and nonpolar solutions, waste, etc. Therefore, one or more of the fluid storage system, fluid control system, or temperature control system may be removably coupled to the bioassay system via the cartridge or other biosensor.

[0144] The illumination system 816 may include a light source (e.g., one or more LEDs) and multiple optical components for illuminating the biosensor. Examples of light sources may include lasers, arc lamps, LEDs, or laser diodes. Optical components may be, for example, reflectors, dichroic mirrors, beam splitters, collimators, lenses, filters, wedge mirrors, prisms, mirrors, detectors, etc. In a specific embodiment using the illumination system, the illumination system 816 may be configured to direct excitation light to the reaction site. As an example, the fluorophore may be excited by light of a green wavelength, thus the wavelength of the excitation light may be approximately 532 nm. In one embodiment, the illumination system 816 is configured to produce illumination parallel to the surface normal of the surface of the biosensor 812. In another embodiment, the illumination system 816 is configured to produce illumination at an angle relative to the surface normal of the surface of the biosensor 812. In yet another embodiment, the illumination system 816 is configured to produce illumination with multiple angles, including some parallel illumination and some angled illumination.

[0145] System socket or interface 810 is configured to engage biosensor 812 in at least one of mechanical, electrical, and fluidic methods. System socket 810 can hold biosensor 812 in a desired orientation to facilitate fluid flow through biosensor 812. System socket 810 may also include electrical contacts configured to engage biosensor 812, enabling sequencing system 800A to communicate with and / or power biosensor 812. Furthermore, system socket 810 may include a fluid port (e.g., a nozzle) configured to engage biosensor 812. In some embodiments, biosensor 812 is removably coupled to system socket 810 in mechanical, electrical, and fluidic methods.

[0146] Furthermore, the sequencing system 800A can communicate remotely with other systems or networks or with other bioassay systems 800A. Detection data obtained by the bioassay system 800A can be stored in a remote database.

[0147] Figure 8B It is possible Figure 8A The block diagram of the system controller 806 used in the system is shown below. In one specific implementation, the system controller 806 includes one or more processors or modules that can communicate with each other. Each of the processors or modules may include an algorithm (e.g., instructions stored on a tangible and / or non-transitory computer-readable storage medium) or a sub-algorithm for performing a particular process. The system controller 806 is conceptually shown as a collection of modules, but can be implemented using any combination of dedicated hardware boards, DSPs, processors, etc. Alternatively, the system controller 806 can be implemented using an off-the-shelf PC with a single processor or multiple processors, where functional operations are distributed among the processors. As a further alternative, the modules described below can be implemented using a hybrid configuration, where some modular functions are executed using dedicated hardware, while other modular functions are executed using an off-the-shelf PC, etc. Modules can also be implemented as software modules within a processing unit.

[0148] During operation, communication port 850 can communicate with biosensor 812 ( Figure 8A ) and / or subsystems 808, 814, 804 ( Figure 8A It transmits information (e.g., commands) or receives information (e.g., data) from or from the communication port 850. In a specific implementation, the communication port 850 can output multiple pixel signal sequences. The communication link 834 can be connected from the user interface 818 ( Figure 8AThe system receives user input and transmits data or information to the user interface 818. Data from the biosensor 812 or subsystems 808, 814, 804 can be processed in real time by the system controller 806 during a bioassay session. Alternatively, data can be temporarily stored in the system memory during a bioassay session and processed at a slower rate than in real-time or offline operation.

[0149] like Figure 8B As shown, the system controller 806 may include multiple modules 824-848 that communicate with the main control module 824 and the central processing unit (CPU) 852. The main control module 824 may communicate with the user interface 818. Figure 8A Communication. Although modules 824-848 are shown communicating directly with the main control module 824, modules 824-848 can also communicate directly with each other, and directly with the user interface 818 and the biosensor 812. Additionally, modules 824-848 can communicate with the main control module 824 through other modules.

[0150] Multiple modules 824-848 include system modules 828-832, 826 that communicate with subsystems 808, 814, 804, and 816, respectively. Fluid control module 828 can communicate with fluid control system 808 to control valves and flow sensors in the fluid network, thereby controlling the flow of one or more fluids through the fluid network. Fluid storage module 830 can notify the user when the fluid level is low or when the waste storage tank is at or near its capacity. Fluid storage module 830 can also communicate with temperature control module 832 to allow fluid to be stored at a desired temperature. Illumination module 826 can communicate with illumination system 816 to illuminate the reaction site at a specified time during the protocol, such as after a desired reaction (e.g., a binding event) has occurred. In some embodiments, illumination module 826 can communicate with illumination system 816 to illuminate the reaction site at a specified angle.

[0151] Multiple modules 824-848 may further include a device module 836 that communicates with the biosensor 812 and an identification module 838 that determines identification information associated with the biosensor 812. The device module 836 may, for example, communicate with a system socket 810 to confirm that the biosensor has established electrical and fluid connections with the sequencing system 800A. The identification module 838 may receive signals that identify the biosensor 812. The identification module 838 may use the identity of the biosensor 812 to provide additional information to the user. For example, the identification module 838 may determine and subsequently display a batch number, manufacturing date, or suggested protocols for operation with the biosensor 812.

[0152] Multiple modules 824-848 also include an analysis module 844 (also referred to as a signal processing module or signal processor) for receiving and analyzing signal data (e.g., image data) from the biosensor 812. Analysis module 844 includes memory (e.g., RAM or flash memory) for storing the detection / image data. The detection data may include multiple pixel signal sequences, enabling the detection of pixel signal sequences from each of millions of sensors (or pixels) within numerous base interpretation cycles. The signal data may be stored for subsequent analysis or may be transmitted to user interface 818 to display desired information to the user. In some specific implementations, the signal data may be processed by a solid-state imaging device (e.g., a CMOS image sensor) before being received by analysis module 844.

[0153] Analysis module 844 is configured to acquire image data from a photodetector at each of a plurality of sequencing cycles. The image data originates from the emission signal detected by the photodetector and is processed by base interpreter 102 for each of the plurality of sequencing cycles, generating base interpretations for at least some of the analytes at each of the plurality of sequencing cycles. The photodetector may be part of one or more top-view cameras (e.g., an Illumina GAIIx CCD camera that captures images of the clusters on biosensor 812 from above), or it may be part of biosensor 812 itself (e.g., an Illumina iSeq CMOS image sensor located below the clusters on biosensor 812 and capturing images of the clusters from the bottom).

[0154] The output of the photodetector is a sequencing image, each depicting the intensity emission of the cluster and its surrounding background. The sequencing image depicts the intensity emission resulting from nucleotide incorporation into the sequence during sequencing. The intensity emission originates from the associated analyte and its surrounding background. The sequencing images are stored in memory 848.

[0155] Protocol modules 840 and 842 communicate with main control module 824 to control the operation of subsystems 808, 814, and 804 during a predetermined assay protocol. Protocol modules 840 and 842 may include a set of instructions for instructing sequencing system 800A to perform specific operations according to a predetermined protocol. As shown, the protocol module may be a sequencing-by-synthesis (SBS) module 840, configured to issue various commands for performing the sequencing-by-synthesis process. In SBS, the extension of nucleic acid primers along a nucleic acid template is monitored to determine the sequence of nucleotides in the template. The underlying chemical process may be polymerization (e.g., catalyzed by a polymerase) or ligation (e.g., catalyzed by a ligase). In a specific polymerase-based SBS implementation, fluorescently labeled nucleotides are added to primers in a template-dependent manner (causing primer extension), such that detection of the sequence and type of nucleotides added to the primers can be used to determine the sequence of the template. For example, to initiate a first SBS cycle, a command may be issued to deliver one or more labeled nucleotides, DNA polymerase, etc., to / through a flow cell containing an array of nucleic acid templates. The nucleic acid template may be located at the corresponding reaction site. Those reaction sites where primer extension leads to the incorporation of labeled nucleotides can be detected by an imaging event. During the imaging event, the illumination system 816 may provide excitation light to the reaction site. Optionally, the nucleotide may also include a reversible termination property that terminates further primer extension once the nucleotide is added to the primer. For example, a nucleotide analog with a reversible termination moiety may be added to the primer such that subsequent extension does not occur until a deblocking agent is delivered to remove that moiety. Thus, for a specific implementation using reversible termination, a command may be issued to deliver the deblocking agent to the flow cell (before or after detection occurs). One or more commands may be issued to perform washing between the various delivery steps. This cycle can then be repeated n times to extend the primer by n nucleotides, thereby detecting a sequence of length n. Exemplary sequencing techniques are described in, for example, Bentley et al., Nature 456:53-59 (2008), WO 04 / 018497, US 7,057,026, WO 91 / 06678, WO 07 / 123744, US 7,329,492, US 7,211,414, US 7,315,019, US 7,405,281 and US 2008 / 014708082, each of which is incorporated herein by reference.

[0156] For the nucleotide delivery step in the SBS cycle, a single type of nucleotide can be delivered at once, or multiple different nucleotide types (e.g., A, C, T, and G together) can be delivered. For nucleotide delivery configurations where only a single type of nucleotide is present at a time, the different nucleotides do not need to have different labels, as they can be distinguished based on the inherent time intervals in individualized delivery. Therefore, sequencing methods or apparatus can use monochromatic detection. For example, the excitation source only needs to provide excitation at a single wavelength or within a single wavelength range. For nucleotide delivery configurations where delivery results in the simultaneous presence of multiple different nucleotides in the flow cell, the sites incorporating different nucleotide types can be distinguished based on different fluorescent labels attached to the corresponding nucleotide types in the mixture. For example, four different nucleotides can be used, each with one of four different fluorophores. In one specific implementation, excitation in four different regions of the spectrum can be used to distinguish the four different fluorophores. For example, four different excitation radiation sources can be used. Alternatively, fewer than four different excitation sources can be used, but optical filtering of excitation radiation from a single source can be used to generate different ranges of excitation radiation at the flow cell.

[0157] In some specific implementations, fewer than four different colors can be detected in a mixture containing four different nucleotides. For example, nucleotide pairs can be detected at the same wavelength, but distinguished based on the intensity difference of one member relative to the other, or based on a change in the presence or absence of a signal in one member that results in a signal that is significantly present or absent compared to the detected signal of the other member of that pair (e.g., by chemical modification, photochemical modification, or physical modification). Exemplary apparatuses and methods for distinguishing four different nucleotides using detection with fewer than four colors are described, for example, in U.S. Patent Application Serial Nos. 61 / 538,294 and 61 / 619,878, which are incorporated herein by reference in their entirety. U.S. Application 13 / 624,200, filed September 21, 2012, is also incorporated herein by reference in its entirety.

[0158] Multiple protocol modules may also include a sample preparation (or generation) module 842, configured to command a fluid control system 808 and a temperature control system 804 to amplify the product within the biosensor 812. For example, the biosensor 812 may be coupled to a sequencing system 800A. The amplification module 842 may instruct the fluid control system 808 to deliver necessary amplification components to the reaction chamber within the biosensor 812. In other embodiments, the reaction site may already contain components for amplification, such as template DNA and / or primers. After the amplification components are delivered to the reaction chamber, the amplification module 842 may instruct the temperature control system 804 to cycle through different temperature phases according to a known amplification protocol. In some embodiments, amplification and / or nucleotide incorporation are performed isothermally.

[0159] The SBS module 840 can issue commands to perform bridged PCR, in which clusters of cloned amplicones are formed on localized regions within the channels of the flow cell. After amplicon generation via bridged PCR, the amplicones can be "linearized" to prepare single-stranded template DNA or sstDNA, and sequencing primers can be hybridized to universal sequences side-linked to regions of interest. For example, sequencing-by-synthesis methods based on reversible terminators can be used as described above or as follows.

[0160] Each base readout or sequencing cycle can be performed by extending the sstDNA with a single base, which can be accomplished, for example, by using a modified DNA polymerase and a mixture of four types of nucleotides. The different types of nucleotides can have unique fluorescent labels, and each nucleotide can also have a reversible terminator that allows only single-base incorporation in each cycle. After adding a single base to the sstDNA, excitation light can be incident on the reaction site and fluorescence emission can be detected. After detection, the fluorescent label and terminator can be chemically cleaved from the sstDNA. This can then be followed by another similar base readout or sequencing cycle. In such sequencing protocols, the SBS module 840 can instruct the fluid control system 808 to guide the flow of reagents and enzyme solutions through the biosensor 812. Exemplary SBS methods based on reversible terminators that can be used with the devices and methods described herein are described in U.S. Patent Application Publication 2007 / 0166705 A1, U.S. Patent Application Publication 2006 / 0188901 A1, U.S. Patent 7,057,026, U.S. Patent Application Publication 2006 / 0240439 A1, U.S. Patent Application Publication 2006 / 02814714709 A1, PCT Publication WO 05 / 065814, U.S. Patent Application Publication 2005 / 014700900 A1, PCT Publication WO 06 / 08B199, and PCT Publication WO 07 / 01470251, each of which is incorporated herein by reference in its entirety. Exemplary reagents for SBS based on reversible terminators are described in: US 7,541,444, US 7,057,026, US 7,414,14716, US 7,427,673, US 7,566,537, US 7,592,435 and WO 07 / 14835368, each of which is incorporated herein by reference in its entirety.

[0161] In some implementations, the amplification module and the SBS module can operate in a single assay protocol, where, for example, template nucleic acids are amplified and subsequently sequenced within the same kit.

[0162] The sequencing system 800A can also allow users to reconfigure the assay protocol. For example, the sequencing system 800A can provide users with options for modifying the determined protocol through the user interface 818. For example, if it is determined that the biosensor 812 will be used for amplification, the sequencing system 800A can request the temperature of the annealing cycle. Furthermore, if the user has provided user input that is generally unacceptable for the selected assay protocol, the sequencing system 800A can issue a warning to the user.

[0163] In a specific implementation, the biosensor 812 comprises millions of sensors (or pixels), each of which generates multiple pixel signal sequences in a subsequent base interpretation cycle. The analysis module 844 detects the multiple pixel signal sequences based on the row-by-row and / or column-by-column positions of the sensors on the sensor array and assigns them to the corresponding sensors (or pixels).

[0164] Figure 9 This is a simplified block diagram of a system for analyzing sensor data (such as base reading sensor output) from a sequencing system 800A. Figure 9 In the example, the system includes a configurable processor 846. The configurable processor 846 can execute a base reader (e.g., a neural network-based base reader 102) in coordination with a runtime program executed by a central processing unit (CPU) 852 (i.e., the host processor). The sequencing system 800A includes a biosensor 812 and a flow cell. The flow cell may include one or more blocks in which clusters of genetic material are exposed to a sequence of analytical streams, the sequence of which is used to elicit a reaction in the clusters to identify bases in the genetic material. The sensor senses the reaction of each cycle of that sequence in each block of the flow cell to provide block data. Genetic sequencing is a data-intensive operation that converts base reading sensor data into base reading sequences for each cluster of genetic material sensed during a base reading operation.

[0165] The system in this example includes a CPU 852 that executes runtime programs to coordinate base read operations, a memory 848B for storing sequences of block data arrays, base read segments generated by the base read operations, and other information used in the base read operations. Additionally, in this illustration, the system includes a memory 848A for storing configuration files (or multiple files) such as FPGA bit files and model parameters for configuring and reconfiguring the neural network of the configurable processor 846, and for executing the neural network. The sequencing system 800A may include programs for configuring the configurable processor and, in some embodiments, a reconfigurable processor to execute the neural network.

[0166] The sequencing system 800A is coupled to the configurable processor 846 via bus 902. Bus 902 can be implemented using high-throughput technologies, such as, in one example, bus technologies compatible with the PCIe standard (Rapid Peripheral Component Interconnect) currently maintained and developed by the PCI-SIG (PCI Special Interest Group). Also in this example, memory 848A is coupled to the configurable processor 846 via bus 906. Memory 848A can be on-board memory disposed on a circuit board with the configurable processor 846. Memory 848A is used by the configurable processor 846 for high-speed access to working data used in base determination operations. Bus 906 can also be implemented using high-throughput technologies such as bus technologies compatible with the PCIe standard.

[0167] Configurable processors, including Field Programmable Gate Arrays (FPGAs), Coarse-grained Reconfigurable Arrays (CGRAs), and other configurable and reconfigurable devices, can be configured to perform a variety of functions more efficiently or faster than could be achieved using a general-purpose processor executing a computer program. Configuring a configurable processor involves compiling a functional description to produce a configuration file, sometimes called a bitstream or bitfile, and distributing the configuration file to configurable elements on the processor. This configuration file defines the logical functions to be performed by the configurable processor by configuring circuitry to set data flow patterns, the use of distributed memory and other on-chip memory resources, lookup table contents, the operation of configurable logic blocks and configurable execution units (such as multiply-accumulate units, configurable interconnects, and other elements of the configurable array). A configurable processor is reconfigurable if the configuration file can be changed in the field by changing the loaded configuration file. For example, the configuration file can be stored in volatile SRAM elements, non-volatile read-write memory elements, and combinations thereof, distributed across an array of configurable elements on the configurable or reconfigurable processor. Many commercially available configurable processors are suitable for base-reading operations as described herein. Examples include Google's Tensor Processing Unit (TPU). ™ Rackmount solutions (such as GX4 Rackmount Series) ™ GX9 Rackmount Series ™ NVIDIA DGX-1 ™ Microsoft's Stratix V FPGA ™ , Graphcore’s Intelligent Processor Unit (IPU) ™ Qualcomm's Snapdragon processors ™ Zeroth Platform ™NVIDIA's Volta ™ NVIDIA's Drive PX ™ , NVIDIA's JETSON TX1 / TX2MODULE ™ Intel's Nirvana ™ Movidius VPU ™ ,Fujitsu DPI ™ ARM's DynamicIQ ™ IBM TrueNorth ™ It has Testa V100s ™ Lambda GPU servers, Xilinx Alveo ™ U200, Xilinx Alveo ™ U250, Xilinx Alveo ™ U280, Intel / Altera Stratix ™ GX2800, Intel / Altera Stratix ™ GX2800 and Intel Stratix ™ GX10M. In some examples, the host CPU can be implemented on the same integrated circuit as the configurable processor.

[0168] The implementation described herein uses a configurable processor 846 to implement a neural network-based base reader 102. The configuration file for the configurable processor 846 can be implemented by specifying the logic functions to be performed using a high-level description language (HDL) or register-transfer level RTL language specification. The specification can be compiled to generate the configuration file using resources designed for the selected configurable processor. The same or similar specifications can be compiled to generate designs for application-specific integrated circuits (ASICs) that may not be configurable processors.

[0169] Therefore, in all embodiments described herein, alternative embodiments of the configurable processor 846 include a configured processor comprising an application-specific ASIC or application-specific integrated circuit or integrated circuit group, or a system-on-a-chip (SOC) device, or a graphics processing unit (GPU) processor or a coarse-grained reconfigurable architecture (CGRA) processor, the configured processor being configured to perform the neural network-based base-reading operation as described herein.

[0170] Generally speaking, the configurable processor and the configured processor described herein, which are configured to execute the operation of a neural network, are referred to herein as neural network processors.

[0171] In this example, the configurable processor 846 is configured by a configuration file loaded by a program executed using CPU 852, or by other sources configuring an array of configurable elements 916 (e.g., configuration logic blocks (CLBs), such as lookup tables (LUTs), flip-flops, compute processing units (PMUs) and compute memory units (CMUs), configurable I / O blocks, programmable interconnects) on the configurable processor to perform base-reading functionality. In this example, the configuration includes data flow logic 908 coupled to buses 902 and 906, and performs functions for distributing data and control parameters among the elements used in the base-reading operation.

[0172] Additionally, the configurable processor 846 is configured with base-reading execution data flow logic 908 to execute a neural network-based base reader 102. The data flow logic 908 includes multiple loop execution clusters (e.g., 914), which in this example include execution clusters 1 through X. The number of multiple loop execution clusters can be selected based on a trade-off between the required throughput of the operations involved and the available resources on the configurable processor 846.

[0173] The multi-loop execution cluster is coupled to the data flow logic 908 via a data flow path 910 implemented using configurable interconnects and memory resources on the configurable processor 846. Additionally, the multi-loop execution cluster is coupled to the data flow logic 908 via a control path 912 implemented, for example, using configurable interconnects and memory resources on the configurable processor 846. This control path provides indication of available execution clusters, prepares input units for providing to the available execution clusters for executing the neural network-based base reader 102, prepares to provide trained parameters to the neural network-based base reader 102, prepares to provide output patches of base reading classification data, and other control data for executing the neural network-based base reader 102.

[0174] Configurable processor 846 is configured to execute the operation of neural network-based base reader 102 using trained parameters to generate classification data for sensing cycles of base interpretation operations. The execution of neural network-based base reader 102 generates classification data for subject sensing cycles of base interpretation operations. The operation of neural network-based base reader 102 operates on a sequence comprising a number of N arrays of block data from corresponding sensing cycles of N sensing cycles, wherein the N sensing cycles, in the examples described herein, provide sensor data for a different base interpretation operation at a base position for each operation in the time series. Optionally, some of the N sensing cycles may be out of order, depending on the specific neural network model being executed, if desired. The number N can be any number greater than 1. In some examples described herein, the sensing cycles in the N sensing cycles represent a set of sensing cycles that precede at least one sensing cycle and follow at least one sensing cycle in the time series. Examples described herein are in which the number N is an integer equal to or greater than five.

[0175] Dataflow logic 908 is configured to move at least some trained parameters of block data and model parameters from memory 848A to a configurable processor 846 for the operation of a neural network-based base reader 102 using input units for a given run. These input units include block data of patches for spatial alignment of N arrays. The input units can be moved via a direct memory access operation in a DMA operation, or in smaller units that move in coordination with the execution of the deployed neural network during available time slots.

[0176] Block data for sensing cycles, as described herein, may include an array of sensor data having one or more features. For example, the sensor data may include two images, which are analyzed to identify one of four bases at a base position in the genetic sequence of DNA, RNA, or other genetic material. The block data may also include metadata about the images and the sensors. For example, in an embodiment of a base interpretation operation, the block data may include information about the alignment of the image with the clusters, such as distance from the center, indicating the distance of each pixel in the sensor data array from the center of the cluster of genetic material on the block.

[0177] During the execution of the neural network-based base reader 102 as described below, the block data may also include data generated during the execution of the neural network-based base reader 102, referred to as intermediate data, which can be reused instead of being recalculated during the operation of the neural network-based base reader 102. For example, during the execution of the neural network-based base reader 102, data flow logic 908 may write intermediate data into memory 848A instead of sensor data for a given patch of the block data array. An embodiment similar to this is described in more detail below.

[0178] As shown in the figure, a system for analyzing the output of a base-reading sensor is described. This system includes a memory (e.g., 848A) accessible by a runtime program, which stores block data, including sensor data from blocks of sensing loops of a base-reading operation. Additionally, the system includes a neural network processor, such as a configurable processor 846 with access to the memory. The neural network processor is configured to execute the operation of a neural network using trained parameters to generate classification data for the sensing loops. As described herein, the operation of the neural network operates on a sequence of N arrays of block data from the corresponding sensing loops (including the subject loop) of N sensing loops to generate classification data for the subject loop. Data flow logic 908 is provided to move the block data and trained parameters from the memory to the neural network processor for the operation of the neural network using input units (including data from the spatial alignment patches of the N arrays of the corresponding sensing loops of the N sensing loops).

[0179] Additionally, a system is described in which a neural network processor has access to memory and includes multiple execution clusters configured to execute a neural network. Dataflow logic 908 can access the memory and the execution clusters to provide input units of block data to available execution clusters within the multiple execution clusters. The input units include N digital spatial alignment patches from an array of block data from a corresponding sensing loop (including a subject sensing loop), and cause the execution clusters to apply the N spatial alignment patches to the neural network to produce output patches of classification data for the spatial alignment patches of the subject sensing loop, where N is greater than 1.

[0180] like Figure 9 and Figure 10As shown, in one specific embodiment, the disclosed technology includes an artificial intelligence-based system for base interpretation. The system includes a host processor, a memory, a configurable processor, and data flow logic. The memory, accessible by the host processor, stores image data for sequencing cycles used in sequencing runs, wherein current image data for the current sequencing cycle depicts the intensity emission of the analyte captured in the current sequencing cycle and its surrounding background. The configurable processor, accessible by the memory, includes multiple execution clusters configured to execute neural networks. The data flow logic, accessible by the memory and the execution clusters, is configured to provide the current image data, prior to the current sequencing cycle, to available execution clusters among the multiple execution clusters. Previous image data from one or more previous sequencing cycles, and subsequent image data for one or more subsequent sequencing cycles after the current sequencing cycle, cause the execution cluster to apply different groups of the current image data, previous image data, and subsequent image data to a neural network to generate a first base readout prediction, a second base readout prediction, and a third base readout prediction for the current sequencing cycle, and feed the first base readout prediction, the second base readout prediction, and the third base readout prediction for the current sequencing cycle back to a memory for generating base readouts for the current sequencing cycle based on the first base readout prediction, the second base readout prediction, and the third base readout prediction.

[0181] In one specific implementation, the different groups include a first group, a second group, and a third group. The first group includes current image data and previous image data; the second group includes current image data, previous image data, and subsequent image data; and the third group includes current image data and subsequent image data.

[0182] In one specific implementation, the execution cluster applies a first group to a neural network to produce a first base readout prediction, applies a second group to the neural network to produce a second base readout prediction, and applies a third group to the neural network to produce a third base readout prediction.

[0183] In one specific implementation, the first base readout prediction, the second base readout prediction, and the third base readout prediction for the current sequencing cycle identify the probability that one or more bases in the analyte incorporated into the analyte are A, C, T, and G at the current sequencing cycle.

[0184] In one specific implementation, the data flow logic is further configured to generate a base read for the current sequencing cycle by summing the probabilities of the first, second, and third base read predictions across the current sequencing cycle, determining the base-by-base average based on the base-by-base summation, and generating a base read for the current sequencing cycle based on the highest of the base-by-base averages.

[0185] In one specific implementation, the data flow logic is further configured to generate a base read for the current sequencing cycle by determining a preliminary base read for each of the first, second, and third base read predictions based on the highest probability among the probabilities, thereby generating a preliminary base read sequence; and to generate a base read for the current sequencing cycle based on the most common base read in the preliminary base read sequence.

[0186] In one specific implementation, the data flow logic is further configured to generate a base read for the current sequencing cycle by determining a preliminary base read for each of the first, second, and third base read predictions based on the highest probability among the probabilities, thereby generating a preliminary base read sequence; applying a base-by-base weight to a corresponding preliminary base read in the preliminary base read sequence to generate a weighted preliminary base read sequence; and generating a base read for the current sequencing cycle based on the most weighted base read in the weighted preliminary base read sequence.

[0187] Figure 10 This is a simplified diagram illustrating an aspect of the base-reading operation, including the functionality of a runtime program executed by the host processor. In this diagram, the output from the image sensor in the flow cell is provided on line 1000 to image processing thread 1001. This image processing thread can perform processing on the image, such as alignment and arrangement of individual blocks in the sensor data array and image resampling. This can be used by a process that calculates a block cluster mask for each block in the flow cell, identifying pixels in the sensor data array corresponding to clusters of genetic material on the corresponding block in the flow cell. Depending on the state of the base-reading operation, the output of image processing thread 1001 is provided on line 1002 to scheduling logic 1010 in the CPU. This scheduling logic routes the block data array to data cache 1004 (e.g., an SSD storage device) on high-speed bus 1003, or to neural network processor hardware 1020 on high-speed bus 1005, such as… Figure 9The configurable processor 846. Processed and transformed images can be stored on data cache 1004 for use in previously used sensing loops. Hardware 1020 returns classification data output from the neural network to scheduling logic 1010, which passes information to data cache 1004, or on thread 1011 to thread 1002, which performs base interpretation and quality score calculation using the classification data, and can arrange the data for base interpretation reads in a standard format. On thread 1012, the output of thread 1002, which performs base interpretation and quality score calculation, is provided to thread 1003, which aggregates base interpretation reads, performs other operations such as data compression, and writes the resulting base interpretation output to a specified destination for client use.

[0188] In some implementations, the host may include final processing of the output of hardware 1020 to support threads (not shown) of the neural network. For example, hardware 1020 may provide outputs of classification data from the final layer of a multi-cluster neural network. The host processor may perform output activation functions such as softmax on the classification data to configure the data for use by the base interpretation and quality scoring threads 1002. Additionally, the host processor may perform input operations (not shown), such as batch normalization of block data before it is input to hardware 1020.

[0189] Figure 11 It is a configurable processor 846 (such as Figure 9 A simplified diagram of the configuration of the configurable processor. Figure 11 In this configuration, the configurable processor 846 includes an FPGA with multiple high-speed PCIe interfaces. The FPGA is configured with a wrapper 1100, which includes a reference... Figure 9The data flow logic 908 is described. The package 1100 manages the interface and coordination with the runtime program in the CPU via CPU communication link 1109, and manages communication with the onboard DRAM 1102 (e.g., memory 848A) via DRAM communication link 1110. The data flow logic 908 in the package 1100 provides patch data retrieved by traversing a block data array of N cycles on the onboard DRAM 1102 to cluster 1101, and retrieves process data 1115 from cluster 1101 for delivery back to the onboard DRAM 1102. The package 1100 also manages data transfer between the onboard DRAM 1102 and host memory for both the input array of block data and the output patches of classified data. The package transfers patch data on line 1113 to the assigned cluster 1101. The package provides trained parameters such as weights and biases to the cluster 1101 retrieved from the onboard DRAM 1102 on line 1112. The wrapper on line 1111 provides configuration and control data to cluster 1101, which is provided from or generated in response to a runtime program on the host via CPU communication link 1109. The cluster can also provide status signals to wrapper 1100 on line 1116, which are used in conjunction with control signals from the host to manage the traversal of the block data array, thereby providing spatially aligned patch data and performing multi-recurrent neural networks on the patch data using the resources of cluster 1101.

[0190] As described above, multiple clusters can exist on a single configurable processor managed by the wrapper 1100, and these clusters are configured to execute on corresponding patches of multiple patches of block data. Each cluster can be configured to provide classification data for base interpretation in a subject sensing cycle using block data from multiple sensing cycles as described herein.

[0191] In system examples, model data (including kernel data such as filter weights and biases) can be sent from the host CPU to a configurable processor, allowing the model to be updated based on the number of loops. As a representative example, a base-reading operation may include approximately several hundred sensing loops. In some implementations, the base-reading operation may include end-reads. For example, model training parameters may be updated every 20 loops (or another number of loops), or according to an update pattern implemented for a particular system and neural network model. In some implementations that include end-reads, where the sequence of a given string in a genetic cluster on a block comprises a first portion extending down (or up) along the string from a first end and a second portion extending up (or down) along the string from a second end, the trained parameters may be updated during the transition from the first portion to the second portion.

[0192] In some examples, image data for multiple loops in the sensing data used for blocks can be sent from the CPU to the package 1100. The package 1100 may optionally perform some preprocessing and transformation on the sensing data and write the information to the onboard DRAM 1102. The input block data for each sensing loop may include a sensor data array comprising approximately 4000 × 3000 pixels or more per block per sensing loop, wherein two features represent the colors of two images of the block, and each feature is one or two bytes per pixel. For an implementation where the number N is three sensing loops to be used in each run of the multi-recurrent neural network, the block data array for each run of the multi-recurrent neural network may consume approximately several hundred megabytes per block. In some implementations of the system, the block data may also include an array of DFC data stored once per block, or other types of metadata about the sensor data and the blocks.

[0193] In operation, when a multi-cycle cluster becomes available, the wrapper allocates a patch to the cluster. The wrapper retrieves the next patch of block data during block traversal and sends it, along with appropriate control and configuration information, to the allocated cluster. The cluster can be configured to have sufficient memory on a configurable processor to hold data patches, including patches from multiple cycles in several systems that are being processed in-situ, as well as data patches to be processed when processing of the current patch is completed using ping-pong buffering or raster scanning techniques in various implementations.

[0194] When an assigned cluster completes its neural network operation on the current patch and produces an output patch, it signals the packager. The packager reads the output patch from the assigned cluster, or alternatively, the assigned cluster pushes data to the packager. The packager then assembles the output patch onto the processed block in DRAM 1102. When the processing of the entire block is complete and the output patch of data has been transferred to DRAM, the packager sends the processed output array of the block back to the host / CPU in a specified format. In some implementations, the onboard DRAM 1102 is managed by memory management logic in the packager 1100. The runtime program can control the sequencing operation to perform analysis of all arrays of block data in all loops during operation in a continuous streaming manner, thereby providing real-time analysis.

[0195] Computer System

[0196] Figure 12A computer system 1200 is available for use by a sequencing system 800A to implement the base interpretation technology disclosed herein. The computer system 1200 includes at least one central processing unit (CPU) 1272 that communicates with a plurality of peripheral devices via a bus subsystem 1255. These peripheral devices may include a storage subsystem 1210, which includes, for example, memory devices and a file storage subsystem 1236, a user interface input device 1238, a user interface output device 1276, and a network interface subsystem 1274. The input and output devices allow users to interact with the computer system 1200. The network interface subsystem 1274 provides an interface to an external network, including interfaces to corresponding interface devices in other computer systems.

[0197] In one implementation, the system controller 806 may be communicatively linked to the storage subsystem 1210 and the user interface input device 1238.

[0198] User interface input device 1238 may include: a keyboard; pointing devices such as a mouse, trackball, touchpad, or graphics tablet; a scanner; a touchscreen integrated into a display; audio input devices such as a voice recognition system and a microphone; and other types of input devices. Generally, the term "input device" is intended to include all possible types of devices and methods for inputting information into computer system 1200.

[0199] User interface output device 1276 may include a display subsystem, a printer, a fax machine, or a non-visual display such as an audio output device. The display subsystem may include an LED display, a cathode ray tube (CRT), a flat panel device such as a liquid crystal display (LCD), a projection device, or some other mechanism for producing visible images. The display subsystem may also provide a non-visual display, such as an audio output device. Generally, the term "output device" is intended to encompass all possible types of devices and methods for outputting information from computer system 1200 to a user or to another machine or computer system.

[0200] The storage subsystem 1210 stores the programming and data construction of some or all of the functions and methods of the modules described herein. These software modules are typically executed by the deep learning processor 1278.

[0201] The deep learning processor 1278 can be a graphics processing unit (GPU), a field-programmable gate array (FPGA), an application-specific integrated circuit (ASIC), and / or a coarse-grained reconfigurable architecture (CGRA). The deep learning processor 1278 can be powered by deep learning cloud platforms such as Google Cloud Platform. ™ Xilinx ™ and Cirrascale™ Hosted. Examples of the Deep Learning Processor 1278 include Google's Tensor Processing Unit (TPU). ™ Rackmount solutions (such as GX4 Rackmount Series) ™ GX12 Rackmount Series ™ NVIDIA DGX-1 ™ Microsoft's Stratix V FPGA ™ , Graphcore’s Intelligent Processor Unit (IPU) ™ Qualcomm's Snapdragon processors ™ Zeroth Platform ™ NVIDIA's Volta ™ NVIDIA's Drive PX ™ NVIDIA's Jetson TX1 / TX2 module ™ Intel's Nirvana ™ Movidius VPU ™ ,Fujitsu DPI ™ ARM's DynamicIQ ™ IBM TrueNorth ™ It has Testa V100s ™ Lambda GPU servers, etc.

[0202] The memory subsystem 1222 used in the storage subsystem 1210 may include multiple memories, including a main random access memory (RAM) 1232 for storing instructions and data during program execution and a read-only memory (ROM) 1234 for storing fixed instructions. The file storage subsystem 1236 may provide persistent storage for program files and data files and may include hard disk drives, floppy disk drives, and associated removable media, CD-ROM drives, optical disk drives, or removable media magnetic tape disks. Modules implementing certain specific implementation functions may be stored by the file storage subsystem 1236 in the storage subsystem 1210 or in other machines accessible to the processor.

[0203] The bus subsystem 1255 provides mechanisms for enabling various components and subsystems of the computer system 1200 to communicate with each other as intended. Although the bus subsystem 1255 is schematically shown as a single bus, alternative implementations of the bus subsystem may use multiple buses.

[0204] The computer system 1200 itself can be of different types, including personal computers, portable computers, workstations, computer terminals, network computers, televisions, mainframes, server clusters, a loosely networked group of widely distributed computers, or any other data processing system or user equipment. Due to the constantly evolving nature of computers and networks, [the following applies]. Figure 12 The description of the computer system 1200 depicted herein is intended only as a specific example to illustrate a preferred embodiment of the invention. Many other configurations of the computer system 1200 are possible, which have... Figure 12 The computer system depicted in the text has more or fewer components.

[0205] Terms and Conditions

[0206] This invention discloses the following provisions:

[0207] 1. An artificial intelligence-based system for base identification, the system comprising:

[0208] A neural network-based base reader that processes at least a right-wing input, a center input, and a left-wing input, and produces at least a right-wing output, a center output, and a left-wing output;

[0209] The right-wing input includes current image data for the current sequencing cycle used in the sequencing run, supplemented with previous image data for one or more previous sequencing cycles prior to the current sequencing cycle, and the right-wing output includes right-wing base readout predictions for the current sequencing cycle and base readout predictions for previous sequencing cycles.

[0210] The center input includes current image data, supplemented by previous image data and subsequent image data for one or more subsequent sequencing cycles after the current sequencing cycle; and the center output includes center base interpretation predictions for the current sequencing cycle and base interpretation predictions for previous and subsequent sequencing cycles.

[0211] The left-wing input includes the current image data, supplemented by subsequent image data, and the left-wing output includes left-wing base interpretation predictions for the current sequencing cycle and base interpretation predictions for subsequent sequencing cycles; and

[0212] A base readout generator coupled to a neural network-based base readout generator is configured to generate base readouts for the current sequencing cycle based on right-wing base readout predictions, center base readout predictions, and left-wing base readout predictions for the current sequencing cycle.

[0213] 2. The artificial intelligence-based system according to Clause 1, wherein current image data for the current sequencing cycle depicts the intensity emission of the analyte captured in the current sequencing cycle and its surrounding background.

[0214] 3. The artificial intelligence-based system according to Clause 2, wherein the right-wing base readout prediction, the center base readout prediction, and the left-wing base readout prediction for the current sequencing cycle identify the probability that one or more bases in the analyte incorporated into the analyte are A, C, T, and G at the current sequencing cycle.

[0215] 4. The artificial intelligence-based system according to Clause 3, wherein the base-reading generator is further configured to include an averager, the averager...

[0216] The probabilities of right-wing base readout prediction, center base readout prediction, and left-wing base readout prediction across the current sequencing cycle are summed base by base;

[0217] The average value for each base is determined by summing up each base; and

[0218] The base interpretation for the current sequencing cycle is generated based on the highest average value among the base-by-base average values.

[0219] 5. The artificial intelligence-based system according to Clause 3, wherein the base-reading generator is further configured to include a co-sensor, the co-sensor...

[0220] Based on the highest probability among the possibilities, preliminary base reads are determined for each of the right-wing, center-wing, and left-wing base read predictions for the current sequencing cycle, thereby generating preliminary base read sequences; and

[0221] Based on the most common bases in the initial bases readout sequence, bases are readouts for the current sequencing cycle are generated.

[0222] 6. The artificial intelligence-based system according to Clause 3, wherein the base-reading generator is further configured to include a weighted co-sensor, the weighted co-sensor...

[0223] Based on the highest probability among the probabilities, a preliminary base read is determined for each of the right wing base read predictions, the center base read predictions, and the left wing base read predictions for the current sequencing cycle, thereby generating a preliminary base read sequence.

[0224] The base-by-base weights are applied to the corresponding preliminary base reads in the preliminary base read sequence, generating a weighted preliminary base read sequence; and

[0225] A base read for the current sequencing cycle is generated based on the most weighted base read in the weighted preliminary base read sequence.

[0226] 7. The AI-based system as described in Clause 3, wherein the probability is an exponentially normalized score generated by a softmax layer.

[0227] 8. The AI-based system as described in Clause 1, further configured to include a trainer that, during training,

[0228] The errors between the base interpretations generated by the base interpretation generator for the current sequencing cycle, previous sequencing cycles, and subsequent sequencing cycles, based on the right-wing output, center output, and left-wing output of the neural network-based base interpreter, and the baseline true base interpretations are calculated.

[0229] Gradients for the current sequencing cycle, previous sequencing cycles, and subsequent sequencing cycles are determined based on error; and

[0230] The parameters of the neural network-based base reader are updated by backpropagating gradients.

[0231] 9. The artificial intelligence-based system according to Clause 1, wherein the right-wing base interpretation prediction for the current sequencing cycle takes into account a predetermined phase effect between the current sequencing cycle and previous sequencing cycles.

[0232] 10. The artificial intelligence-based system described in Clause 9, wherein the prediction of the central base readout for the current sequencing cycle takes into account the predetermined phase effect between the current sequencing cycle and the previous sequencing cycle, as well as the phase fixation effect between the current sequencing cycle and the subsequent sequencing cycle.

[0233] 11. The artificial intelligence-based system according to Clause 10, wherein the left-wing base interpretation prediction for the current sequencing cycle takes into account the phase-fixing effect between the current sequencing cycle and subsequent sequencing cycles.

[0234] 12. An artificial intelligence-based system for base identification, the system comprising:

[0235] Host processor;

[0236] A memory, accessible to the host processor, stores image data of sequencing cycles used in sequencing runs, wherein the current image data of the current sequencing cycle for the sequencing run depicts the intensity emission of the analyte captured in the current sequencing cycle and its surrounding background; and

[0237] A configurable processor that can access memory, the configurable processor comprising:

[0238] Multiple execution clusters, wherein the execution clusters are configured as execution neural networks; and

[0239] A data flow logic, which can access memory and execution clusters in multiple execution clusters, is configured to provide current image data, previous image data for one or more previous sequencing cycles preceding the current sequencing cycle, and subsequent image data for one or more subsequent sequencing cycles following the current sequencing cycle to available execution clusters in the multiple execution clusters. This causes the execution clusters to apply different groups of the current image data, previous image data, and subsequent image data to a neural network to generate a first base readout prediction, a second base readout prediction, and a third base readout prediction for the current sequencing cycle. The first base readout prediction, the second base readout prediction, and the third base readout prediction for the current sequencing cycle are then fed back to memory for use in generating base readouts for the current sequencing cycle based on the first base readout prediction, the second base readout prediction, and the third base readout prediction.

[0240] 13. The artificial intelligence-based system as described in Clause 12, wherein the different groups include a first group, a second group, and a third group, the first group including current image data and previous image data; the second group including current image data, previous image data, and subsequent image data; and the third group including current image data and subsequent image data.

[0241] 14. The artificial intelligence-based system according to Clause 13, wherein the execution cluster applies a first group to a neural network to produce a first base readout prediction, applies a second group to a neural network to produce a second base readout prediction, and applies a third group to a neural network to produce a third base readout prediction.

[0242] 15. The artificial intelligence-based system according to Clause 12, wherein the first base readout prediction, the second base readout prediction, and the third base readout prediction for the current sequencing cycle identify the probability that one or more bases in the analyte incorporated into the analyte are A, C, T, and G at the current sequencing cycle.

[0243] 16. The AI-based system as described in Clause 15, wherein the data stream logic is further configured to generate base interpretations for the current sequencing cycle.

[0244] The probability of the first, second, and third base readouts across the current sequencing cycle is summed base by base;

[0245] The average value for each base is determined by summing up each base; and

[0246] The base interpretation for the current sequencing cycle is generated based on the highest average value among the base-by-base average values.

[0247] 17. The AI-based system as described in Clause 15, wherein the data stream logic is further configured to generate base interpretations for the current sequencing cycle.

[0248] A preliminary base-read sequence is generated by determining the preliminary base reads for each of the first, second, and third base read predictions for the current sequencing cycle based on the highest probability among the possible base reads; and

[0249] Based on the most common bases in the initial bases readout sequence, bases are readouts for the current sequencing cycle are generated.

[0250] 18. The AI-based system as described in Clause 15, wherein the data stream logic is further configured to generate base interpretations for the current sequencing cycle.

[0251] A preliminary base read sequence is generated by determining a preliminary base read for each of the first, second, and third base read predictions for the current sequencing cycle based on the highest probability among the probabilities.

[0252] The base-by-base weights are applied to the corresponding preliminary base reads in the preliminary base read sequence, generating a weighted preliminary base read sequence; and

[0253] A base read for the current sequencing cycle is generated based on the most weighted base read in the weighted preliminary base read sequence.

[0254] 19. An artificial intelligence-based method for base identification, the method comprising:

[0255] Access the current image data for the current sequencing cycle used in the sequencing run, the previous image data for one or more previous sequencing cycles before the current sequencing cycle, and the subsequent image data for one or more subsequent sequencing cycles after the current sequencing cycle.

[0256] A neural network-based base reader processes different groups of current, previous, and subsequent image data, generating first, second, and third base read predictions for the current sequencing cycle; and

[0257] Base interpretations for the current sequencing cycle are generated based on the first base interpretation prediction, the second base interpretation prediction, and the third base interpretation prediction.

[0258] 20. The artificial intelligence-based method as described in Clause 19, wherein the different groupings include:

[0259] The first group includes current image data and previous image data;

[0260] The second group includes current image data, previous image data, and subsequent image data; and

[0261] The third group includes current image data and subsequent image data.

[0262] 21. The artificial intelligence-based method according to Clause 20, further comprising:

[0263] The first group is processed by a neural network-based base reader to generate a first base reading prediction;

[0264] The second group is processed by a neural network-based base reader to generate a second base reading prediction; and

[0265] The third group is processed by a neural network-based base reader to generate a third base reading prediction.

[0266] 22. The artificial intelligence-based method according to Clause 19, wherein the first base readout prediction, the second base readout prediction, and the third base readout prediction for the current sequencing cycle identify the probability that one or more bases in the analyte incorporated into the analyte are A, C, T, and G at the current sequencing cycle.

[0267] 23. The artificial intelligence-based method as described in Clause 22, further comprising generating base interpretations for the current sequencing cycle,

[0268] The probability of the first, second, and third base readouts across the current sequencing cycle is summed base by base;

[0269] The average value for each base is determined by summing up each base; and

[0270] The base interpretation for the current sequencing cycle is generated based on the highest average value among the base-by-base average values.

[0271] 24. The artificial intelligence-based method as described in Clause 22, further comprising generating base interpretations for the current sequencing cycle,

[0272] A preliminary base-read sequence is generated by determining the preliminary base reads for each of the first, second, and third base read predictions for the current sequencing cycle based on the highest probability among the possible base reads; and

[0273] Based on the most common bases in the initial bases readout sequence, bases are readouts for the current sequencing cycle are generated.

[0274] 25. The artificial intelligence-based method as described in Clause 22, further comprising generating base interpretations for the current sequencing cycle.

[0275] A preliminary base read sequence is generated by determining a preliminary base read for each of the first, second, and third base read predictions for the current sequencing cycle based on the highest probability among the probabilities.

[0276] The base-by-base weights are applied to the corresponding preliminary base reads in the preliminary base read sequence, generating a weighted preliminary base read sequence; and

[0277] A base read for the current sequencing cycle is generated based on the most weighted base read in the weighted preliminary base read sequence.

[0278] 26. An artificial intelligence-based method for base identification, the method comprising:

[0279] The base interpreter based on a neural network processes at least the right wing input, the center input, and the left wing input, and produces at least the right wing output, the center output, and the left wing output.

[0280] The right-wing input includes current image data for the current sequencing cycle used in the sequencing run, supplemented with previous image data for one or more previous sequencing cycles prior to the current sequencing cycle, and the right-wing output includes right-wing base readout predictions for the current sequencing cycle and base readout predictions for previous sequencing cycles.

[0281] The center input includes current image data, supplemented by previous image data and subsequent image data for one or more subsequent sequencing cycles after the current sequencing cycle; and the center output includes center base interpretation predictions for the current sequencing cycle and base interpretation predictions for previous and subsequent sequencing cycles.

[0282] The left-wing input includes current image data, supplemented by subsequent image data, and the left-wing output includes left-wing base interpretation predictions for the current sequencing cycle and base interpretation predictions for subsequent sequencing cycles; and

[0283] Base interpretations for the current sequencing cycle are generated based on right-wing base interpretation predictions, center base interpretation predictions, and left-wing base interpretation predictions for the current sequencing cycle.

[0284] 27. The artificial intelligence-based method according to Clause 26, wherein current image data for the current sequencing cycle depicts the intensity emission of the analyte captured in the current sequencing cycle and its surrounding background.

[0285] 28. The artificial intelligence-based method according to Clause 26, wherein the right-wing base readout prediction, the center base readout prediction, and the left-wing base readout prediction for the current sequencing cycle identify the probability that one or more bases in the analyte incorporated into the analyte are A, C, T, and G at the current sequencing cycle.

[0286] 29. The artificial intelligence-based method according to Clause 28, further comprising generating base interpretations for the current sequencing cycle.

[0287] The probability of right wing base readout prediction, center base readout prediction, and left wing base readout prediction across the current sequencing cycle is summed base by base;

[0288] The average value for each base is determined by summing up each base; and

[0289] The base interpretation for the current sequencing cycle is generated based on the highest average value among the base-by-base average values.

[0290] 30. The artificial intelligence-based method as described in Clause 28, further comprising generating base interpretations for the current sequencing cycle.

[0291] A preliminary base-read sequence is generated by determining the preliminary base reads for each of the right-wing, center-wing, and left-wing base read predictions for the current sequencing cycle based on the highest probability among the possible predictions; and

[0292] Based on the most common bases in the initial bases readout sequence, bases are readouts for the current sequencing cycle are generated.

[0293] 31. The artificial intelligence-based method as described in Clause 28, further comprising generating base interpretations for the current sequencing cycle.

[0294] A preliminary base read sequence is generated by determining a preliminary base read for each of the right wing, center, and left wing base read predictions for the current sequencing cycle based on the highest probability among the probabilities.

[0295] The base-by-base weights are applied to the corresponding preliminary base reads in the preliminary base read sequence, generating a weighted preliminary base read sequence; and

[0296] A base read for the current sequencing cycle is generated based on the most weighted base read in the weighted preliminary base read sequence.

[0297] 32. The artificial intelligence-based method as described in Clause 28, wherein the probability is an exponentially normalized score generated by a softmax layer.

[0298] 33. The artificial intelligence-based method according to Clause 26, further comprising: during training,

[0299] The errors between the base interpretations generated by the base interpretation generator for the current sequencing cycle, previous sequencing cycles, and subsequent sequencing cycles, based on the right-wing output, center output, and left-wing output of the neural network-based base interpreter, and the baseline true base interpretations are calculated.

[0300] Gradients for the current sequencing cycle, previous sequencing cycles, and subsequent sequencing cycles are determined based on error; and

[0301] The parameters of the neural network-based base reader are updated by backpropagating gradients.

[0302] 34. The artificial intelligence-based method according to Clause 26, wherein the right-wing base interpretation prediction for the current sequencing cycle takes into account a predetermined phase effect between the current sequencing cycle and previous sequencing cycles.

[0303] 35. The artificial intelligence-based method according to Clause 34, wherein the prediction of the central base readout for the current sequencing cycle takes into account the predetermined phase effect between the current sequencing cycle and the previous sequencing cycle, as well as the phase fixation effect between the current sequencing cycle and the subsequent sequencing cycle.

[0304] 36. The artificial intelligence-based method according to Clause 35, wherein the left-wing base interpretation prediction for the current sequencing cycle takes into account the phase-fixing effect between the current sequencing cycle and subsequent sequencing cycles.

[0305] 37. An artificial intelligence-based method for base identification, the method comprising:

[0306] At least the first input, the second input, and the third input are processed by a neural network-based base reader, and at least the first output, the second output, and the third output are generated.

[0307] The first input includes specific image data for a specific sequencing cycle used in the sequencing run, supplemented by previous image data for one or more previous sequencing cycles preceding the specific sequencing cycle, and the first output includes a first base readout prediction for the specific sequencing cycle and a base readout prediction for the previous sequencing cycles.

[0308] The second input includes specific image data, supplemented by previous image data and subsequent image data for one or more subsequent sequencing cycles following a specific sequencing cycle; and the second output includes a second base readout prediction for a specific sequencing cycle and a base readout prediction for both the previous and subsequent sequencing cycles.

[0309] The third input includes specific image data, supplemented by subsequent image data, and the third output includes a third base interpretation prediction for a specific sequencing cycle and a base interpretation prediction for subsequent sequencing cycles; and

[0310] Base interpretations for a specific sequencing cycle are generated based on the first, second, and third base interpretation predictions for that cycle.

[0311] 38. The artificial intelligence-based method described in Clause 37, the implementation of which ultimately depends on each of the provisions of Clause 1.

[0312] 39. A non-transitory computer-readable storage medium having computer program instructions applied to it to perform artificial intelligence-based base determination, the instructions, when executed on a processor, implementing a method comprising:

[0313] Access the current image data for the current sequencing cycle used in the sequencing run, the previous image data for one or more previous sequencing cycles before the current sequencing cycle, and the subsequent image data for one or more subsequent sequencing cycles after the current sequencing cycle.

[0314] A neural network-based base reader processes different groups of current, previous, and subsequent image data, generating first, second, and third base read predictions for the current sequencing cycle; and

[0315] Base interpretations for the current sequencing cycle are generated based on the first base interpretation prediction, the second base interpretation prediction, and the third base interpretation prediction.

[0316] 40. The nontransitory computer-readable storage medium as described in Clause 39, the implementation of which ultimately depends on each of the clauses in Clause 1.

[0317] 41. A non-transitory computer-readable storage medium having computer program instructions applied to it to perform artificial intelligence-based base determination, the instructions, when executed on a processor, implementing a method comprising:

[0318] At least the first input, the second input, and the left input are processed by a neural network-based base reader, and at least the first output, the second output, and the left output are generated.

[0319] The first input includes specific image data for a specific sequencing cycle used in the sequencing run, supplemented by previous image data for one or more previous sequencing cycles preceding the specific sequencing cycle, and the first output includes a first base readout prediction for the specific sequencing cycle and a base readout prediction for the previous sequencing cycles.

[0320] The second input includes specific image data, supplemented by previous image data and subsequent image data for one or more subsequent sequencing cycles following a specific sequencing cycle; and the second output includes a second base readout prediction for a specific sequencing cycle and a base readout prediction for both the previous and subsequent sequencing cycles.

[0321] The left-hand input includes specific image data, supplemented by subsequent image data, and the left-hand output includes left-hand base interpretation predictions for a specific sequencing cycle and base interpretation predictions for subsequent sequencing cycles; and

[0322] Base interpretations for a specific sequencing cycle are generated based on the first base interpretation prediction, the second base interpretation prediction, and the left base interpretation prediction for that specific sequencing cycle.

[0323] 44. The non-transitory computer-readable storage medium as described in Clause 43, the implementation of which ultimately depends on each of the clauses in Clause 1.

[0324] 45. An artificial intelligence-based method for base identification, the method comprising:

[0325] Access the progress of the per-cycle analyte channel set generated for the sequencing cycle of the sequencing run;

[0326] By using a neural network-based base reader, the window for each sequencing cycle during sequencing operation is processed, along with the window for each analyte channel in the progress cycle, thus enabling...

[0327] Neural Network-Based Base Reader

[0328] The object window for the sequencing cycle in the sequencing run, the object window for processing the analyte channel set per cycle in progress; and

[0329] Generate provisional base read predictions for three or more sequencing cycles within the object window of the sequencing cycle;

[0330] From multiple windows appearing at different positions in a specific sequencing cycle, a neural network-based base reader is used to generate temporary base read predictions for that specific sequencing cycle; and

[0331] Base interpretation is determined for a specific sequencing cycle based on multiple base interpretation predictions.

[0332] 46. ​​The artificial intelligence-based method described in Clause 45, the implementation of which ultimately depends on each of the provisions of Clause 1.

[0333] 47. A system comprising one or more processors coupled to a memory loaded with computer instructions to perform artificial intelligence-based base reading, the instructions, when executed on the processor, performing the following actions:

[0334] Access the progress of the per-cycle analyte channel set generated for the sequencing cycle of the sequencing run;

[0335] By using a neural network-based base reader, the window for each sequencing cycle during sequencing operation is processed, along with the window for each analyte channel in the progress cycle, thus enabling...

[0336] Neural Network-Based Base Reader

[0337] The object window for the sequencing cycle in the sequencing run, the object window for processing the analyte channel set per cycle in progress; and

[0338] Generate provisional base read predictions for three or more sequencing cycles within the object window of the sequencing cycle;

[0339] From multiple windows appearing at different positions in a specific sequencing cycle, a neural network-based base reader is used to generate temporary base read predictions for that specific sequencing cycle; and

[0340] Base interpretation is determined for a specific sequencing cycle based on multiple base interpretation predictions.

[0341] 48. The system described in Clause 47 is ultimately dependent on each of the provisions of Clause 1.

[0342] 49. A non-transitory computer-readable storage medium having computer program instructions applied to it to perform artificial intelligence-based base determination, the instructions, when executed on a processor, implementing a method comprising:

[0343] Access the progress of the per-cycle analyte channel set generated for the sequencing cycle of the sequencing run;

[0344] By using a neural network-based base reader, the window for each sequencing cycle during sequencing operation is processed, along with the window for each analyte channel in the progress cycle, thus enabling...

[0345] Neural Network-Based Base Reader

[0346] The object window for the sequencing cycle in the sequencing run, the object window for processing the analyte channel set per cycle in progress; and

[0347] Generate provisional base read predictions for three or more sequencing cycles within the object window of the sequencing cycle;

[0348] From multiple windows appearing at different positions in a specific sequencing cycle, a neural network-based base reader is used to generate temporary base read predictions for that specific sequencing cycle; and

[0349] Base interpretation is determined for a specific sequencing cycle based on multiple base interpretation predictions.

[0350] 50. The non-transitory computer-readable storage medium as described in Clause 49, the implementation of which ultimately depends on each of the clauses in Clause 1.

[0351] 51. An artificial intelligence-based method for base identification, the method comprising:

[0352] Access a set of analyte channels generated per cycle for the sequencing run;

[0353] By using a neural network-based base reader, the window for each analyte channel in the sequencing cycle is processed within the sequence, enabling...

[0354] Neural Network-Based Base Reader

[0355] The object window for the sequencing cycle in the sequencing run, and the object window for processing the analyte channel set for each cycle in the series.

[0356] And generate base reading predictions for two or more sequencing cycles in the object window of the sequencing cycle;

[0357] Using a neural network-based base reader,

[0358] For multiple windows of the sequencing cycle in a sequencing run, multiple windows of the analyte channel set in each cycle of the series are processed.

[0359] And generate output for each window in the multiple windows.

[0360] Each of the multiple windows includes a specific set of per-cycle analyte channels for a specific sequencing cycle used in the sequencing run, and

[0361] The output for each of the multiple windows includes:

[0362] (i) Base reading prediction for specific sequencing cycles and

[0363] (ii) One or more additional base read predictions for one or more additional sequencing cycles of a sequencing run, thereby generating multiple base read predictions for a specific sequencing cycle across multiple windows; and

[0364] Base interpretation is determined for a specific sequencing cycle based on multiple base interpretation predictions.

[0365] 52. A system comprising one or more processors coupled to a memory loaded with computer instructions to perform artificial intelligence-based base reading, the instructions, when executed on the processor, performing the following actions:

[0366] Access a set of analyte channels generated per cycle for the sequencing run;

[0367] By using a neural network-based base reader, the window for each analyte channel in the sequencing cycle is processed within the sequence, enabling...

[0368] Neural Network-Based Base Reader

[0369] The object window for the sequencing cycle in the sequencing run, and the object window for processing the analyte channel set for each cycle in the series.

[0370] And generate base reading predictions for two or more sequencing cycles in the object window of the sequencing cycle;

[0371] Using a neural network-based base reader,

[0372] For multiple windows of the sequencing cycle in a sequencing run, multiple windows of the analyte channel set in each cycle of the series are processed.

[0373] And generate output for each window in the multiple windows.

[0374] Each of the multiple windows includes a specific set of per-cycle analyte channels for a specific sequencing cycle used in the sequencing run, and

[0375] The output for each of the multiple windows includes:

[0376] (i) Base reading prediction for specific sequencing cycles and

[0377] (ii) One or more additional base read predictions for one or more additional sequencing cycles of a sequencing run, thereby generating multiple base read predictions for a specific sequencing cycle across multiple windows; and

[0378] Base interpretation is determined for a specific sequencing cycle based on multiple base interpretation predictions.

[0379] 53. The system described in Clause 52 is ultimately dependent on each of the provisions of Clause 1.

[0380] 54. A non-transitory computer-readable storage medium having computer program instructions applied to it to perform artificial intelligence-based base determination, the instructions, when executed on a processor, implementing a method comprising:

[0381] Access a set of analyte channels generated per cycle for the sequencing run;

[0382] By using a neural network-based base reader, the window for each analyte channel in the sequencing cycle is processed within the sequence, enabling...

[0383] Neural Network-Based Base Reader

[0384] The object window for the sequencing cycle in the sequencing run, and the object window for processing the analyte channel set for each cycle in the series.

[0385] And generate base reading predictions for two or more sequencing cycles in the object window of the sequencing cycle;

[0386] Using a neural network-based base reader,

[0387] For multiple windows of the sequencing cycle in a sequencing run, multiple windows of the analyte channel set in each cycle of the series are processed.

[0388] And generate output for each window in the multiple windows.

[0389] Each of the multiple windows includes a specific set of per-cycle analyte channels for a specific sequencing cycle used in the sequencing run, and

[0390] The output for each of the multiple windows includes:

[0391] (i) Base reading prediction for specific sequencing cycles and

[0392] (ii) One or more additional base read predictions for one or more additional sequencing cycles of a sequencing run, thereby generating multiple base read predictions for a specific sequencing cycle across multiple windows; and

[0393] Base interpretation is determined for a specific sequencing cycle based on multiple base interpretation predictions.

[0394] 55. The non-transitory computer-readable storage medium as described in Clause 54, the implementation of which ultimately depends on each of the clauses in Clause 1.

[0395] Other embodiments of the above methods may include a non-transitory computer-readable storage medium storing instructions that can be executed by a processor to perform any of the methods described above. Yet another embodiment of the methods described in this section may include a system comprising a memory and one or more processors operable to execute instructions stored in the memory to perform any of the methods described above.

Claims

1. A system comprising: At least one processor; as well as A non-transitory computer-readable medium, comprising instructions that, when executed by the at least one processor, cause the system to: Access a series of per-cycle analyte datasets generated for the sequencing cycle of the sequencing run; The initial object window of the per-cycle analyte dataset, corresponding to the initial object window of the sequencing cycle for the sequencing run, is processed by a base reader and from a series of per-cycle analyte datasets. Using a base reader, temporary base read predictions are generated for a set of sequencing cycles in the initial object window of the sequencing cycle, where the target sequencing cycle appears at the target position; as well as Base reads for the target sequencing cycle are determined based on specific temporary base read predictions for the target sequencing cycle derived from temporary base read predictions.

2. The system of claim 1, further comprising instructions that, when executed by the at least one processor, cause the system to: Access a series of per-cycle analyte datasets for one or more image channels at the sequencing cycle during a sequencing run; and The initial object window for each cycle of analyte datasets for one or more image channels is processed using a base interpreter.

3. The system of claim 1, further comprising instructions that, when executed by the at least one processor, cause the system to: The subsequent object window of the sequencing cycle corresponding to the subsequent object window of the sequencing cycle that does not contain the initial object window of the sequencing cycle is processed from a series of per-cycle analyte datasets by a base interpreter. Using a base reader, subsequent provisional base read predictions are generated for the subsequent set of sequencing cycles within the subsequent object window of the sequencing cycle, where the subsequent target sequencing cycle appears at the target position; and Base interpretations for subsequent target sequencing cycles are determined based on specific provisional base interpretation predictions for subsequent provisional base interpretation predictions.

4. The system of claim 1, further comprising instructions that, when executed by the at least one processor, cause the system to: The previous object window of the sequencing cycle corresponding to the previous object window of the sequencing cycle that does not contain the initial object window of the sequencing cycle is processed from a series of per-cycle analyte datasets by a base interpreter. Using a base reader, a previous provisional base read prediction is generated for the previous set of sequencing cycles in the previous object window of the sequencing cycle, where the previous target sequencing cycle appears at the target position. as well as Base reads for previous target sequencing cycles are determined based on specific provisional base read predictions for previous target sequencing cycles derived from previous provisional base read predictions.

5. The system according to claim 1, wherein: The target location includes the center position within the initial object window of the sequencing cycle; and The initial object window of the sequencing cycle includes flanking elements adjacent to the center position.

6. The system of claim 1, wherein the set of sequencing cycles in the initial object window of the sequencing cycle comprises three or more sequencing cycles.

7. The system of claim 1, further comprising instructions that, when executed by the at least one processor, cause the system to access a series of per-cycle analyte datasets by accessing per-cycle images of intensity emission from nucleotide sequencing clusters.

8. The system of claim 1, wherein the base reader comprises a neural network-based base reader.

9. The system of claim 8, further comprising instructions, when executed by the at least one processor, causing the system to generate provisional base read predictions for a sequencing cycle set within an initial object window of the sequencing cycle by: A spatial convolutional layer based on a neural network base reader is applied to a series of per-cycle analyte datasets to generate a series of spatial convolutional representations of the per-cycle analyte datasets; Temporal convolutional layers of a neural network-based base reader are applied to combine spatial convolutional representations of a series of per-cycle analyte datasets; as well as The output layer of a neural network-based base reader is applied to a combined convolutional representation of a series of per-cycle analyte datasets to generate the probability of each nucleotide base for the target sequencing cycle.

10. A non-transitory computer-readable storage medium storing instructions that, when executed by at least one processor, cause a computer system to: Access a series of per-cycle analyte datasets generated for the sequencing cycle of the sequencing run; The initial object window of the per-cycle analyte dataset, corresponding to the initial object window of the sequencing cycle for the sequencing run, is processed by a base reader and from a series of per-cycle analyte datasets. Using a base reader, temporary base read predictions are generated for a set of sequencing cycles in the initial object window of the sequencing cycle, where the target sequencing cycle appears at the target position; as well as Base reads for the target sequencing cycle are determined based on specific temporary base read predictions for the target sequencing cycle derived from temporary base read predictions.

Images