Indel pathogenicity determination

Abstract

Described herein are techniques for converting the context of an ANN or another type of computing system that is trainable through machine learning. In some embodiments, the techniques convert a first context of a computing system (such as an ANN) that is intended to provide pathogenicity of variants in the genome of a population to a second context of a computing system that is intended to provide pathogenicity of indels in the genome of a population.

Description

[Technical field]

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS This application claims the benefit of and priority to U.S. Provisional Patent Application No. 63 / 304,308, entitled "INDEL PATHOGENICITY DETERMINATION," filed January 28, 2022, which is incorporated herein by reference in its entirety.

[0002] The disclosed technology relates to artificial intelligence-based computers and digital data processing systems and corresponding data processing methods and products for mimicking intelligence (i.e., knowledge-based systems, inference systems, and knowledge acquisition systems), including systems for reasoning with uncertainty (e.g., fuzzy logic systems), adaptive systems, machine learning systems, and artificial neural networks. In particular, the disclosed technology relates to using techniques to transform the context of an artificial neural network (ANN) or another type of computing system that is trainable through machine learning.

[0003] Built-in The following are incorporated by reference for all purposes as if fully set forth herein: U.S. Provisional Patent Application No. 63 / 253,122, entitled “PROTEIN STRUCTURE-BASED PROTEIN LANGUAGE MODELS,” filed on October 6, 2021 (Attorney Docket No. ILLM1050-1 / IP-2164-PRV); U.S. Provisional Patent Application No. 63 / 281,579, entitled “PREDICTING VARIANT PATHOGENICITY FROM EVOLUTIONARY CONSERVATION USING THREE-DIMENSIONAL (3D) PROTEIN STRUCTURE VOXELS,” filed on November 19, 2021 (Attorney Docket No. ILLM1060-1 / IP-2270-PRV); U.S. Provisional Patent Application No. 63 / 281,592, entitled “COMBINED AND TRANSFER LEARNING OF A VARIANT PATHOGENICITY PREDICTOR USING GAPED AND NON-GAPED PROTEIN SAMPLES,” filed on November 19, 2021 (Attorney Docket No. ILLM1061-1 / IP-2271-PRV); U.S. patent application Ser. No. 62 / 573,144, entitled “TRAINING A DEEP PATHOGENICITY CLASSIFIER USING LARGE-SCALE BENIGN TRAINING DATA,” filed on October 16, 2017 (Attorney Docket No. ILLM 1000-1 / IP-1611-PRV); U.S. Patent Application No. 62 / 573,149, entitled “PATHOGENICITY CLASSIFIER BASED ON DEEP CONVOLUTIONAL NEURAL NETWORKS (CNNs),” filed on October 16, 2017 (Attorney Docket No. ILLM 1000-2 / IP-1612-PRV); U.S. patent application Ser. No. 62 / 573,153, entitled “DEEP SEMI-SUPERVISED LEARNING THAT GENERATES LARGE-SCALE PATHOGENIC TRAINING DATA,” filed on October 16, 2017 (Attorney Docket No. ILLM 1000-3 / IP-1613-PRV); U.S. Patent Application No. 62 / 582,898, entitled “PATHOGENICITY CLASSIFICATION OF GENOMIC DATA USING DEEP CONVOLUTIONAL NEURAL NETWORKS (CNNs),” filed on November 7, 2017 (Attorney Docket No. ILLM 1000-4 / IP-1618-PRV); U.S. Patent Application No. 16 / 160,903, entitled “DEEP LEARNING-BASED TECHNIQUES FOR TRAINING DEEP CONVOLUTIONAL NEURAL NETWORKS,” filed on October 15, 2018 (Attorney Docket No. ILLM1000-5 / IP-1611-US); U.S. Patent Application Serial No. 16 / 160,986, entitled "DEEP CONVOLUTIONAL NEURAL NETWORKS FOR VARIANT CLASSIFICATION," filed on October 15, 2018 (Attorney Docket No. ILLM 1000-6 / IP-1612-US); U.S. Patent Application Serial No. 16 / 160,968, entitled “SEMI-SUPERVISED LEARNING FOR TRAINING AN ENSEMBLE OF DEEP CONVOLUTIONAL NEURAL NETWORKS,” filed on October 15, 2018 (Attorney Docket No. ILLM 1000-7 / IP-1613-US); U.S. Patent Application No. 16 / 407,149, entitled “DEEP LEARNING-BASED TECHNIQUES FOR PRE-TRAINING DEEP CONVOLUTIONAL NEURAL NETWORKS,” filed May 8, 2019 (Attorney Docket No. ILLM 1010-1 / IP-1734-US); U.S. Patent Application No. 17 / 232,056, entitled “DEEP CONVOLUTIONAL NEURAL NETWORKS TO PREDICT VARIANT PATHOGENICITY USING THREE-DIMENSIONAL (3D) PROTEIN STRUCTURES,” filed on April 15, 2021 (Attorney Docket No. ILLM 1037-2 / IP-2051-US); U.S. Patent Application No. 63 / 175,495, entitled “MULTI-CHANNEL PROTEIN VOXELIZATION TO PREDICT VARIANT PATHOGENICITY USING DEEP CONVOLUTIONAL NEURAL NETWORKS,” filed on April 15, 2021 (Attorney Docket No. ILLM 1047-1 / IP-2142-PRV); Sundaram, L. et al. Predicting the clinical impact of human mutation with deep neural networks. Nat. Genet. 50, 1161-1170 (2018) (hereinafter referred to as "PrimateAI"), Jaganathan,K.et al.Predicting splicing from primary sequence with deep learning.Cell 176,535-548(2019), U.S. Patent Application No. 63 / 175,767, entitled “EFFICIENT VOXELIZATION FOR DEEP LEARNING,” filed on April 16, 2021 (Attorney Docket No. ILLM 1048-1 / IP-2143-PRV); and U.S. Patent Application No. 17 / 468,411, entitled “ARTIFICIAL INTELLIGENCE-BASED ANALYSIS OF PROTEIN THREE-DIMENSIONAL (3D) STRUCTURES,” filed on September 7, 2021 (Attorney Docket No. ILLM 1037-3 / IP-2051A-US). [Background technology]

[0004] The subject matter discussed in this section should not be assumed to be prior art merely as a result of its mention in this section. Similarly, it should not be assumed that the problems mentioned in this section, or associated with the subject matter provided as background, have been previously recognized in the prior art. The subject matter in this section merely represents different approaches, which as such may also correspond to embodiments of the claimed technology.

[0005] Neural Networks FIG. 1 illustrates one embodiment of an artificial neural network (ANN) with multiple layers. An ANN (or neural network as it is also described herein) is a system of interconnected artificial neurons (e.g., a1, a2, a3) that exchange messages with each other. The illustrated neural network has three inputs, two neurons in the hidden layer, and two neurons in the output layer. The hidden layer has an activation function f(●) and the output layer has an activation function g(●). The connections are weighted by numerical weights (e.g., w 11 , w 21 , w 12 , w 31 , w 22 , w 32 , v 11 , v 22 ), so that a properly trained network will respond correctly when fed an image to be recognized. The input layer processes the raw input, and the hidden layer processes the output from the input layer based on the weights of the connections between the input layer and the hidden layer. The output layer receives the output from the hidden layer and processes it based on the weights of the connections between the hidden layer and the output layer. The network contains multiple layers of feature detection neurons, each with a number of neurons that respond to different combinations of inputs from the previous layer. These layers are constructed such that the first layer detects a set of primitive patterns in the input image data, the second layer detects patterns of patterns, and the third layer detects patterns of those patterns. Summary of the Invention

[0006] Described herein are techniques for converting the context of an ANN or another type of computing system that is trainable through machine learning. In some embodiments, the techniques convert a first context of a computing system (such as an ANN) that is intended to provide pathogenicity of variants (e.g., missense variants) in the genome of a population to a second context of a computing system that is intended to provide pathogenicity of indels in the genome of a population.

[0007] In providing such techniques, the systems and methods described herein overcome several technical problems in obtaining scores from a computing system where the context of the computing system is changed, and the techniques described herein provide specific technical solutions to overcome at least the technical problems mentioned herein, as well as other technical problems not described herein but recognized by those skilled in the art.

[0008] In accordance with certain embodiments, disclosed herein is a computerized method for transforming the context of an ANN or another type of computing system, and a non-transitory computer-readable storage medium for performing the technical operations of the computerized method. The non-transitory computer-readable storage medium tangibly stores or tangibly encodes computer-readable instructions that, when executed by one or more devices (e.g., one or more personal computers or servers), cause at least one processor to perform the method for transforming the context of an ANN or another type of computing system. [Brief description of the drawings]

[0009] In the drawings, like reference characters generally refer to like parts throughout the different views. Also, the drawings are not necessarily to scale, emphasis instead being placed upon illustrating the principles of the disclosed technology. In the following description, various embodiments of the disclosed technology are described with reference to the following drawings, in which: [Figure 1] 1 shows an embodiment of a feedforward neural network having multiple layers, which is a type of ANN. [Diagram 2] 1 illustrates a method for transforming the context of an artificial neural network (ANN) or another type of computing system that is trainable through machine learning, according to some embodiments of the present disclosure. [Diagram 3] Each of Figures 3 and 4 illustrates a method for converting the context of an ANN or another type of computing system that is trainable through machine learning, according to some embodiments of the present disclosure. Specifically, each of Figures 3 and 4 illustrates converting a first context of a computing system that is intended to provide pathogenicity of variants (e.g., missense variants) in the genomes of a population to a second context of a system that is intended to provide pathogenicity of indels in the genomes of a population. [Figure 4] Each of Figures 3 and 4 illustrates a method for converting the context of an ANN or another type of computing system that is trainable through machine learning, according to some embodiments of the present disclosure. Specifically, each of Figures 3 and 4 illustrates converting a first context of a computing system that is intended to provide pathogenicity of variants (e.g., missense variants) in the genomes of a population to a second context of a system that is intended to provide pathogenicity of indels in the genomes of a population. [Diagram 5] 5A-5D depict methods, each of which may be part of the method depicted in FIG. 4, according to some embodiments of the present disclosure. [Figure 6] 5A-5D depict methods, each of which may be part of the method depicted in FIG. 4, according to some embodiments of the present disclosure. [Figure 7]5A-5D depict methods, each of which may be part of the method depicted in FIG. 4, according to some embodiments of the present disclosure. [Figure 8] 5 illustrates two operations that may be combined with the method illustrated in FIG. 3 or the method illustrated in FIG. 4 according to some embodiments of the present disclosure. [Figure 9] Specifically, a method for converting the context of an ANN according to some embodiments of the present disclosure is shown in FIG. 9. Also shown is converting a first context of an ANN for providing pathogenicity of variants in the genome of a population to a second context of an ANN for providing pathogenicity of indels in the genome of a population. [Figure 10] 1 illustrates a block diagram of an exemplary aspect of a computing system according to some embodiments of the present disclosure. [Figure 11] Illustrated is a plot in a two-dimensional graph showing the relationship between binned PrimateAI scores for variants and insertion variants versus natural depletion (i.e., more depletion indicates stronger selection (i.e., propensity for variants or insertions in the genome of a population)). In FIG. 11, the natural depletion value (or propensity value) is represented on the y-axis, and the bins of PrimateAI scores are represented on the x-axis. [Figure 12] Figure 12 illustrates a scatter plot in a two-dimensional graph showing the relationship between the binned PrimateAI score for variants, insertion variants, and deletion variants versus the proportion of observed variants (i.e., the propensity of variants or indels in the genome of a population). In Figure 12, the proportion of observed variants is represented on the y-axis. Also, the bins of PrimateAI score are represented on the x-axis. [Figure 13]Each scatter plot is illustrated in a two-dimensional graph, and each plot shows the relationship between the binned PrimateAI score for variants, insertion variants, and deletion variants versus the adjusted proportion of observed variants (i.e., the propensity of variants or indels in the genome of the population). In Figures 13 and 14, the adjusted proportion (or adjusted ratio) of observed variants is represented on the y-axis, and the bins of PrimateAI score are represented on the x-axis. Specifically, Figure 13 relates to variants that occur 3 base pairs in frame in the exome. Specifically, Figure 14 relates to variants that occur 6 base pairs in frame in the exome. [Figure 14] Each scatter plot is illustrated in a two-dimensional graph, and each plot shows the relationship between the binned PrimateAI score for variants, insertion variants, and deletion variants versus the adjusted proportion of observed variants (i.e., the propensity of variants or indels in the genome of the population). In Figures 13 and 14, the adjusted proportion (or adjusted ratio) of observed variants is represented on the y-axis, and the bins of PrimateAI score are represented on the x-axis. Specifically, Figure 13 relates to variants that occur 3 base pairs in frame in the exome. Specifically, Figure 14 relates to variants that occur 6 base pairs in frame in the exome. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] The following discussion is presented to enable those skilled in the art to make and use the disclosed technology and is provided in the context of a particular application and its requirements. Various modifications to the disclosed embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be applied to other embodiments and applications without departing from the spirit and scope of the disclosed technology. Thus, the disclosed technology is not intended to be limited to the embodiments shown, but is to be accorded the widest scope consistent with the principles and features disclosed herein.

[0011] Primate AI PrimateAI is a deep residual neural network for classifying the pathogenicity of missense mutations. In at least one version, PrimateAI is trained on a dataset of approximately 380,000 common variants from humans and six non-human primate species using a semi-supervised benign-versus-unlabeled training regime. In such a version, the input to the network is the amino acid sequence flanking the variant of interest and ortholog sequence alignments in other species without any additional artificial features, and the output is a pathogenicity score ranging from 0 (low pathogenicity) to 1 (higher pathogenicity). In such a version, to incorporate information about protein structure, PrimateAI can learn to predict secondary structure and solvent accessibility from amino acid sequence and include these as sub-networks in the full model. Also, in such a version, the total size of the network in which the protein structure is included is 36 layers of convolutions with approximately 400,000 trainable parameters.

[0012] gnomAD The Genome Aggregation Database (gnomAD) is a resource developed by an international coalition of researchers with the goal of aggregating and harmonizing both exome and genome sequencing data from a wide variety of large-scale sequencing projects and making summary data available to the broader scientific community. Multiple versions of gnomAD have been released.

[0013] Described herein are techniques for transforming the context of an artificial neural network or another type of computing system that is trainable through machine learning. An example of the techniques disclosed herein transforms a first context for a computing system (such as an ANN) to a second context for the computing system. Specifically, the first context for the computing system is the pathogenicity of variants (e.g., missense variants) in the genome of a population, and the second context for the computing system is the pathogenicity of indels in the genome of a population. In other words, some of the techniques disclosed herein provide an operation for transforming a computing system or an output of a computing system, which means first providing the pathogenicity of variants (e.g., missense variants) in the genome of a population to a computing system or an output of a computing system that provides the pathogenicity of indels in the genome of a population.

[0014] The actions of Figures 2-9 may be performed at least in part using and / or by one or more processors configured to receive or retrieve information, process the information, store the results, and transmit the results. Other embodiments may perform the actions in a different order and / or with different, fewer, or additional actions than those illustrated in Figures 2-9. In some embodiments, multiple actions may be combined. For convenience, the figures are described with reference to a system that performs the method. A system is not necessarily part of the method. The actions of Figures 2-9 may be performed in parallel or sequentially.

[0015] FIG. 2 illustrates a method 100 for transforming a first context for a computing system (eg, an ANN) to a second context for the computing system.

[0016] In one embodiment, the ANN is a multilayer perceptron (MLP). In another embodiment, the ANN is a feedforward neural network. In yet another embodiment, the ANN is a fully-connected neural network. In a further embodiment, the ANN is a fully convolution neural network. In yet a further embodiment, the ANN is a semantic segmentation neural network. In yet another further embodiment, the ANN is a generative adversarial network (GAN) (e.g., CycleGAN, StyleGAN, pixelRNN, text-2-image, DiscoGAN, IsGAN).

[0017] In one embodiment, the ANN is a convolutional neural network (CNN) having multiple convolutional layers. In another embodiment, the ANN is a recurrent neural network (RNN), such as a long short-term memory network (LSTM), a bi-directional LSTM (Bi-LSTM), or a gated recurrent unit (GRU). In yet another embodiment, the ANN includes both a CNN and an RNN.

[0018] In yet other embodiments, the ANN may use 1D convolution, 2D convolution, 3D convolution, 4D convolution, 5D convolution, dilated or expanded convolution, transposed convolution, depth separable convolution, point-wise convolution, 1×1 convolution, group convolution, flattened convolution, spatial and cross-channel convolution, shuffled grouped convolution, spatially separable convolution, and deconvolution. The ANN may 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. ANNs can use any parallel, efficient, and compaction schemes, such as TFRecords, compression encoding (e.g., PNG), sharding, parallel calls to map transforms, batching, prefetching, model parallelism, data parallelism, and synchronous / asynchronous stochastic gradient descent (SGD). ANNs can include upsampling layers, downsampling layers, recurrent connections, gates and gated memory units (such as LSTM or GRU), residual blocks, residual connections, highway connections, skip connections, peephole connections, activation functions (e.g., nonlinear transformation functions such as rectified linear unit (ReLU), leaky ReLU, exponential linear unit (ELU), sigmoid, and hyperbolic tangent function (tanh)), batch normalization layers, normalization layers, dropout, pooling layers (e.g., max or average pooling), global average pooling layers, and attention mechanisms (e.g., self-attention).

[0019] The ANN can be a rule-based model, a linear regression model, a logistic regression model, an Elastic Net model, a support vector machine (SVM), a random forest (RF), a decision tree, and a boosted decision tree (e.g., XGBoost), or some other tree-based logic (e.g., metric trees, kd trees, R trees, universal B trees, X trees, ball trees, locality sensitive hashing, and inverted index). The ANN can be an ensemble of multiple models in some implementations.

[0020] The ANN is trained using a backpropagation-based gradient update technique. Exemplary gradient descent techniques that may be used to train the ANN include stochastic gradient descent, batch gradient descent, and mini-batch gradient descent. Some examples of gradient descent optimization algorithms that may be used to train the ANN include Momentum, Nesterov accelerated gradient, Adagrad, Adadelta, RMSprop, Adam, AdaMax, Nadam, and AMSGrad.

[0021] The ANN is a Transformer, Vision Transformer(ViT), Bidirectional Transformer(BERT), Detection Transformer(DETR), Deformable DETR, UP-DETR, DeiT, Swin, GPT, iGPT, GPT-2, GPT-3, BERT, SpanBERT, RoBERTa, XLNet, ELECTRA, UniLM, BART, T5, ER NIE(THU), KnowBERT, DeiT-Ti, DeiT-S, DeiT-B, T2T-ViT-14, T2T-ViT-19, T2T-ViT-24, PVT-Small, PVT-Medium, PV T-Large, TNT-S, TNT-B, CPVT-S, CPVT-S-GAP, CPVT-B, Swin-T, Swin-S, Swin-B, Twins-SVT-S, Twins-SVT-B, Twin- SVT-L, Shuffle-T, Shuffle-S, Shuffle-B, XCiT-S12 / 16, CMT-S, CMT-B, VOLO-D1, VOLO-D2, VOLO-D3, VOLO-D4, MoCo v3、ACT、TSP、Max-DeepLab、VisTR、SETR、Hand-Transformer、HOT-Net、METRO、Image Transformer、Taming transformer、TransGAN、IPT、TTSR、STTN、Masked Transformer, CLIP, DALL-E, Cogview, UniT, ASH, TinyBert, FullyQT, ConvBert, FCOS, Faster R-CNN+FPN、DETR-DC5、TSP-FCOS、TSP-RCNN、ACT+MKDD(L=32)、ACT+MKDD(L=16)、SMCA、Efficient DETR, UP-DETR, UP-DETR, ViTB / 16-FRCNN, ViT-B / 16-FRCNN, PVT-Small+RetinaNet, Swin-T+Retin aNet, Swin-T+ATSS, PVT-Small+DETR, TNT-S+DETR, YOLOS-Ti, YOLOS-S, and YOLOS-B are all slightly different.

[0022] FIG. 3 illustrates a method 200 of converting a first context of a computing system context intended to provide pathogenicity of variants (e.g., missense variants) in the genome of a population to a second context that provides pathogenicity of indels in the genome of a population.

[0023] 4 illustrates a method 300 for converting a first context of a computing system context intended to provide pathogenicity of variants (e.g., missense variants) in the genome of a population to a second context that provides pathogenicity of indels in the genome of the population. However, as shown in FIG. 4, indels specifically include multiple insertions and multiple deletions in the genome of the population.

[0024] For purposes of this disclosure, multiple indels should generally be understood to include multiple insertions and / or multiple deletions. Also, for purposes of this disclosure, variant should be understood to be a generic term for variants or indel variants (i.e., indels). Also, for purposes of this disclosure, indel variants (i.e., indels) should be understood to be a generic term for insertion variants (i.e., insertions) or deletion variants (i.e., deletions). Unless otherwise specified herein, the term "variant" refers to a nucleic acid sequence that differs from a nucleic acid reference. Exemplary nucleic acid sequence variants include, but are not limited to, single nucleotide polymorphisms (SNPs), short deletion and insertion polymorphisms (indels), copy number variations (CNVs), microsatellite markers, or short tandem repeats and structural mutations. Somatic variant calling is an effort to identify variants present at low frequency in a DNA sample. Somatic variant calling is of interest in the context of cancer treatment. Cancer is caused by the accumulation of mutations in DNA. DNA samples from tumors are generally heterogeneous, containing some normal cells, early stages of cancer progression (with fewer mutations), and some late stages (with more mutations). Due to this heterogeneity, when sequencing tumors (e.g., from FFPE samples), somatic mutations often appear at low frequency. For example, SNVs may be found in only 10% of the reads that cover a given base. Variants that are classified as somatic or germline by variant classifiers are also referred to herein as "variants under test".

[0025] Method 100 begins at step 102 with processing a plurality of first variations of an object to generate a plurality of first scores for respective quantifiable attributes for the variations among the plurality of first variations of the object. Method 100 then continues to step 104 with generating a first context curve based on the plurality of first scores according to one or more curve forming functions.

[0026] As used herein, a "function" or "logic" (e.g., a curve forming function) may be realized in the form of a computer product including a non-transitory computer readable storage medium with computer usable program code to perform the method steps described herein. "Logic" may be realized in the form of an apparatus including a memory and at least one processor coupled to the memory and operative to perform the exemplary method steps. "Logic" may be realized in the form of a means for performing one or more of the method steps described herein. The means may include (i) hardware modules, (ii) software modules running on one or more hardware processors, or (iii) a combination of hardware and software modules, any of (i)-(iii) implementing certain techniques described herein, the software modules being stored in a computer readable storage medium (or multiple such media). In one embodiment, logic performs a data processing function. The logic may be a general-purpose, single-core or multi-core processor with a computer program specifying the function, a digital signal processor with a computer program, configurable logic such as an FPGA with a configuration file, special purpose circuitry such as a state machine, or any combination thereof. Also, the computer program product may embody computer program and configuration file portions of logic.

[0027] Method 100 also begins at step 106, which includes processing a plurality of second variations of the object to generate a plurality of second scores for respective quantifiable attributes for the variations in the plurality of second variations of the object. Method 100 then continues to step 108, which includes generating a second context curve based on the plurality of second scores according to one or more curve forming functions.

[0028] The method 100 then continues at step 110, which includes determining a selection pattern difference between the first context curve and the second context curve. The method 100 then continues at step 112, which includes determining one or more scaling functions to reduce the selection pattern difference between the first context curve and the second context curve. Finally, at step 114, the method 100 continues with augmenting / calibrating / recalibrating / updating / optimizing / modifying the plurality of second scores according to the scaling functions to provide increased accuracy of the respective quantifiable attribute for each second variation of the plurality of second variations of the object.

[0029] In some embodiments of method 100 (such as method 200 shown in FIG. 3), the plurality of first variations of the object are a plurality of variants in the genome of the population, and the plurality of second variations of the object are a plurality of indels in the genome. Also, in such embodiments of method 100, the plurality of first scores are a plurality of missense pathogenicity scores for each variant of the plurality of variants, and the plurality of second scores are a plurality of indel pathogenicity scores for each indel of the plurality of indels. Also, in such embodiments of method 100, the first context curve is a missense curve based on the plurality of missense pathogenicity scores, and the second context curve is an indel curve based on the plurality of indel pathogenicity scores.

[0030] Method 200 begins at step 202, which includes processing a plurality of variants to generate a plurality of missense pathogenicity scores for each variant of the plurality of variants. Method 200 then continues to step 204, which includes generating a missense curve based on the plurality of missense pathogenicity scores according to one or more curve-forming functions.

[0031] Method 200 also begins at step 206, which includes processing the plurality of indels to generate a plurality of indel pathogenicity scores for each indel of the plurality of indels. Method 200 then continues to step 208, which includes generating an indel curve based on the plurality of indel pathogenicity scores according to a curve-forming function.

[0032] The method 200 then continues at step 210, which includes determining a selection pattern difference between the indel curves and the missense curves. The method 200 then continues at step 212, which includes determining one or more scaling functions to reduce the selection pattern difference between the missense curves and the indel curves. Finally, at step 214, the method 200 continues with enhancing / calibrating / recalibrating / updating / optimizing / modifying the plurality of indel pathogenicity scores according to the scaling functions to provide a recalibrated accuracy of the indel pathogenicity score for each indel of the plurality of indels.

[0033] In some embodiments of the above-mentioned method, the curve-forming function comprises a function that describes the proportion of different indels and the proportion of different variants in the genome of the population.For example, in some embodiments, the curve-forming function comprises a function that describes the natural selection of different indels and the natural selection of different variants in the genome of the population.See FIG. 11 for an example of the results of such a function that describes the natural selection of variants.

[0034] In a more general embodiment, the curve-forming function includes a function that describes the proportion of a first variation of objects and the proportion of a second variation of objects in a population of objects.

[0035] In some embodiments of the methods described above (such as method 300 shown in FIG. 4), the multiple indels include multiple insertions and multiple deletions, and the multiple indel pathogenicity scores include multiple insertion scores and multiple deletion scores, respectively.

[0036] Method 300 begins at step 302, which includes processing a plurality of variants to generate a plurality of missense pathogenicity scores for each variant of the plurality of variants. Method 300 then continues to step 304, which includes generating a missense curve based on the plurality of missense pathogenicity scores according to one or more curve-forming functions.

[0037] Method 300 also begins at step 306a, which includes processing the multiple insertions to generate multiple insertion scores for each insertion of the multiple insertions. Method 300 also begins at step 306b, which includes processing the multiple deletions to generate multiple deletion scores for each deletion of the multiple deletions. Method 300 then continues at step 308a, which includes generating an insertion curve based on the multiple insertion scores according to a curve-fitting function. Method 300 also continues at step 308b, which includes generating a deletion curve based on the multiple deletion scores according to a curve-fitting function.

[0038] Next, the method 300 continues at step 310a, which includes determining a selection pattern difference between the insertion curve and the missense curve. Next, the method 300 continues at step 310b, which includes determining a selection pattern difference between the deletion curve and the missense curve. Next, the method 300 continues at step 312a, which includes determining one or more scaling functions to reduce the selection pattern difference between the missense curve and the insertion curve. Further, the method 300 continues at step 312b, which includes determining an additional one or more scaling functions to reduce the selection pattern difference between the missense curve and the deletion curve. Finally, at step 314, the method 300 continues to augment / calibrate / recalibrate / update / optimize / modify the plurality of insertion scores and the plurality of deletion scores according to the respective scaling functions to provide a recalibrated accuracy of the insertion pathogenicity score for each insertion of the plurality of insertions and each deletion of the plurality of deletions.

[0039] In some embodiments of the above-described methods (see, e.g., FIG. 4), the insertion curve includes a first plurality of data points including an insertion propensity score for each bin of the group of bins. Also in such embodiments, the deletion curve includes a second plurality of data points including a deletion propensity score for each bin of the group of bins. Also in such examples, the missense curve includes a third plurality of data points including a missense propensity score for each bin of the group of bins. See FIGS. 12-14 for examples of such data points displayed on a graph.

[0040] In a more general example (see, e.g., FIG. 3), the indel curve includes a plurality of data points including an indel propensity score for each bin of the group of bins. Also, in such a more general example, the missense curve includes a plurality of data points including a missense propensity score for each bin of the group of bins. In an even more general example (see, e.g., FIG. 1), the first context curve includes a plurality of data points including a first context propensity score for each bin of the group of bins. Also, in such an even more general example, the second context curve includes a plurality of data points including a second context propensity score for each bin of the group of bins.

[0041] In some embodiments of the above-mentioned method (see, e.g., FIG. 4), the insertion propensity score for a bin of the group of bins is related to the proportion of different insertions in the genome of a population having an insertion score of the plurality of insertion scores associated with the bin. In such an example, the deletion propensity score for a bin of the group of bins is related to the proportion of different deletions in the genome of a population having a deletion score of the plurality of deletion scores associated with the bin, and the missense propensity score for a bin of the group of bins is related to the proportion of variants in the genome of a population having a missense pathogenicity score of the plurality of missense pathogenicity scores associated with the bin. In a more general example (see, e.g., FIG. 3), the indel propensity score for a bin of the group of bins is related to the proportion of different indels in the genome of a population having an indel pathogenicity score of the plurality of indel pathogenicity scores associated with the bin. In such a more general example, the missense propensity score for a bin of the group of bins is related to the proportion of variants in the genome of a population having a missense pathogenicity score of the plurality of missense pathogenicity scores associated with the bin. In an even more general example (see, e.g., FIG. 1 ), a first context propensity score for a bin of the group of bins is related to the proportion of different first variations of objects in the population having a first context score of a plurality of first context scores associated with the bin. In such an even more general example, a second context propensity score for a bin of the group of bins is related to the proportion of different second variations of objects in the population having a second context score of a plurality of second context scores associated with the bin.

[0042] 5 illustrates a method 400 that, in some embodiments, is part of step 308a (including generating an insertion curve) of method 300. In such embodiments, generating an insertion curve in step 308a includes grouping a plurality of insertions into a group of bins in step 402 of method 400. Step 308a also includes step 404, which includes, for each bin in the group of bins, measuring a central tendency distribution of insertion scores for the bin. And step 308a also includes step 406, which includes, for each bin in the group of bins, applying the central tendency distribution of insertion scores for the bin to identify an insertion propensity score for the bin.

[0043] 6 illustrates a method 500 that, in some embodiments, is part of step 308b (including generating the deletion curve) of method 300. In such embodiments, generating the deletion curve in step 308b includes grouping the multiple deletions into groups of bins in step 502 of method 500. Step 308b also includes step 504, which includes, for each bin in the group of bins, measuring a central tendency distribution of the deletion scores in the bin. And step 308b also includes step 506, which includes, for each bin in the group of bins, applying a central tendency distribution of the insertion scores in the bin to identify an insertion propensity score for the bin.

[0044] 7 illustrates a method 600 that, in some embodiments, is part of step 304 (including generating the missense curve) of method 300. In such embodiments, generating the missense curve in step 304 includes grouping the plurality of variants into groups of bins in step 602 of method 600. Step 304 also includes step 604 that includes, for each bin in the group of bins, determining a central tendency distribution of the missense pathogenicity scores in the bin. And step 304 also includes step 606 that includes, for each bin in the group of bins, applying the central tendency distribution of the missense pathogenicity scores in the bin to identify a missense propensity score for the bin.

[0045] Techniques similar to those shown in Figures 5-7 may be applied to more general embodiments using multiple indels and multiple variants. Techniques similar to those shown in Figures 5-7 may also be applied to even more general embodiments using multiple first variations of objects of a population and multiple second variations of objects. For example, in some embodiments (such as those with respect to Figure 3), generating the indel curve includes grouping the multiple indels into a group of bins. Also, generating the missense curve includes grouping the multiple variants into a group of bins. Also, in some examples, generating the indel curve includes, for each bin of the group of bins, measuring a central tendency distribution of indel pathogenicity scores at the bin and applying the central tendency distribution of indel pathogenicity scores at the bins to identify an indel propensity score for the bin. Also, in such examples, generating the missense curve includes, for each bin of the group of bins, measuring a central tendency distribution of missense pathogenicity scores at the bin and applying the central tendency distribution of missense pathogenicity scores at the bins to identify a missense propensity score for the bin.

[0046] In some embodiments, measuring the central tendency distribution of the indel pathogenicity scores comprises determining the mean of the indel pathogenicity scores. For example, measuring the central tendency distribution of the insertion scores comprises determining the mean of the insertion scores, and measuring the central tendency distribution of the deletion scores comprises determining the mean of the deletion scores. Also, in some embodiments, measuring the central tendency distribution of the missense pathogenicity scores comprises determining the mean of the missense pathogenicity scores. In some embodiments, measuring the central tendency distribution of the indel pathogenicity scores comprises determining the mode of the indel pathogenicity scores. For example, measuring the central tendency distribution of the insertion scores comprises determining the mode of the insertion scores, and measuring the central tendency distribution of the deletion scores comprises determining the mode of the deletion scores. Also, in some embodiments, measuring the central tendency distribution of the missense pathogenicity scores comprises determining the mode of the missense pathogenicity scores. In some embodiments, measuring the central tendency distribution of the indel pathogenicity scores comprises determining the median of the indel pathogenicity scores. For example, measuring the central tendency distribution of the insertion scores includes determining the median of the insertion scores, and measuring the central tendency distribution of the deletion scores includes determining the median of the deletion scores. Also, in some embodiments, measuring the central tendency distribution of the missense pathogenicity scores includes determining the median of the missense pathogenicity scores. Also, such techniques apply to more general embodiments as well. For example, measuring the central tendency distribution of the first context scores includes determining the mean, mode, or median of the first context scores. Also, measuring the central tendency distribution of the second context scores includes determining the mean, mode, or median of the second context scores.

[0047] In some embodiments of the above-mentioned method (see, for example, FIG. 4), the insertion propensity score for a bin of a group of bins represents the probability that one of the multiple insertions associated with the bin occurs in the genome of the population given the set of observed insertions. In such embodiments, the deletion propensity score for a bin represents the probability that one of the multiple deletions associated with the bin occurs in the genome of the population given the set of observed deletions, and the missense propensity score for a bin represents the probability that one of the multiple variants associated with the bin occurs in the genome of the population given the set of observed variants. In some of the above embodiments, the propensity score reduces selection bias by equalizing the groups based on covariates, where the covariates are the set of observed insertions, the set of observed deletions, and the set of observed variants, respectively.

[0048] In a more general example (see, e.g., FIG. 3), the indel propensity score for a bin of a group of bins represents the probability that one of the indels associated with the bin will occur in the genome of the population given the set of observed indels. Also, in such more general embodiments, the missense propensity score for a bin represents the probability that one of the variants associated with the bin will occur in the genome of the population given the set of observed variants. Also, in some of the above embodiments, the propensity score reduces selection bias by equalizing groups based on covariates, where the covariates are the set of observed indels and the set of observed variants, respectively.

[0049] In an even more general example (see, e.g., FIG. 1 ), the first contextual propensity score for a bin of the group of bins represents a probability that one of a plurality of first variations of an object associated with the bin occurs in the population given a set of observed first variations. And in such more general implementations, the second contextual propensity score for a bin of the group of bins represents a probability that one of a plurality of second variations of an object associated with the bin occurs in the population given a set of observed second variations. And in some of the above implementations, the propensity scores reduce selection bias by equating groups based on covariates, the covariates being the set of observed first variations and the set of observed second variations, respectively.

[0050] In some embodiments of the above-mentioned method (see, e.g., FIG. 4), an insertion curve is generated when a first plurality of data points are plotted on a two-dimensional graph with one axis for the propensity score and the other axis for the group of bins. In such an example, a deletion curve is generated when a second plurality of data points are plotted on a two-dimensional graph, and a missense curve is generated when a third plurality of data points are plotted on a two-dimensional graph. In a more general example (see, e.g., FIG. 3), an indel curve is generated when a corresponding plurality of data points for an indel are plotted on a two-dimensional graph with one axis for the propensity score and the other axis for the group of bins. In such a more general example, a missense curve is generated when a corresponding plurality of data points for a variant are plotted on a two-dimensional graph. In an even more general example (see, e.g., FIG. 1), a first context curve is generated when a corresponding plurality of data points for a first variation of an object are plotted on a two-dimensional graph with one axis for the propensity score and the other axis for the group of bins. In such more general examples, the second context curve is generated when a corresponding plurality of data points for a second variation of the object are plotted on a two-dimensional graph. In some of the above embodiments, the two-dimensional graph includes a set of ordered pairs (x, y), where f(x)=y, where x is a group of bins and y is a propensity score. See Figures 12-14 for examples of such data points displayed on the graph.

[0051] In some embodiments of the above-mentioned method (see, e.g., FIG. 4), one or more scaling functions (for variants), one or more scaling functions (for insertions), and one or more scaling functions (for deletions) are part of the scaling function. Such scaling functions also include functions that scale the proportion of different insertions, different deletions, and different variants in the genome of the population, respectively, since indels and single nucleotide variants have different variability. In a more general example (see, e.g., FIG. 3), one or more scaling functions (for variants) and one or more scaling functions (for indels) are part of the scaling function. Such scaling functions also include functions that scale the proportion of different indels and different variants in the genome of the population, respectively. In an even more general example (see, e.g., FIG. 1), one or more scaling functions (for a first variation of an object) and one or more scaling functions (for a second variation of an object) are part of the scaling function. Such scaling functions also include functions that scale the proportion of different first variations of an object and different second variations of an object, respectively.

[0052] In some embodiments (see, e.g., FIG. 3 and FIG. 4), the scaling function obtains the scaling factor from a variant that is equivalent under natural selection. See FIG. 11 for an example of a function result that accounts for natural selection of a variant. For example, in some embodiments (see, e.g., FIG. 4), the enhancing / calibrating / recalibrating / updating / optimizing / correcting the plurality of insertion scores includes scaling an insertion propensity score according to a first scaling factor of the scaling factors associated with an insertion in the genome of the population. In such an example, the enhancing / calibrating / recalibrating / updating / optimizing / correcting the plurality of deletion scores includes scaling a deletion propensity score according to a second scaling factor of the scaling factors associated with a deletion in the genome of the population, and the enhancing / calibrating / recalibrating / updating / optimizing / correcting the plurality of missense pathogenicity scores includes scaling a missense propensity score according to a third scaling factor of the scaling factors associated with a variant in the genome of the population. Also, for example, in some embodiments (see, e.g., FIG. 3), enhancing / calibrating / recalibrating / updating / optimizing / correcting the plurality of indel pathogenicity scores includes scaling the indel propensity score according to a first one of the scaling factors associated with the indels in the genome of the population. In such examples, enhancing / calibrating / recalibrating / updating / optimizing / correcting the plurality of missense pathogenicity scores includes scaling the missense propensity score according to a second one of the scaling factors associated with the variants in the genome of the population.

[0053] In some embodiments, the equivalent variant of the variant is a synonymous mutation of the variant, and in some such instances, enhancing / calibrating / recalibrating / updating / optimizing / modifying the multiple missense pathogenicity scores comprises calibrating a missense propensity score based on a synonymous mutation for the variant.

[0054] In some embodiments, the equivalent variants of the indels are indels in coding and non-coding regions of the genome of the population. Also, in some of such examples, enhancing / calibrating / recalibrating / updating / optimizing / correcting the plurality of insertion scores comprises calibrating the insertion propensity score based on observed to expected ratios based on insertions occurring in coding regions versus non-coding regions of the genome of the population, respectively. Also, in such examples, enhancing / calibrating / recalibrating / updating / optimizing / correcting the plurality of deletion scores comprises calibrating the deletion propensity score based on observed to expected ratios based on deletions occurring in coding regions versus non-coding regions of the genome of the population, respectively.

[0055] In some embodiments, the group of bins represents all scores, with each bin of the group of bins representing a different range of scores in the all scores. Also, in some examples, the all scores include a plurality of insertion scores, a plurality of deletion scores, and a plurality of missense pathogenic scores. In some examples, the all scores include a plurality of indel pathogenic scores and a plurality of missense pathogenic scores. And, in yet still more general examples, the all scores include a plurality of first context scores and a plurality of second context scores.

[0056] In some embodiments, each bin of the group of bins is associated with a particular amount of insertions having a score within the respective score range associated with the bin. Also, each bin of the group of bins is associated with a particular amount of deletions having a score within the respective score range associated with the bin. Also, each bin of the group of bins is associated with a particular amount of variants having a score within the respective score range associated with the bin. The same is true for the more general example where the bins are associated with indel and missense pathogenic scores. The same is true for the even more general example where the bins are associated with a first context score and a second context score.

[0057] In some embodiments, the group of bins includes a group of percentile bins. For example, the group of percentile bins includes 100 bins, where the first bin of the 100 bins represents a score range of 0 to 0.01, and the 100th bin represents a score range of 0.99 to 1. The bins between the first bin and the 100th bin each include a percentile range of scores.

[0058] In some embodiments (see, e.g., FIG. 4), the indel pathogenicity score is generated by an artificial neural network (ANN) and the processing of the multiple insertions and multiple deletions is performed by the ANN. In some embodiments (see, e.g., FIG. 3), the indel pathogenicity score is generated by an ANN and the processing of the multiple indels is performed by the ANN. Also, in some embodiments (see, e.g., FIG. 2), the processing of the multiple first variations of the object is performed by the ANN and the processing of the multiple second variations of the object is performed by the ANN.

[0059] In some embodiments (see, e.g., Figures 3 and 4), the ANN is configured to classify the pathogenicity of the variant. In some examples of such embodiments, the ANN comprises a deep residual neural network for classifying the pathogenicity of missense mutations. Even more specifically, in some examples, the ANN comprises a version of PrimateAI.

[0060] FIG. 8 illustrates a method 700 of converting a first context of a computing system context intended to provide pathogenicity of variants in the genome of a population to a second context providing pathogenicity of indels in the genome of a population. For the sake of brevity, additional diagrams similar to FIG. 8 that separate the processing steps for indels into similar steps for insertions and deletions are not provided. Method 700 begins with step 702, which includes identifying a plurality of variants in a first genome database. Method 700 also begins with step 704, which includes identifying a plurality of indels in a second genome database. After steps 702 and 704, method 700 continues with the steps of method 200 or the steps of method 300, depending on the implementation of method 700.

[0061] As mentioned herein and with respect to Figure 8, it should be understood that multiple indels generally include multiple insertions and / or multiple deletions. It should also be understood that Figure 3 is a generalization of Figure 4. In other words, Figure 4 illustrates a more specific method also disclosed by Figure 3. Figure 3 relates to indels that can be insertions and / or deletions, while Figure 4 relates to embodiments involving both insertions and deletions.

[0062] In some embodiments, the first genome database comprises a version of the Genome Aggregation Database (gnomAD). In some of such embodiments, the second genome database comprises a version of gnomAD. In some cases, the second genome database and the first genome database are the same version of gnomAD, and in some other embodiments, the second and first genome databases are different versions of gnomAD.

[0063] 9 illustrates a method 800 of converting a first context of a computing system context intended to provide pathogenicity of variants in the genome of a population to a second context providing pathogenicity of indels in the genome of a population. For the sake of brevity, additional diagrams similar to FIG. 9 that separate the processing steps for indels into similar steps for insertions and deletions are not provided.

[0064] The method 800 begins at step 802, which includes identifying a plurality of variants in a first genomic database. The method 800 also begins at step 804, which includes identifying a plurality of indels in a second genomic database. The method 800 continues with an artificial neural network (ANN) generating a plurality of missense pathogenicity scores for each variant of the plurality of variants (at step 806). The method 800 also continues with the ANN generating a plurality of indel pathogenicity scores for each indel of the plurality of indels (at step 808). At step 810, the method 800 continues with further processing the plurality of indel pathogenicity scores and the plurality of missense pathogenicity scores to be applied to one or more curve-forming functions. At step 812, the method 800 continues with applying the further processed scores to a curve-forming function to generate an indel curve and a missense curve. At step 814, the method 800 continues with determining a selection pattern difference between the indel curve and the missense curve. At step 816, the method 800 continues with determining one or more scaling functions to reduce the selection pattern differences between the curves. At step 818, the method 800 continues with updating the coefficients of the ANN according to the scaling functions. In some embodiments, updating the coefficients of the ANN according to the scaling functions includes enhancing / calibrating / recalibrating / updating / optimizing / modifying the plurality of indel pathogenicity scores according to the scaling functions to provide a recalibrated accuracy of the indel pathogenicity score for each indel of the plurality of indels.

[0065] 9, in some embodiments, further processing the plurality of indel pathogenicity scores and the plurality of missense pathogenicity scores includes grouping the plurality of variants into groups of bins and grouping the plurality of indels into groups of bins. Further processing the plurality of indel pathogenicity scores and the plurality of missense pathogenicity scores also includes, for each bin of the group of bins, determining a central tendency distribution of the indel pathogenicity scores in the bin and determining a central tendency distribution of the missense pathogenicity scores in the bin. Applying the further processed scores to a curve-forming function to generate indel and missense curves includes applying the central tendencies of the indel pathogenicity scores and the missense pathogenicity scores to a curve-forming function to generate indel and missense curves.

[0066] With respect to Figure 9, in some embodiments, the curve-forming function comprises a function that accounts for the proportion of different indels and the proportion of different variants in the genome of the population. In some of such embodiments, the curve-forming function comprises a function that accounts for the natural selection of different indels and the natural selection of different variants in the genome of the population. For examples of the results of such a function that accounts for the natural selection of variants, see Figure 11.

[0067] 9, in some embodiments, the plurality of indels comprises a plurality of insertions and a plurality of deletions, and the plurality of indel pathogenicity scores comprises a plurality of insertion scores and a plurality of deletion scores, respectively. In some of such embodiments, applying the further processed scores in step 812 or method 800 generally includes: (1) generating an insertion curve based on the plurality of insertion scores according to a curve forming function; (2) generating a deletion curve based on the plurality of deletion scores according to a curve forming function; and (3) generating a missense curve based on the plurality of missense pathogenicity scores according to a curve forming function.

[0068] In some such embodiments, the insertion curve includes a first plurality of data points including an insertion propensity score for each bin of the group of bins. The deletion curve includes a second plurality of data points including a deletion propensity score for each bin of the group of bins. The missense curve includes a third plurality of data points including a missense propensity score for each bin of the group of bins. See Figures 12-14 for examples of such data points displayed on a graph. In some such embodiments, the insertion propensity score for a bin of the group of bins is related to a proportion of different insertions in the genome of a population having an insertion score of the plurality of insertion scores associated with the bin. The deletion propensity score for a bin of the group of bins is related to a proportion of different deletions in the genome of a population having a deletion score of the plurality of deletion scores associated with the bin. The missense propensity score for a bin of the group of bins is related to a proportion of variants in the genome of a population having a missense pathogenicity score of the plurality of missense pathogenicity scores associated with the bin.

[0069] With respect to FIG. 9 , in some such embodiments, generating the insertion curve includes grouping the plurality of insertions into a group of bins. Generating the insertion curve also includes, for each bin of the group of bins, (1) measuring a central tendency distribution of insertion scores in the bin, and (2) applying the central tendency distribution of insertion scores in the bin to identify an insertion propensity score for the bin. Generating the deletion curve also includes grouping the plurality of deletions into a group of bins. Generating the deletion curve also includes, for each bin of the group of bins, (1) measuring a central tendency distribution of deletion scores in the bin, and (2) applying the central tendency distribution of deletion scores in the bin to identify a deletion propensity score for the bin. Generating the missense curve also includes grouping the plurality of variants into a group of bins. Generating the missense curve also includes, for each bin of the group of bins, (1) measuring a central tendency distribution of missense pathogenicity scores in the bin, and (2) applying the central tendency distribution of missense pathogenicity scores in the bin to identify an insertion propensity score for the bin.

[0070] Also, with reference to FIG. 9, in some embodiments, the insertion propensity score for a bin of a group of bins represents the probability that one of the multiple insertions associated with the bin will occur in the genome of the population given the set of observed insertions. Also, the deletion propensity score for a bin represents the probability that one of the multiple deletions associated with the bin will occur in the genome of the population given the set of observed deletions. Also, the missense propensity score for a bin represents the probability that one of the multiple variants associated with the bin will occur in the genome of the population given the set of observed variants. In such an example, the propensity score reduces selection bias by equating groups based on covariates, where the covariates are the set of observed insertions, the set of observed deletions, and the set of observed variants, respectively.

[0071] Also, with reference to FIG. 9, in some embodiments, an insertion curve is generated when a first plurality of data points are plotted on a two-dimensional graph with one axis for the propensity score and the other axis for the group of bins. And, a deletion curve is generated when a second plurality of data points are plotted on the two-dimensional graph. And, a missense curve is generated when a third plurality of data points are plotted on the two-dimensional graph. In some such examples, the two-dimensional graph includes a set of ordered pairs (x, y), where f(x)=y, where x is the group of bins and y is the propensity score. See FIGS. 12-14 for examples of such data points displayed on the graphs.

[0072] 9 but inferred from some steps of method 800, the method includes determining a selection pattern difference between an insertion curve and a missense curve, determining one or more second scaling functions to reduce the selection pattern difference between the insertion curve and the missense curve, and enhancing / calibrating / recalibrating / updating / optimizing / modifying the plurality of insertion scores according to the second scaling functions to change the output of the ANN. Also, in such embodiments, method 800 includes determining a selection pattern difference between a deletion curve and a missense curve, determining one or more third scaling functions to reduce the selection pattern difference between the deletion curve and the missense curve, and enhancing / calibrating / recalibrating / updating / optimizing / modifying the plurality of deletion scores according to the third scaling functions to change the output of the ANN. In some embodiments, the one or more second scaling functions and the one or more third scaling functions are part of a scaling function that includes functions that scale the proportion of different insertions, different deletions, and different variants in the genomes of the population, respectively, since indels and single nucleotide variants have different variability.

[0073] Also with respect to Figure 9, in some embodiments, the enhancing / calibrating / recalibrating / updating / optimizing / correcting the plurality of insertion scores includes scaling an insertion propensity score according to a first scaling factor of the scaling factors associated with insertions in the genome of the population, the enhancing / calibrating / recalibrating / updating / optimizing / correcting the plurality of deletion scores includes scaling a deletion propensity score according to a second scaling factor of the scaling factors associated with deletions in the genome of the population, and the enhancing / calibrating / recalibrating / updating / optimizing / correcting the plurality of missense pathogenicity scores includes scaling a missense propensity score according to a third scaling factor of the scaling factors associated with variants in the genome of the population.

[0074] Also, with respect to FIG. 9, in some embodiments, the scaling function obtains the scaling factor from an equivalent variant under natural selection. See FIG. 11 for an example of the function result that describes natural selection of such variants. In some embodiments, the equivalent variant of the variant is a synonymous mutation of the variant. In such an example, enhancing / calibrating / recalibrating / updating / optimizing / correcting the multiple missense pathogenicity scores includes calibrating the missense propensity score based on a synonymous mutation for the variant. Also, in some embodiments, the equivalent variant of the indel is an indel in a coding and non-coding region of the genome of the population. In such an example, enhancing / calibrating / recalibrating / updating / optimizing / correcting the multiple insertion scores includes calibrating the insertion propensity score based on an observed to expected ratio based on insertions occurring in coding to non-coding regions of the genome of the population, respectively. Also, enhancing / calibrating / recalibrating / updating / optimizing / correcting multiple deletion scores includes calibrating deletion propensity scores based on observed vs. expected ratios based on deletions occurring in coding vs. non-coding regions of the genome of a population, respectively.

[0075] Also, with respect to FIG. 9, in some embodiments, the group of bins represents all scores. Each bin of the group of bins represents a different range of scores across all scores. And, all scores include multiple indel pathogenic scores and multiple missense pathogenic scores. In such and other examples, each bin of the group of bins is associated with a particular amount of multiple indels having scores within the respective range of scores associated with the bin. And, each bin of the group of bins is associated with a particular amount of multiple variants having scores within the respective range of scores associated with the bin.

[0076] 9, in some embodiments, the group of bins includes a group of percentile bins, and in some examples, the group of percentile bins includes 100 bins, where a first bin of the 100 bins represents a score in the range of 0 to 0.01, the 100th bin represents a score in the range of 0.99 to 1, and the bins between the first bin and the 100th bin each include a percentile range of scores.

[0077] Also, with reference to Figure 9, in some embodiments, the ANN is configured to classify the pathogenicity of the variant. In some examples of such embodiments, the ANN comprises a deep residual neural network for classifying the pathogenicity of the missense mutation. Even more specifically, in some examples, the ANN comprises a version of PrimateAI.

[0078] Also, with respect to Figure 9, in some embodiments, the first genome database includes a version of the Genome Aggregation Database (gnomAD). In some of such embodiments, the second genome database includes a version of gnomAD. In some cases, the second genome database and the first genome database are the same version of gnomAD, and in some other embodiments, the second and first genome databases are different versions of gnomAD.

[0079] Also with respect to FIG. 9 , in some embodiments, measuring the central tendency distribution of the indel pathogenicity scores comprises determining the mean of the indel pathogenicity scores. For example, measuring the central tendency distribution of the insertion scores comprises determining the mean of the insertion scores, and measuring the central tendency distribution of the deletion scores comprises determining the mean of the deletion scores. Also, in some embodiments, measuring the central tendency distribution of the missense pathogenicity scores comprises determining the mean of the missense pathogenicity scores. In some embodiments, measuring the central tendency distribution of the indel pathogenicity scores comprises determining the mode of the indel pathogenicity scores. For example, measuring the central tendency distribution of the insertion scores comprises determining the mode of the insertion scores, and measuring the central tendency distribution of the deletion scores comprises determining the mode of the deletion scores. Also, in some embodiments, measuring the central tendency distribution of the missense pathogenicity scores comprises determining the mode of the missense pathogenicity scores. In some embodiments, measuring the central tendency distribution of the indel pathogenicity scores comprises determining the median of the indel pathogenicity scores. For example, measuring the central tendency distribution of the insertion scores includes determining the median of the insertion scores, and measuring the central tendency distribution of the deletion scores includes determining the median of the deletion scores. Also, in some embodiments, measuring the central tendency distribution of the missense pathogenicity scores includes determining the median of the missense pathogenicity scores. Also, such techniques apply to more general embodiments as well. For example, measuring the central tendency distribution of the first context scores includes determining the mean, mode, or median of the first context scores. Also, measuring the central tendency distribution of the second context scores includes determining the mean, mode, or median of the second context scores.

[0080] 10 shows a block diagram of an exemplary embodiment of a computing system 900 that may include, be, or be part of any one of the electronic or computing systems described herein. FIG 10 illustrates a portion of the computing system 900 on which a set of instructions for causing a machine to perform any one or more of the methodologies discussed herein are executed.

[0081] In some embodiments, the computing system 900 corresponds to a host system that includes, is coupled to, or utilizes memory, or is used to perform operations performed by any one of the computing devices, data processors, and user interface devices described herein. In alternative embodiments, the machine is connected (e.g., networked) to other machines in a LAN, an intranet, an extranet, or the Internet. In some embodiments, the machine operates in the capacity of a server or a client machine in a client-server network environment, as a peer machine in a peer-to-peer (or distributed) network environment, or as a server or a client machine in a cloud computing infrastructure or environment. In some embodiments, the machine is a personal computer (PC), a tablet PC, a cellular phone, a web appliance, a server, or any machine capable of executing (in sequence or otherwise) a set of instructions that specify actions to be taken by the machine. Furthermore, although a single machine is illustrated, the term "machine" is also intended to include any collection of machines that individually or jointly execute a set (or sets) of instructions to perform any one or more of the methodologies discussed herein.

[0082] The computing system 900 includes a processing device 902, a main memory 904 (e.g., read-only memory (ROM), flash memory, dynamic random-access memory (DRAM), etc.), a static memory 906 (e.g., flash memory, static random-access memory (SRAM), etc.), and a data storage system 910, which communicate with each other via a bus 930. The processing device 902 represents one or more general-purpose processing devices, such as a microprocessor, a central processing unit, etc. More specifically, the processing device is a microprocessor or a processor that implements other instruction sets, or a processor that implements a combination of instruction sets. Alternatively, the processing device 902 is one or more special-purpose processing devices, such as an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), a digital signal processor (DSP), a network processor, etc. The processing device 902 is configured to execute instructions 914 to perform the operations or steps discussed herein. In some implementations, the computing system 900 includes a network interface device 908 for communicating over a communications network 940 shown in FIG.

[0083] The data storage system 910 includes a machine-readable storage medium 912 (also known as a computer-readable medium) on which one or more sets of instructions 914 or software embodying any one or more of the methodologies or functions described herein are stored. The instructions 914 also reside, completely or at least partially, within the main memory 904 or within the processing device 902 during execution thereof by the computing system 900, with the main memory 904 and the processing device 902 also constituting machine-readable storage media.

[0084] In some implementations, the instructions 914 include instructions for performing functions corresponding to any one of a computing device, a data processor, a user interface device, and an I / O device described herein. Although the machine-readable storage medium 912 is shown in the exemplary implementation to be a single medium, the term "machine-readable storage medium" should be interpreted to include a single medium or multiple media that store one or more sets of instructions. The term "machine-readable storage medium" should also be interpreted to include any medium capable of storing or encoding a set of instructions for execution by a machine, causing the machine to perform any one or more of the methodologies of the present disclosure. Thus, the term "machine-readable storage medium" should be interpreted to include solid-state memory, optical media, magnetic media, and the like.

[0085] Also as shown, the computing system 900 includes a user interface 920, which in some implementations includes a display, e.g., performing functions corresponding to any one of the user interface devices disclosed herein. A user interface, such as the user interface 920, or a user interface device as described herein, includes any space or equipment where interaction between a human and a machine occurs. A user interface as described herein allows operation and control of a machine from a human user, while the machine provides feedback information to the user. Examples of user interfaces (UIs) or user interface devices include computer operating systems (such as graphical user interfaces or GUIs), machine operator controls, and interactive aspects of process control. A UI as described herein includes one or more layers, including a human-machine interface (HMI) that interfaces the machine with physical input and output hardware.

[0086] It should also be appreciated that the methodologies discussed herein are computer-implemented methods and, in some embodiments, may be implemented by the computing system 900. For example, the computer-implemented method includes an artificial neural network (ANN) that generates a plurality of missense pathogenicity scores for each variant of the plurality of variants. The computer-implemented method also includes an ANN that generates a plurality of indel pathogenicity scores for each indel of the plurality of indels. The computer-implemented method further includes applying the plurality of indel pathogenicity scores and the plurality of missense pathogenicity scores to one or more curve-forming functions. The computer-implemented method also includes applying the processed scores to the curve-forming functions to generate an indel curve and a missense curve, and determining a selection pattern difference between the indel curve and the missense curve. The computer-implemented method also includes determining one or more scaling functions to reduce the selection pattern difference between the curves, and updating the coefficients of the ANN according to the scaling functions. Updating the coefficients of the ANN according to the scaling function includes augmenting / calibrating / recalibrating / updating / optimizing / modifying the multiple indel pathogenicity scores according to the scaling function to provide a recalibrated accuracy of the indel pathogenicity score for each indel of the multiple indels.

[0087] FIG. 11 illustrates a plot in a two-dimensional graph showing the relationship between binned PrimateAI scores for variants and insertion variants versus natural selection (i.e., the propensity of the variant or insertion in the genome of a population). In FIG. 11, natural selection values ​​(or propensity values) are represented on the y-axis. Also, the bins of the PrimateAI scores are represented on the x-axis, where the bins are ranges of PrimateAI variant pathogenicity score predictions. The different propensity scores described herein may be or include natural selection values. The bins of the PrimateAI scores may be or include any one of the groups of bins described herein.

[0088] FIG. 12 illustrates a scatter plot in a two-dimensional graph showing the relationship between the binned PrimateAI score for variants (green dots), insertion variants (blue dots), and deletion variants (orange dots) versus the proportion of observed variants (i.e., the propensity of variants or indels in the genome of a population). In FIG. 12, the proportion of observed variants is represented on the y-axis. And, the bins of the PrimateAI score are represented on the x-axis. The different propensity scores described herein may be or include the proportion of observed variants. The bins of the PrimateAI score may be or include any one of the groups of bins described herein.

[0089] FIG. 13 and FIG. 14 illustrate respective scatter plots in respective two-dimensional graphs, each plot showing the relationship between the binned PrimateAI score for variants (green dots), insertion variants (blue dots), and deletion variants (orange dots) versus the adjusted proportion of observed variants (i.e., the propensity of variants or indels in the genome of the population). In FIG. 13 and FIG. 14, the adjusted proportion (or adjusted ratio) of observed variants is represented on the y-axis. Also, the bins of the PrimateAI score are represented on the x-axis. Specifically, FIG. 13 relates to variants occurring 3 base pairs in-frame in the exome. Specifically, FIG. 14 relates to variants occurring 6 base pairs in-frame in the exome. The different scaled propensity scores described herein may be or include the adjusted proportion of observed variants. The bins of the PrimateAI score may be or include any one of the groups of bins described herein.

[0090] Some portions of the above detailed descriptions are presented in terms of algorithms and symbolic representations of operations on data bits within a computer memory. These algorithmic descriptions and representations are the way used by those skilled in the data processing arts to most effectively convey the substance of their work to others skilled in the art. An algorithm is here, and generally, conceived to be a self-consistent sequence of operations leading to a certain result. The operations require physical manipulations of physical quantities. Usually, although not necessarily, these quantities take the form of electrical or magnetic signals capable of being stored, combined, compared, and otherwise manipulated. It has proven convenient at times, primarily for reasons of common usage, to refer to these signals as bits, values, elements, symbols, characters, terms, numbers, or the like.

[0091] It should be noted, however, that these and similar terms are to be associated with the appropriate physical quantities and are merely convenient labels applied to these quantities. This disclosure may refer to actions and processes of a computing system, or similar electronic computing device, that manipulate and convert data represented as physical (electronic) quantities in the computing system's registers and memory into other data that is similarly represented as physical quantities in the computing system's memory or registers or other such information storage systems.

[0092] The present disclosure also relates to an apparatus for performing the operations herein. The apparatus may be specially constructed for the intended purposes or may include a general-purpose computer selectively activated or reconfigured by a computer program stored in the computer. Such a computer program may be stored in a computer-readable storage medium coupled to a computer system bus, such as any type of disk, including optical disks, CD-ROMs, and magneto-optical disks, read-only memory (ROM), random access memory (RAM), EPROM, EEPROM, magnetic or optical cards, or any type of medium suitable for storing electronic instructions.

[0093] The algorithms and displays presented herein are not inherently related to any particular computer or other apparatus. Various general-purpose systems may be used with programs in accordance with the teachings herein, or it may prove convenient to construct a more specialized apparatus to implement the methods. The structure of a variety of these systems will become apparent as described below. In addition, the present disclosure is not described with reference to any particular programming language. It will be understood that a variety of programming languages ​​can be used to implement the teachings of the present disclosure as described herein.

[0094] The present disclosure may be provided as a computer program product or software that may include a machine-readable medium having stored thereon instructions that may be used to program a computing system (or other electronic device) to perform a process according to the present disclosure. A machine-readable medium includes any mechanism for storing information in a form readable by a machine (e.g., a computer). In some embodiments, the machine-readable (e.g., computer-readable) medium includes a machine (e.g., computer) readable storage medium, such as a read-only memory ("ROM"), a random access memory ("RAM"), a magnetic disk storage medium, an optical storage medium, a flash memory component, or the like.

[0095] While the present invention has been described in conjunction with specific embodiments illustrated herein, it is apparent that many alternatives, combinations, modifications, and variations will be apparent to those skilled in the art. Accordingly, the exemplary embodiments of the invention described herein are intended to be illustrative only and not in a limiting sense. Various changes may be made without departing from the spirit and scope of the invention.

[0096] The inventors disclose the following provisions: Clause 1. A computer-implemented method comprising: processing the plurality of variants to generate a plurality of missense pathogenicity scores for each variant of the plurality of variants; generating a missense curve based on the plurality of missense pathogenicity scores according to one or more curve-forming functions; processing the plurality of indels to generate a plurality of indel pathogenicity scores for each indel of the plurality of indels; generating an indel curve based on the plurality of indel pathogenicity scores according to one or more curve-forming functions; determining a selection pattern difference between the indel and missense curves; determining one or more scaling functions to reduce the selection pattern difference between the missense and indel curves; augmenting / calibrating / recalibrating / updating / optimizing / modifying the plurality of indel pathogenicity scores according to one or more scaling functions to provide a recalibrated accuracy of the indel pathogenicity score for each indel of the plurality of indels. Clause 2. The computer-implemented method of clause 1, wherein the one or more curve-fitting functions include a function that describes the proportion of different indels and the proportion of different variants in the genomes of the population. Clause 3. The computer-implemented method of clause 2, wherein the one or more curve-forming functions include a function that accounts for natural selection of different indels and natural selection of different variants in the genome of the population. Clause 4. The computer-implemented method of clause 2, wherein the plurality of indels comprises a plurality of insertions and a plurality of deletions, and the plurality of indel pathogenicity scores comprises a plurality of insertion scores and a plurality of deletion scores, respectively. Article 5. generating an insertion curve based on the plurality of insertion scores according to one or more curve forming functions; and generating a deletion curve based on the plurality of deletion scores according to one or more curve-forming functions. Article 6. the inset curve comprises a first plurality of data points comprising an inset propensity score for each bin of the group of bins; the loss curve comprises a second plurality of data points comprising a loss propensity score for each bin of the group of bins; The computer-implemented method of clause 5, wherein the missense curve comprises a third plurality of data points comprising a missense propensity score for each bin of the group of bins. Article 7. an insertion propensity score for a bin of a group of bins related to a proportion of different insertions in the genome of a population having an insertion score among the plurality of insertion scores associated with the bin; a deletion propensity score for a bin of the group of bins is related to a proportion of different deletions in the genome of the population having a deletion score among the plurality of deletion scores associated with the bin; 7. The computer-implemented method of clause 6, wherein the missense propensity score for a bin of the group of bins is related to the proportion of variants in the genome of the population that have a missense pathogenicity score among the plurality of missense pathogenicity scores associated with the bin. Clause 8. The generation of insertion curves Grouping multiple insertions into groups of bins; For each bin in the group of bins, Measuring the central tendency distribution of insertion scores in the bins; and applying a central tendency distribution of the insertion scores across the bins to identify an insertion propensity score for the bin. Clause 9. The generation of deletion curves Grouping multiple deletions into groups of bins; For each bin in the group of bins, Measuring the central tendency distribution of deletion scores in the bins; and applying a central tendency distribution of the deletion scores in the bins to identify a deletion propensity score for the bin. Clause 10. The generation of a missense curve Grouping multiple variants into groups of bins; For each bin in the group of bins, Measuring the central tendency distribution of the missense pathogenicity scores in the bins; and and applying a central tendency distribution of the missense pathogenicity scores in the bins to identify an insertion propensity score for the bin. Article 11. The insertion propensity score for a bin of a group of bins represents the probability that one of the multiple insertions associated with the bin will occur in the genome of the population given the set of observed insertions; The deletion propensity score for a bin represents the probability that one of the deletions associated with the bin will occur in the genome of the population given the set of observed deletions; 8. The computer-implemented method of clause 7, wherein the missense propensity score for a bin represents the probability that one of a plurality of variants associated with the bin will occur in the genome of a population given the set of observed variants. Clause 12. The computer-implemented method of clause 11, wherein the insertion propensity score, the deletion propensity score, and the missense propensity score reduce selection bias by equating groups based on covariates, the covariates being the set of observed insertions, the set of observed deletions, and the set of observed variants, respectively. Article 13. The intercalation curve is generated when the first plurality of data points are plotted on a two-dimensional graph having one axis for the propensity score and another axis for the group of bins; A deletion curve is generated when the second plurality of data points are plotted on a two-dimensional graph; The computer-implemented method of clause 7, wherein a missense curve is generated when the third plurality of data points is plotted on a two-dimensional graph. Article 14. a two-dimensional graph includes a set of ordered pairs (x, y), f(x)=y, x is the group of bins, 14. The computer-implemented method of clause 13, wherein y is a propensity score. Article 15. determining a selection pattern difference between the insertion curve and the missense curve; determining one or more second scaling functions to reduce the selection pattern difference between the insertion and missense curves; and augmenting / calibrating / recalibrating / updating / optimizing / modifying the plurality of insertion scores according to one or more second scaling functions to provide a recalibrated accuracy of the insertion pathogenicity score for each insertion of the plurality of insertions. Article 16. determining a selection pattern difference between the deletion curve and the missense curve; determining one or more third scaling functions to reduce the selection pattern difference between the deletion curves and the missense curves; augmenting / calibrating / recalibrating / updating / optimizing / modifying the plurality of deletion scores according to one or more third scaling functions to provide a recalibrated accuracy of the deletion pathogenicity score for each deletion of the plurality of deletions. Article 17. the one or more second scaling functions and the one or more third scaling functions are part of the one or more scaling functions; 17. The computer-implemented method of clause 16, wherein the one or more scaling functions include functions that scale the proportion of distinct insertions, distinct deletions, and distinct variants, respectively, in the genomes of the population, since indels and single nucleotide variants have different mutabilities. Clause 18. The computer-implemented method of clause 17, wherein the one or more scaling functions derive their scaling factors from equivalent variants under natural selection. Article 19. augmenting / calibrating / recalibrating / updating / optimizing / correcting the plurality of insertion scores comprises scaling the plurality of insertion propensity scores according to a first scaling factor among the scaling factors associated with insertions in the genome of the population; augmenting / calibrating / recalibrating / updating / optimizing / correcting the plurality of deletion scores comprises scaling the plurality of deletion propensity scores according to a second one of the scaling factors associated with deletions in the genome of the population; 19. The computer-implemented method of clause 18, wherein enhancing / calibrating / recalibrating / updating / optimizing / correcting the plurality of missense pathogenicity scores comprises scaling the plurality of missense propensity scores according to a third scaling factor among the scaling factors associated with the variants in the genome of the population. Clause 20. The computer-implemented method of clause 19, wherein an equivalent variant of a variant is a synonymous mutation to the variant. Clause 21. The computer-implemented method of clause 20, wherein the enhancing / calibrating / recalibrating / updating / optimizing / modifying the plurality of missense pathogenicity scores comprises calibrating a missense propensity score based on synonymous mutations for the variant. Clause 22. The computer-implemented method of clause 19, wherein the equivalent variants of an indel are indels in coding and non-coding regions of the genome of the population. Article 23. augmenting / calibrating / recalibrating / updating / optimizing / correcting the plurality of insertion scores, each comprising calibrating an insertion propensity score based on observed vs. expected ratios based on insertions occurring in coding vs. non-coding regions of the genome of the population; 23. The computer-implemented method of clause 22, wherein enhancing / calibrating / recalibrating / updating / optimizing / correcting the plurality of deletion scores comprises calibrating deletion propensity scores based on observed vs. expected ratios based on deletions occurring in coding vs. non-coding regions of the genome of the population, respectively. Article 24. A group of bins represents the total score, each bin in the group of bins represents a different range of scores across all scores; 7. The computer-implemented method of clause 6, wherein all scores include a plurality of insertion scores, a plurality of deletion scores, and a plurality of missense pathogenic scores. Article 25. each bin of the group of bins is associated with a particular quantity of a plurality of insertions having scores within a respective range of scores associated with the bin; each bin of the group of bins is associated with a particular amount of deletions having scores within a range of respective scores associated with the bin; 25. The computer-implemented method of claim 24, wherein each bin of the group of bins is associated with a particular quantity of a plurality of variants having scores within a respective range of scores associated with the bin. Clause 26. The computer-implemented method of clause 25, wherein the group of bins comprises a group of percentile bins. Clause 27. The computer-implemented method of clause 26, wherein the group of percentile bins includes 100 bins, a first bin of the 100 bins representing scores in the range of 0 to 0.01, a 100th bin representing scores in the range of 0.99 to 1, and bins between the first bin and the 100th bin each comprising a range of percentile scores. Clause 28. The computer-implemented method of clause 1, wherein the plurality of indel pathogenicity scores are generated by an artificial neural network (ANN) and the processing of the plurality of indels is performed by the ANN. Clause 29. The computer-implemented method of clause 28, wherein the ANN is configured to classify the pathogenicity of the variant. Clause 30. The computer-implemented method of clause 29, wherein the ANN comprises a deep residual neural network for classifying pathogenicity of missense mutations. Clause 31. The computer-implemented method of clause 30, wherein the ANN includes a version of PrimateAI. Clause 32. The computer-implemented method of clause 1, further comprising identifying a plurality of variants in the first genome database. Clause 33. The computer-implemented method of clause 32, wherein the first genome database comprises a version of the Genome Aggregation Database (gnomAD). Clause 34. The computer-implemented method of clause 32, further comprising identifying a plurality of indels in a second genome database. Clause 35. The computer-implemented method of clause 34, wherein the second genome database comprises a version of gnomAD. Article 36. Identifying a plurality of variants in a first genome database; and identifying a plurality of indels in a second genome database, 2. The computer-implemented method of claim 1, wherein the first and second genome databases are part of one or more versions of the Genome Aggregation Database (gnomAD). Clause 37. The computer-implemented method of clause 1, wherein generating the indel curve comprises grouping a plurality of indels into groups of bins. Clause 38. The computer-implemented method of clause 37, wherein generating the missense curve comprises grouping the plurality of variants into groups of bins. Clause 39. The generation of an indel curve comprises, for each bin of a group of bins: Measuring the central tendency distribution of indel pathogenicity scores in the bins; and and applying a central tendency distribution of the missense pathogenicity scores in the bins to identify an insertion propensity score for the bin. Clause 40. The generation of a missense curve comprises, for each bin of a group of bins: Measuring the central tendency distribution of the missense pathogenicity scores in the bins; and and applying a central tendency distribution of the missense pathogenicity scores in the bins to identify a missense propensity score for the bin. Clause 41. The computer-implemented method of clause 39, wherein measuring the central tendency distribution of the indel pathogenicity scores comprises determining a mean of the indel pathogenicity scores. Clause 42. The computer-implemented method of clause 40, wherein measuring the central tendency distribution of the missense pathogenicity scores comprises determining a mean of the missense pathogenicity scores. Clause 43. The computer-implemented method of clause 39, wherein measuring the central tendency distribution of the indel pathogenicity scores comprises determining a median of the indel pathogenicity scores. Clause 44. The computer-implemented method of clause 40, wherein measuring the central tendency distribution of the missense pathogenicity scores comprises determining a median indel pathogenicity score. Clause 45. The computer-implemented method of clause 39, wherein measuring the central tendency distribution of the indel pathogenicity scores comprises determining a mode of the indel pathogenicity scores. Clause 46. The computer-implemented method of clause 40, wherein measuring the central tendency distribution of the missense pathogenicity scores comprises determining a mode of the indel pathogenicity scores. Clause 47. The computer-implemented method of clause 8, wherein measuring the central tendency distribution of the insertion scores includes determining a mean of the insertion scores. Clause 48. The computer-implemented method of clause 9, wherein measuring the central tendency distribution of the deletion scores comprises determining a mean of the deletion scores. Clause 49. The computer-implemented method of clause 10, wherein measuring the central tendency distribution of the missense pathogenicity scores comprises determining a mean of the missense pathogenicity scores. Clause 50. The computer-implemented method of clause 10, wherein measuring the central tendency of the insertion scores, deletion scores, and missense pathogenicity scores comprises determining the mode or median of the scores. Clause 51. A computer-implemented method comprising: Identifying a plurality of variants in a first genome database; Identifying a plurality of indels in a second genome database; and generating a plurality of missense pathogenicity scores for each variant of the plurality of variants by an artificial neural network (ANN); generating a plurality of indel pathogenicity scores for each indel of the plurality of indels by the ANN; applying the plurality of indel pathogenicity scores and the plurality of missense pathogenicity scores to one or more curve-fitting functions; further processing the plurality of missense pathogenicity scores and the plurality of indel pathogenicity scores using one or more curve-forming functions to generate indel and missense curves; determining a selection pattern difference between the indel and missense curves; determining one or more scaling functions to reduce the selection pattern difference between the indel and missense curves; and updating coefficients of the ANN according to one or more scaling functions. Article 52. Further processing the plurality of indel pathogenicity scores and the plurality of missense pathogenicity scores using one or more curve-fitting functions. Grouping multiple variants into groups of bins; Grouping multiple indels into groups of bins; For each bin in the group of bins, Measuring the central tendency distribution of indel pathogenicity scores in the bins; and Measuring the central tendency distribution of the missense pathogenicity scores in the bins; and and applying the central tendencies of the indel pathogenicity scores and the missense pathogenicity scores to one or more curve-forming functions to generate indel and missense curves. Clause 53. The computer-implemented method of clause 51, wherein the one or more curve-fitting functions include a function that describes the proportion of different indels and the proportion of different variants in the genomes of the population. Clause 54. The computer-implemented method of clause 53, wherein the one or more curve-forming functions comprise a function that accounts for natural selection of different indels and natural selection of different variants in the genome of the population. Clause 55. The computer-implemented method of clause 53, wherein the plurality of indels comprises a plurality of insertions and a plurality of deletions, and the plurality of indel pathogenicity scores comprises a plurality of insertion scores and a plurality of deletion scores, respectively. Article 56. generating an insertion curve based on the plurality of insertion scores according to one or more curve forming functions; generating a deletion curve based on the plurality of deletion scores according to one or more curve-fitting functions; and generating a missense curve based on the plurality of missense pathogenicity scores according to one or more curve-forming functions. Article 57. the inset curve comprises a first plurality of data points comprising an inset propensity score for each bin of the group of bins; the loss curve comprises a second plurality of data points comprising a loss propensity score for each bin of the group of bins; 57. The computer-implemented method of clause 56, wherein the missense curve comprises a third plurality of data points comprising a missense propensity score for each bin of the group of bins. Article 58. an insertion propensity score for a bin of a group of bins related to a proportion of different insertions in the genome of a population having an insertion score among the plurality of insertion scores associated with the bin; a deletion propensity score for a bin of the group of bins is related to a proportion of different deletions in the genome of the population having a deletion score among the plurality of deletion scores associated with the bin; 58. The computer-implemented method of clause 57, wherein the missense propensity score for a bin of the group of bins is related to the proportion of variants in the genome of the population that have a missense pathogenicity score among the plurality of missense pathogenicity scores associated with the bin. Article 59. The generation of insertion curves Grouping multiple insertions into groups of bins; For each bin in the group of bins, Measuring the central tendency distribution of insertion scores in the bins; and applying a central tendency distribution of the insertion scores across the bins to identify an insertion propensity score for the bin. Article 60. The generation of deletion curves Grouping multiple deletions into groups of bins; For each bin in the group of bins, Measuring the central tendency distribution of deletion scores in the bins; and applying a central tendency distribution of the deletion scores in the bins to identify a deletion propensity score for the bin. Article 61. The generation of missense curves Grouping multiple variants into groups of bins; For each bin in the group of bins, Measuring the central tendency distribution of the missense pathogenicity scores in the bins; and and applying a central tendency distribution of the missense pathogenicity scores in the bins to identify an insertion propensity score for the bin. Article 62. The insertion propensity score for a bin of a group of bins represents the probability that one of the multiple insertions associated with the bin will occur in the genome of the population given the set of observed insertions; The deletion propensity score for a bin represents the probability that one of the deletions associated with the bin will occur in the genome of the population given the set of observed deletions; 59. The computer-implemented method of clause 58, wherein the missense propensity score for a bin represents the probability that one of a plurality of variants associated with the bin will occur in the genome of a population given the set of observed variants. Clause 63. The computer-implemented method of clause 62, wherein the insertion propensity score, the deletion propensity score, and the missense propensity score reduce selection bias by equating groups based on covariates, the covariates being the set of observed insertions, the set of observed deletions, and the set of observed variants, respectively. Article 64. The intercalation curve is generated when the first plurality of data points are plotted on a two-dimensional graph having one axis for the propensity score and another axis for the group of bins; A deletion curve is generated when the second plurality of data points are plotted on a two-dimensional graph; 59. The computer-implemented method of clause 58, wherein the missense curve is generated when the third plurality of data points is plotted on a two-dimensional graph. Article 65. a two-dimensional graph includes a set of ordered pairs (x, y), f(x)=y, x is the group of bins, 65. The computer-implemented method of clause 64, wherein y is a propensity score. Article 66. determining a selection pattern difference between the insertion curve and the missense curve; determining one or more second scaling functions to reduce the selection pattern difference between the insertion and missense curves; and augmenting / calibrating / recalibrating / updating / optimizing / modifying the plurality of insertion scores according to one or more second scaling functions to alter an output of the ANN. Article 67. determining a selection pattern difference between the deletion curve and the missense curve; determining one or more third scaling functions to reduce the selection pattern difference between the deletion curves and the missense curves; and augmenting / calibrating / recalibrating / updating / optimizing / modifying the plurality of deletion scores according to one or more third scaling functions to alter the output of the ANN. Article 68. the one or more second scaling functions and the one or more third scaling functions are part of the one or more scaling functions; 68. The computer-implemented method of clause 67, wherein the one or more scaling functions include functions that scale the proportion of distinct insertions, distinct deletions, and distinct variants, respectively, in the genomes of the population, since indels and single nucleotide variants have different mutabilities. Clause 69. The computer-implemented method of clause 68, wherein the one or more scaling functions derive scaling factors from equivalent variants under natural selection. Article 70. augmenting / calibrating / recalibrating / updating / optimizing / correcting the plurality of insertion scores comprises scaling the plurality of insertion propensity scores according to a first scaling factor among the scaling factors associated with insertions in the genome of the population; augmenting / calibrating / recalibrating / updating / optimizing / correcting the plurality of deletion scores comprises scaling the plurality of deletion propensity scores according to a second one of the scaling factors associated with deletions in the genome of the population; 70. The computer-implemented method of clause 69, wherein enhancing / calibrating / recalibrating / updating / optimizing / correcting the plurality of missense pathogenicity scores comprises scaling the plurality of missense propensity scores according to a third scaling factor among the scaling factors associated with the variants in the genome of the population. Clause 71. The computer-implemented method of clause 68, wherein an equivalent variant of a variant is a synonymous mutation to the variant. Clause 72. The computer-implemented method of clause 71, wherein the enhancing / calibrating / recalibrating / updating / optimizing / modifying the multiple missense pathogenicity scores comprises calibrating a missense propensity score based on synonymous mutations for the variant. Clause 73. The computer-implemented method of clause 72, wherein the equivalent variants of an indel are indels in coding and non-coding regions of the genome of the population. Article 74. augmenting / calibrating / recalibrating / updating / optimizing / correcting the plurality of insertion scores, each comprising calibrating an insertion propensity score based on observed vs. expected ratios based on insertions occurring in coding vs. non-coding regions of the genome of the population; 23. The computer-implemented method of clause 22, wherein enhancing / calibrating / recalibrating / updating / optimizing / correcting the plurality of deletion scores comprises calibrating deletion propensity scores based on observed vs. expected ratios based on deletions occurring in coding vs. non-coding regions of the genome of the population, respectively. Article 75. A group of bins represents the total score, each bin in the group of bins represents a different range of scores across all scores; 53. The computer-implemented method of clause 52, wherein all scores include a plurality of indel pathogenicity scores and a plurality of missense pathogenicity scores. Article 76. each bin of the group of bins is associated with a particular amount of a plurality of indels having scores within a respective range of scores associated with the bin; 76. The computer-implemented method of claim 75, wherein each bin of the group of bins is associated with a particular quantity of a plurality of variants having scores within a respective range of scores associated with the bin. Clause 77. The computer-implemented method of clause 76, wherein the group of bins comprises a group of percentile bins. Clause 78. The computer-implemented method of clause 77, wherein the group of percentile bins includes 100 bins, a first bin of the 100 bins representing scores in the range of 0 to 0.01, a 100th bin representing scores in the range of 0.99 to 1, and bins between the first bin and the 100th bin each comprising a range of percentile scores. Clause 79. The computer-implemented method of clause 51, wherein the ANN is configured to classify the pathogenicity of the variant. Clause 80. The computer-implemented method of clause 79, wherein the ANN comprises a deep residual neural network for classifying pathogenicity of missense mutations. Clause 81. The computer-implemented method of clause 80, wherein the ANN includes a version of PrimateAI. Clause 82. The computer-implemented method of clause 51, wherein the first genome database comprises a version of the Genome Aggregation Database (gnomAD). Clause 83. The computer-implemented method of clause 82, wherein the second genome database comprises a version of gnomAD. Clause 84. The computer-implemented method of clause 52, wherein measuring the central tendency distribution of the indel pathogenicity scores in the bins comprises determining a mean of the indel pathogenicity scores. Clause 85. The computer-implemented method of clause 52, wherein measuring the central tendency distribution of the missense pathogenicity scores in the bins comprises determining a mean of the missense pathogenicity scores. Clause 86. The computer-implemented method of clause 52, wherein measuring the central tendency of the indel pathogenicity scores and missense pathogenicity scores in the bins comprises determining the mode or median of the scores. Clause 87. The computer-implemented method of clause 59, wherein measuring the central tendency distribution of the insertion scores includes determining a mean of the insertion scores. Clause 88. The computer-implemented method of clause 60, wherein measuring the central tendency distribution of the deletion scores comprises determining a mean of the deletion scores. Clause 89. The computer-implemented method of clause 61, wherein measuring the central tendency distribution of the missense pathogenicity scores in the bins comprises determining a mean of the missense pathogenicity scores. Clause 90. The computer-implemented method of clause 59, wherein measuring the central tendency distribution of the insertion scores comprises determining a mode or a median of the insertion scores. Clause 91. The computer-implemented method of clause 60, wherein measuring the central tendency distribution of the deletion scores comprises determining the mode or median of the deletion scores. Clause 92. The computer-implemented method of clause 61, wherein measuring the central tendency distribution of the missense pathogenicity scores in the bins comprises determining the mode or median of the missense pathogenicity scores. Clause 93. The computer-implemented method of clause 51, wherein updating the coefficients of the ANN according to the one or more scaling functions includes augmenting / calibrating / recalibrating / updating / optimizing / modifying the plurality of indel pathogenicity scores according to the one or more scaling functions to provide a recalibrated accuracy of the indel pathogenicity score for each indel of the plurality of indels. Clause 94. A computer-implemented method comprising: generating a plurality of missense pathogenicity scores for each variant of the plurality of variants by an artificial neural network (ANN); generating a plurality of indel pathogenicity scores for each indel of the plurality of indels by the ANN; applying the plurality of indel pathogenicity scores and the plurality of missense pathogenicity scores to one or more curve-fitting functions; further processing the plurality of indel pathogenicity scores and the plurality of missense pathogenicity scores using one or more curve-forming functions to generate indel curves and missense curves; determining a selection pattern difference between the indel and missense curves; determining one or more scaling functions to reduce the selection pattern difference between the indel and missense curves; updating coefficients of the ANN according to one or more scaling functions; A computer-implemented method, wherein updating the coefficients of the ANN according to one or more scaling functions includes augmenting / calibrating / recalibrating / updating / optimizing / modifying the plurality of indel pathogenicity scores according to one or more scaling functions to provide a recalibrated accuracy of the indel pathogenicity score for each indel of the plurality of indels.

Claims

1. 1. A computer-implemented method comprising: processing the plurality of variants to generate a plurality of missense pathogenicity scores for each variant in the plurality of variants; generating a missense curve based on the plurality of missense pathogenicity scores according to one or more curve-forming functions; processing the plurality of indels to generate a plurality of indel pathogenicity scores for each indel of the plurality of indels; generating an indel curve based on the plurality of indel pathogenicity scores according to the one or more curve-forming functions; determining a selection pattern difference between the indel curve and the missense curve; determining one or more scaling functions to reduce the selection pattern difference between the missense and indel curves; and modifying the plurality of indel pathogenicity scores according to the one or more scaling functions to provide a recalibrated accuracy of indel pathogenicity score for each indel of the plurality of indels.

2. 2. The computer-implemented method of claim 1, wherein the one or more curve-fitting functions comprise a function that describes the proportion of different indels and the proportion of different variants in the genomes of a population.

3. 3. The computer-implemented method of claim 2, wherein the one or more curve-forming functions comprise a function that accounts for natural selection of different indels and natural selection of different variants in the genomes of the population.

4. 3. The computer-implemented method of claim 2, wherein the plurality of indels comprises a plurality of insertions and a plurality of deletions, and the plurality of indel pathogenicity scores comprises a plurality of insertion scores and a plurality of deletion scores, respectively.

5. generating an insertion curve based on the plurality of insertion scores according to the one or more curve forming functions; 5. The computer-implemented method of claim 4, further comprising: generating a deletion curve based on the plurality of deletion scores according to the one or more curve-fitting functions.

6. the inset curve comprises a first plurality of data points comprising an inset propensity score for each bin of a group of bins; the loss curve comprises a second plurality of data points comprising a loss propensity score for each bin of the group of bins; 6. The computer-implemented method of claim 5, wherein the missense curve comprises a third plurality of data points comprising a missense propensity score for each bin of the group of bins.

7. the insertion propensity score for a bin of the group of bins is related to a proportion of different insertions in the genome of the population having an insertion score among the plurality of insertion scores associated with the bin; the deletion propensity score for a bin of the group of bins is related to a proportion of different deletions in the genome of the population having a deletion score of the plurality of deletion scores associated with the bin; 7. The computer-implemented method of claim 6, wherein the missense propensity score for a bin of the group of bins is related to the proportion of variants in the genome of the population that have a missense pathogenicity score among the plurality of missense pathogenicity scores associated with the bin.

8. said generating said insertion curve comprising: grouping the plurality of insertions into groups of bins; For each bin in the group of bins, measuring a central tendency distribution of the insertion scores in the bins; and and applying the central tendency distribution of the insertion scores in the bins to identify the insertion propensity score for the bin.

9. said generating said deletion curve comprising: grouping said plurality of deletions into said groups of bins; For each bin in the group of bins, measuring a central tendency distribution of the deletion scores in the bins; and and applying the central tendency distribution of the deletion scores in the bins to identify the deletion propensity score for the bin.

10. said generating said missense curve further comprising: grouping said plurality of variants into said group of bins; For each bin in the group of bins, determining a central tendency distribution of the missense pathogenicity scores in the bins; and and applying the central tendency distribution of the missense pathogenicity scores in the bins to identify the insertion propensity score for the bin.

11. the insertion propensity score for a bin of the group of bins represents the probability that one of the multiple insertions associated with the bin will occur in the genome of the population given a set of observed insertions; the deletion propensity score for the bin represents the probability that one of the deletions associated with the bin will occur in the genome of the population given a set of observed deletions; 8. The computer-implemented method of claim 7, wherein the missense propensity score for the bin represents a probability that one of the multiple variants associated with the bin occurs in the genome of the population given a set of observed variants.

12. 12. The computer-implemented method of claim 11, wherein the insertion propensity score, the deletion propensity score, and the missense propensity score reduce selection bias by equalizing groups based on covariates, the covariates being the set of observed insertions, the set of observed deletions, and the set of observed variants, respectively.

13. the interpolation curve is generated when the first plurality of data points are plotted on a two-dimensional graph having one axis for propensity score and another axis for the group of bins; the defect curve is generated when the second plurality of data points are plotted on the two-dimensional graph; 8. The computer-implemented method of claim 7, wherein the missense curve is generated when the third plurality of data points are plotted on the two-dimensional graph.

14. the two-dimensional graph comprises a set of ordered pairs (x, y); f(x)=y, x is a group of bins; The computer-implemented method of claim 13 , wherein y is the propensity score.

15. determining a selection pattern difference between the insertion curve and the missense curve; determining one or more second scaling functions to reduce the selection pattern difference between the insertion curve and the missense curve; 14. The computer-implemented method of claim 13, further comprising: modifying the plurality of insertion scores according to the one or more second scaling functions to provide a recalibrated accuracy of an insertion pathogenicity score for each insertion of the plurality of insertions.

16. determining a selection pattern difference between the deletion curve and the missense curve; determining one or more third scaling functions to reduce the selection pattern difference between the deletion curves and the missense curves; 16. The computer-implemented method of claim 15, further comprising: modifying the plurality of deletion scores according to the one or more third scaling functions to provide a recalibrated accuracy of deletion pathogenicity score for each deletion of the plurality of deletions.

17. the one or more second scaling functions and the one or more third scaling functions are part of the one or more scaling functions; 17. The computer-implemented method of claim 16, wherein the one or more scaling functions comprise functions that scale the proportion of distinct insertions, distinct deletions, and distinct variants in the genomes of the population, respectively, since indels and single nucleotide variants have different variability.

18. 20. The computer-implemented method of claim 17, wherein the one or more scaling functions obtain scaling factors from equivalent variants under natural selection.

19. wherein the modifying the plurality of insertion scores comprises scaling a plurality of insertion propensity scores according to a first one of the scaling factors associated with insertions in the genome of the population; wherein the modifying the plurality of deletion scores comprises scaling a plurality of deletion propensity scores according to a second one of the scaling factors associated with deletions in the genome of the population; 20. The computer-implemented method of claim 18, wherein the modifying the plurality of missense pathogenicity scores comprises scaling a plurality of missense propensity scores according to a third one of the scaling factors associated with variants in the genome of the population.

20. the group of bins represents the total score; each bin of said group of bins represents a different range of scores across all of said scores; 7. The computer-implemented method of claim 6, wherein said total scores comprise said plurality of insertion scores, said plurality of deletion scores, and said plurality of missense pathogenic scores.

21. each bin of the group of bins is associated with a particular quantity of the plurality of insertions having scores within a respective range of scores associated with the bin; each bin of the group of bins is associated with a particular amount of the plurality of deletions having a score within a respective range of scores associated with the bin; 21. The computer-implemented method of claim 20, wherein each bin of the group of bins is associated with a particular amount of the plurality of variants having scores within a respective range of scores associated with the bin.

22. 22. The computer-implemented method of claim 21, wherein the group of bins comprises a group of percentile bins.

23. 23. The computer-implemented method of claim 22, wherein the group of percentile bins includes 100 bins, a first bin of the 100 bins represents a score range of 0 to 0.01, a 100th bin represents a score range of 0.99 to 1, and bins between the first bin and the 100th bin each include a percentile range of scores.

24. 2. The computer-implemented method of claim 1, wherein the plurality of indel pathogenicity scores are generated by an artificial neural network (ANN) and the processing of the plurality of indels is performed by the ANN.

25. 25. The computer-implemented method of claim 24, wherein the ANN is configured to classify pathogenicity of a variant.

26. 26. The computer-implemented method of claim 25, wherein the ANN comprises a deep residual neural network for classifying pathogenicity of missense mutations.

27. 10. The computer-implemented method of claim 1, further comprising identifying the plurality of variants in a first genome database.

28. 28. The computer-implemented method of claim 27, further comprising identifying the plurality of indels in a second genome database.

29. 1. A computer-implemented method comprising: Identifying a plurality of variants in a first genome database; Identifying a plurality of indels in a second genome database; generating a plurality of missense pathogenicity scores for each variant of the plurality of variants by an artificial neural network (ANN); generating a plurality of indel pathogenicity scores for each indel of the plurality of indels by the ANN; applying the plurality of indel pathogenicity scores and the plurality of missense pathogenicity scores to one or more curve-forming functions; further processing the plurality of missense pathogenicity scores and the plurality of indel pathogenicity scores using the one or more curve-forming functions to generate indel and missense curves; determining a selection pattern difference between the indel curve and the missense curve; determining one or more scaling functions to reduce the selection pattern difference between the indel and missense curves; and updating coefficients of the ANN according to the one or more scaling functions.

30. further processing the plurality of indel pathogenicity scores and the plurality of missense pathogenicity scores using the one or more curve-fitting functions; grouping the plurality of variants into groups of bins; grouping the plurality of indels into the group of bins; For each bin in the group of bins, determining a central tendency distribution of indel pathogenicity scores in the bins; and determining a central tendency distribution of missense pathogenicity scores in the bins; and and applying the central tendencies of the indel pathogenicity scores and the missense pathogenicity scores to the one or more curve-forming functions to generate the indel curves and the missense curves.

31. 30. The computer-implemented method of claim 29, wherein the one or more curve-fitting functions comprise a function that describes the proportion of different indels and the proportion of different variants in the genomes of a population.

32. 32. The computer-implemented method of claim 31 , wherein the one or more curve-forming functions comprise a function that accounts for natural selection of different indels and natural selection of different variants in the genomes of the population.

33. 32. The computer-implemented method of claim 31 , wherein the plurality of indels comprises a plurality of insertions and a plurality of deletions, and the plurality of indel pathogenicity scores comprises a plurality of insertion scores and a plurality of deletion scores, respectively.

34. generating an insertion curve based on the plurality of insertion scores according to the one or more curve forming functions; generating a deletion curve based on the plurality of deletion scores according to the one or more curve-fitting functions; 34. The computer-implemented method of claim 33, further comprising generating the missense curve based on the plurality of missense pathogenicity scores according to the one or more curve-forming functions.

35. the inset curve comprises a first plurality of data points comprising an inset propensity score for each bin of a group of bins; the loss curve comprises a second plurality of data points comprising a loss propensity score for each bin of the group of bins; 35. The computer-implemented method of claim 34, wherein the missense curve comprises a third plurality of data points comprising a missense propensity score for each bin of the group of bins.

36. the insertion propensity score for a bin of the group of bins is related to a proportion of different insertions in the genome of the population having an insertion score among the plurality of insertion scores associated with the bin; the deletion propensity score for a bin of the group of bins is related to a proportion of different deletions in the genome of the population having a deletion score of the plurality of deletion scores associated with the bin; 36. The computer-implemented method of Claim 35, wherein the missense propensity score for a bin of the group of bins is related to the proportion of variants in the genome of the population that have a missense pathogenicity score among the plurality of missense pathogenicity scores associated with the bin.

37. said generating said insertion curve comprising: grouping the plurality of insertions into groups of bins; For each bin in the group of bins, measuring a central tendency distribution of the insertion scores in the bins; and and applying the central tendency distribution of the insertion scores in the bins to identify the insertion propensity score for the bin.

38. said generating said deletion curve comprising: grouping said plurality of deletions into said groups of bins; For each bin in the group of bins, measuring a central tendency distribution of the deletion scores in the bins; and and applying the central tendency distribution of the deletion scores in the bins to identify the deletion propensity score for the bin.

39. said generating said missense curve further comprising: grouping said plurality of variants into said group of bins; For each bin in the group of bins, determining a central tendency distribution of missense pathogenicity scores in the bins; and and applying the central tendency distribution of the missense pathogenicity scores in the bins to identify the insertion propensity score for the bin.

40. the insertion propensity score for a bin of the group of bins represents the probability that one of the multiple insertions associated with the bin will occur in the genome of the population given a set of observed insertions; the deletion propensity score for the bin represents the probability that one of the deletions associated with the bin will occur in the genome of the population given a set of observed deletions; 37. The computer-implemented method of Claim 36, wherein the missense propensity score for the bin represents a probability that one of the multiple variants associated with the bin occurs in the genome of the population given a set of observed variants.

41. 41. The computer-implemented method of claim 40, wherein the insertion propensity score, the deletion propensity score, and the missense propensity score reduce selection bias by equalizing groups based on covariates, the covariates being the set of observed insertions, the set of observed deletions, and the set of observed variants, respectively.

42. the interpolation curve is generated when the first plurality of data points are plotted on a two-dimensional graph having one axis for propensity score and another axis for the group of bins; the defect curve is generated when the second plurality of data points are plotted on the two-dimensional graph; 37. The computer-implemented method of claim 36, wherein the missense curve is generated when the third plurality of data points are plotted on the two-dimensional graph.

43. the two-dimensional graph comprises a set of ordered pairs (x, y); f(x)=y, x is a group of bins; 43. The computer-implemented method of claim 42, wherein y is the propensity score.

44. determining a selection pattern difference between the insertion curve and the missense curve; determining one or more second scaling functions to reduce the selection pattern difference between the insertion curve and the missense curve; 43. The computer-implemented method of claim 42, further comprising modifying the plurality of insertion scores according to the one or more second scaling functions to alter an output of the ANN.

45. determining a selection pattern difference between the deletion curve and the missense curve; determining one or more third scaling functions to reduce the selection pattern difference between the deletion curves and the missense curves; 45. The computer-implemented method of claim 44, further comprising modifying the plurality of deletion scores according to the one or more third scaling functions to alter the output of the ANN.

46. the one or more second scaling functions and the one or more third scaling functions are part of the one or more scaling functions; 46. ​​The computer-implemented method of claim 45, wherein the one or more scaling functions comprise functions that scale the proportion of distinct insertions, distinct deletions, and distinct variants in the genomes of the population, respectively, since indels and single nucleotide variants have different variability.

47. 47. The computer-implemented method of claim 46, wherein the one or more scaling functions obtain scaling factors from equivalent variants under natural selection.

48. 1. A computer-implemented method comprising: generating a plurality of missense pathogenicity scores for each variant of the plurality of variants by an artificial neural network (ANN); generating a plurality of indel pathogenicity scores for each indel of the plurality of indels by the ANN; applying the plurality of indel pathogenicity scores and the plurality of missense pathogenicity scores to one or more curve-forming functions; further processing the plurality of indel pathogenicity scores and the plurality of missense pathogenicity scores using the one or more curve-forming functions to generate indel and missense curves; determining a selection pattern difference between the indel curve and the missense curve; determining one or more scaling functions to reduce the selection pattern difference between the indel and missense curves; updating coefficients of the ANN according to the one or more scaling functions; The computer-implemented method, wherein updating the coefficients of the ANN according to the one or more scaling functions comprises modifying the plurality of indel pathogenicity scores according to the one or more scaling functions to provide a recalibrated accuracy of an indel pathogenicity score for each indel of the plurality of indels.

Images