identifying gene sequence expression profiles according to classification features

Abstract

According to sequence features associated with gene expression, classifying gene sequences by receiving gene sequence data, determining a gene sequence feature set, determining a first classification for the gene sequence feature set according to a machine learning model, defining a causal feature set associated with the first classification for the gene sequence according to the machine learning model, altering the causal feature set for the gene sequence resulting in an altered causal feature set, determining a second classification for the altered causal feature set according to the machine learning model, wherein the second classification is different from the first classification, and defining a target feature set, wherein the target features comprise causal features of the altered causal feature set.

Description

BACKGROUND

[0001] The present disclosure relates generally to the detection and identification of gene sequence expression profiles. The present disclosure relates specifically to identifying gene sequence features associated with gene expression.

[0002] Understanding gene expression, also known as the transcriptome, is essential to understanding biological development and disease in organisms. Machine learning (ML) has been used to predict transcriptome profiles using DNA base sequence and / or epigenetic data. DNA base sequence data often includes transcription factor binding sites (TFBS) and / or enhancers. These attributes are believed to contribute to the control of gene expression, and attributes such as DNA base sequence features can be identified from a wide and publicly available pre-existing resource of many species. Current approaches utilize experimental gene expression data and / or prior knowledge of gene expression rule elements. SUMMARY

[0003] The following presents a summary to provide a basic understanding of one or more embodiments of the present disclosure. This summary is not intended to identify key or important elements or to delineate any scope of any scope or claims of specific embodiments. Its sole purpose is to present concepts in a simplified form as a prelude to the more detailed description that is presented later. In one or more embodiments described herein, an apparatus, system, computer- implemented method, device, and / or computer program product is capable of classifying gene sequence data related to complex patterns of gene expression.

[0004] Aspects of the invention disclose methods, systems, and computer readable media associated with classifying a gene sequence by receiving gene sequence data according to sequence features associated with gene expression, determining a set of gene sequence features, determining a first classification of the set of gene sequence features according to a machine learning model, defining a set of causal features for the gene sequence associated with the first classification according to the machine learning model, altering the set of causal features of the gene sequence resulting in an altered set of causal features, determining a second classification of the altered set of causal features according to a machine learning model, wherein the second classification is different from the first classification, and defining a target feature set, wherein the target feature includes a causal feature from the altered set of causal features. BRIEF DESCRIPTION OF DRAWINGS

[0005] The above and other objects, features and advantages of the present disclosure will become more apparent from the following detailed description when taken in conjunction with the accompanying drawings in which like reference characters refer to like parts throughout the several views.

[0006] Figure 1 A schematic diagram of a computing environment in accordance with embodiments of the invention is provided.

[0007] Figure 2A flow diagram depicting an operational sequence in accordance with embodiments of the present application is provided.

[0008] Figure 3 A cloud computing environment in accordance with embodiments of the present application is depicted.

[0009] Figure 4 An abstraction model layer in accordance with embodiments of the present application is depicted. DETAILED DESCRIPTION

[0010] Some embodiments will be described in greater detail with reference to the accompanying drawings, in which embodiments of the disclosure are shown. The disclosure can be implemented in various ways, however, and should not be construed as limited to the embodiments disclosed herein.

[0011] In one embodiment, one or more components of the system can employ hardware and / or software to address problems that are highly technical in nature (e.g., determining a set of features of a genetic sequence, determining a first classification of the set of features of the genetic sequence according to a machine learning model, defining a causal set of features of the genetic sequence according to the machine learning model, producing an altered causal set of features of the genetic sequence by altering the causal set of features of the genetic sequence, determining a second classification of the altered causal set of features according to the machine learning model, wherein the second classification is different from the first classification, and defining a target set of features, etc.). These solutions are not abstract and cannot be performed by a human as a set of mental acts due to, for example, the processing power required to facilitate classification of genetic sequences. Moreover, some of the processes performed can be performed by a special purpose computer that is used to perform a defined task related to classifying genetic sequences. For example, a specialized computer can be used to perform tasks related to classification of genetic sequences, etc.

[0012] Accurate classification of genetic sequences leads to an understanding of the properties of the genetic sequences that are associated with patterns of gene expression. Identifying sequences that are associated with patterns of gene expression over the course of a day-night circadian rhythm enables control and manipulation of such expression patterns through gene editing using tools such as clustered regularly interspaced short palindromic repeats (CRISPR / Cas9). Applications include gene expression therapy and agricultural improvements. The disclosed embodiments enable classification of genetic sequences that are associated with patterns of gene expression.

[0013] In one embodiment, the method utilizes a trained machine learning (ML) model to classify genetic sequences. The method trains the model according to the nature of the classification required. For example, for classification of genetic sequences or genetic promoter sequences that are associated with physiologically rhythmic or non-physiologically rhythmic sequences, the method utilizes labeled data comprising genetic sequences known to express as physiologically rhythmic or non-physiologically rhythmic as training and testing data for developing the ML classification model.

[0014] The method evaluates time series transcriptome data for a set of genes and a set of associated gene promoter sequences. In one embodiment, the method collects the relevant promoter sequence for an input gene as a set of base pairs immediately upstream of the base pair sequence of the gene. For example, the method collects 1500 base pairs upstream of the gene as the promoter sequence for the gene. The transcriptome includes messenger RNA data related to the activity of the genes / gene promoters. The time series transcriptome data provides data related to the changes in messenger RNA of the genes / gene promoters over the observed time period. Changes in the transcriptome over time indicate changes in gene / promoter activity or gene / promoter expression over the observed time period.

[0015] In one embodiment, the transcriptome analysis of individual genes / promoters of the set of genes / promoters occurs every two hours over a total observation period of 48 hours. The gene / promoter sequences used include known and publicly available gene / promoter sequences. Circadian genes exhibit regular periodic changes in expression over a 24 hour period - and concomitant changes in transcriptome data. Non-circadian genes lack this regular periodic change in expression. The analysis produces a training data set of 50,000 genes / promoters, with 25,000 genes / promoters labeled as circadian due to changes in transcriptome data over the observed time period, and another 25,000 genes / promoters labeled as non-circadian based on the time series transcriptome data. The method labels the genes / promoters of the training set according to the expression data observed in the time series transcriptome data. Genes / promoters with time series data including a periodic expression pattern over 24 cycles labeled as circadian and genes / promoters lacking this periodic expression pattern labeled as non-circadian. Similarly, the method can be adapted using time series transcriptome data for other complex expression patterns to classify and label training data sets for those complex expression patterns. Once classified and labeled, the set of training gene sequences does not need to be generated again.

[0016] After using time series transcriptome analysis of available gene sequences to generate a training data set, the method processes each gene of the 50,000 gene training data set. The method produces a set of genetic nucleotide subsequences or k-mers. In one embodiment, the method utilizes k-mers of length 6 nucleotides. Other k-mer lengths can be selected and used, for example 4, 8, 10, 12 or more. For k-mers, the method produces a set of all possible combinations of the nucleotide options of A, T, G, and C (adenine, thymine, guanine, and cytosine). There are a total of 4096 possible combinations for the 4 nucleotide bases in the 6 groups of k-mers.

[0017] For each possible k-mer combination, the method analyzes the training set of genes and determines the number of occurrences of the k-mer in each gene of the training data set. In one embodiment, the analysis produces a matrix indicating the number of occurrences of each k-mer in each gene. For each gene, the matrix entries constitute the features of the gene.

[0018] In one embodiment, the method counts the number of occurrences of a feature across the base pair sequence of the gene and additionally counts the number of occurrences of the feature across the associated gene promoter base pair sequence. The matrix includes the distribution of feature count values for each of the gene and gene promoter individually. For this embodiment, the total number of possible features is doubled to 8192, 4096 possible gene features and 4096 possible gene promoter features.

[0019] In one embodiment, the method counts the number of occurrences of a feature across the combined sequence of the gene and gene promoter. In this embodiment, the matrix includes the feature count value for each of the 4096 possible features.

[0020] In one embodiment, the method reduces the number of features for each gene from the possible 4096 to a lesser number of features, such as 100 features. As an example, the method can use a chi squared test to identify the 100 most significant features from the overall set of features in the matrix.

[0021] In embodiments, the method utilizes a classification algorithm to predict the classification of the labeled data of the training set. Exemplary classification algorithms include logistic regression, random forest, XGBoost, decision tree, K-NN (K nearest neighbor), Gaussian process, LightGBM (Gradient Boosting Method), and SVM (Support Vector Machine). The method splits the training data set using 80% of the data for training and 20% of the data for testing the developed algorithm. In this embodiment, the method utilizes the k nearest neighbor algorithm and achieves 77% accuracy in classifying the labeled training data with a k value of 2. Depending on the desired accuracy in the fitting and prediction of the training data, the method can utilize other k values. The developed model relies only on the k-mer distribution within the training set sequences and does not use experimental data associated with the gene sequences. For example, the trained model classifies a feature set derived from an input data sequence as circadian or non-circadian. The classification dichotomy results from the nature of the training data set. By analogy, labeled training data associated with other complex gene expression patterns produces a model suitable for classifying a feature set from an input sequence as conforming or not conforming to the complex gene expression pattern.

[0022] In practice, the method receives genetic sequence data, processes the sequence data as described to produce a feature set for the sequence, and passes the feature set to a classification model for analysis. The model returns a classification for the feature set and the associated genetic sequence.

[0023] In embodiments, a user interface, such as a graphical user interface (GUI), provides user access to the disclosed method. The method receives genetic sequence data from a user. The user can download or otherwise provide publicly available genomic (and epigenetic, if available) resources for their species of interest, or use a private user-defined dataset. In one embodiment, the method uses application program interfaces (APIs) associated with publicly available genomic databases to provide links to these databases. The provided genetic sequence resources will be in the form of genomic sequences with gene annotations and / or DNA methylation and / or histone modifications, etc.

[0024] The method processes the provided sequence data, analyzes the provided data to count the number of occurrences of each of the 4096 possible k-mer A-G-T-C for nucleotide combinations with a k-mer of 6 bases. In one embodiment, the method utilizes epigenetic data to ignore known heavily methylated transcription factor binding sites (TFBS) from the set of features captured in the feature matrix. Ignoring such sites reduces the number of matrix values and limits the feature matrix to features / attributes associated with sequence differences (associated with expression differences). TFBS serve a utilitarian function for expression, rather than as a gene attribute. The method captures the respective feature counts as a matrix of values associated with each gene analyzed.

[0025] The method provides the feature matrix to a trained ML model for classification. The method can reduce the number of matrix values from the full 4096 to a smaller number, such as 100, before passing the feature set to the ML model for classification. The ML model, such as a k-nearest neighbor model, classifies each input feature set. The method provides an explanation for the classification in the form of the input feature set’s feature vector and the nearest neighbor that resulted in the classification. The method compares the input feature vector to the nearest neighbor feature vector, and this comparison results in identifying members of a candidate causal feature set - those features of the input feature set that are most likely responsible for classifying the input as the final classification assigned to it.

[0026] In embodiments, the method ranks the features of the candidate causal feature set using data from the comparison of the input feature vector to the k nearest neighbor feature vectors.

[0027] In one embodiment, the method selectively evolves the input gene "computer simulation." For each feature in the set of candidate causal features, the method selectively edits the input gene sequence, removing the candidate feature from the sequence and from the set of features of the sequence. The method then classifies the edited feature set. The method classifies the editing features that result in a change in classification (e.g., features that change a sequence from a physiologically rhythmic to a non-physiologically rhythmic) as members of the target feature set. The method compiles the complete set of target features as all candidate causal features that, when edited, result in a change in classification. The complete target feature set provides candidates for actual gene editing to change the gene expression pattern of the original input gene. Selective removal of the candidate target features by a means such as CRISPR / Cas9 should change the expression pattern of the gene as indicated by the change in classification of the evolved sequence that is edited.

[0028] In one embodiment, the final target feature set provides a means to identify gene homologs in related species to the input gene sequence from a first species. As one example, a user of the method can apply the classification results associated with bread wheat, common wheat, a related wheat species such as durum wheat, or a related grain species such as a barley or oat species to a related wheat species such as durum wheat, or a related grain species such as a barley or oat species. As another example, a user can apply the gene expression classification results associated with a first subject's genome to the genomes of other subjects of the same species. The disclosed embodiments assume that a human donor has consented or otherwise decided that their genetic sequence data be used by users of the disclosed methods and systems.

[0029] In an embodiment, the method maintains a set of candidate causal features for each classification of the model. In this embodiment, the method selects features from the set of candidate causal features for the first classification by computer evolution for addition to the input gene sequence identified by the model as a different classification. Similarly, the method selects features from the set of candidate causal features for removal from the input gene sequence identified by the model with the classification by computer simulation evolution.

[0030] In an embodiment, the method begins computer in-silico evolution of the input sequence with the highest ranked candidate causal feature and proceeds from the highest ranked candidate to the lowest ranked candidate. In this embodiment, the method stops computer simulation evolution of candidate causal features after a threshold number of consecutively ranked candidate causal features fail to result in a change in classification; for example, the method stops computer evolution of the input gene sequence with a candidate causal feature after 10 consecutively ranked candidates each fail to result in a change in classification.

[0031] Figure 1A diagram of exemplary network resources associated with practicing the disclosed invention is provided. The invention can be practiced in a processor of any of the disclosed elements that process instruction streams. As shown, networked client devices 110 are wirelessly connected to server subsystem 102. Client device 104 is wirelessly connected to server subsystem 102 via network 114. Client devices 104 and 110 include a genetic sequence classification program (not shown) and sufficient computing resources (processors, memory, network communication hardware) to execute the program. Client devices 104 and 110 act as user interface devices that enable a user to provide input genetic sequence and epigenetic data to the disclosed methods and systems. Client devices 104 and 110 also act as output devices of the disclosed embodiments to provide output data to the user.

[0032] As Figure 1 shown, server subsystem 102 includes server computer 150. Figure 1 A block diagram showing components of server computer 150 within networked computer system 1000 according to embodiments of the invention. It should be appreciated that Figure 1 only one illustration of an implementation is provided and it is not implied to be of any specific architecture. Many modifications are possible to the depicted environments.

[0033] Server computer 150 can include processor(s) 154, memory 158, persistent storage 170, communication unit 152, input / output (I / O) interface(s) 156, and communication fabric 140. Communication fabric 140 provides communications among cache 162, memory 158, persistent storage 170, communication unit 152, and input / output (I / O) interface(s) 156. Communication fabric 140 can be implemented with any architecture designed for passing data and / or control information between processors (such as microprocessors, communications and network processors, etc.), system memory, peripheral devices, and any other hardware components within a system. For example, communication fabric 140 can be implemented with one or more buses.

[0034] Memory 158 and persistent storage 170 are computer readable storage media. In this embodiment, memory 158 includes random access memory (RAM). Generally, memory 158 can include any suitable volatile or non-volatile computer readable storage media. Cache 162 is a fast memory that enhances the performance of processor(s) 154 by very quickly holding recently accessed data and instructions operating on that data.

[0035] Program instructions and data (e.g., the genetic sequence classification program 175) used to practice embodiments of the present application are stored in the persistent storage 170 for execution and / or access by one or more of the respective processor(s) 154 of the server computer 150 via the cache 162. In this embodiment, the persistent storage 170 includes a magnetic hard disk drive. Alternatively, or in addition to the magnetic hard disk drive, the persistent storage 170 can include a solid-state hard drive, semiconductor memory, read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, or any other computer-readable storage media that is capable of storing program instructions or digital information.

[0036] The media used by the persistent storage 170 can also be removable. For example, a removable hard drive can be used for the persistent storage 170. Other examples include optical and magnetic disks, thumb drives, and smart cards, which are inserted into a drive for transfer onto another computer-readable storage medium, also a portion of the persistent storage 170.

[0037] In these examples, the communication unit 152 provides communication with other data processing systems or devices, including the resources of the client computing devices 104 and 110. In these instances, the communication unit 152 includes one or more network interface cards. The communication unit 152 can provide communication through the use of either or both physical and wireless communications links. Software distributions, as well as other programs and data used in implementing the present application, can be downloaded to the persistent storage 170 of the server computer 150 through the communication unit 152.

[0038] The I / O interface(s) 156 allow for input and output of data with other devices that can be connected to the server computer 150. For example, the I / O interface(s) 156 can provide a connection to an external device(s) 190 such as a keyboard, a keypad, a touchscreen, a microphone, a digital camera, and / or some other suitable input device. The external device(s) 190 can also include portable computer-readable storage media, such as, for example, thumb drives, portable optical or magnetic disks, and memory cards. Software and data used to implement embodiments of the present application, for example the genetic sequence classification program 175 on the server computer 150, can be stored on such portable computer-readable storage media, and can be loaded onto the persistent storage 170 via the I / O interface(s) 156. The I / O interface(s) 156 also connect to a display 180.

[0039] The display 180 provides a mechanism to display data to a user, and can be, for example, a computer monitor. The display 180 can also act as a touchscreen, such as the display of a tablet computer.

[0040] Figure 2 A flowchart 200 showing exemplary activities associated with the practice of the present disclosure is provided. After the procedure begins, a user provides a genetic sequence classification program 175 with genetic sequence data acquired from public sources, private sources, or a combination of public and private sources. The input data includes genomic sequence data 214 as well as genetic annotation and DNA methylation and / or histone modification data. The input data can also include epigenetic data such as prior domain knowledge of the genomic sequence, e.g., heavily methylated TFBS sites of the sequence, 218.

[0041] At 220, the method of the genetic sequence classification program 175 processes the input genetic data 214 to produce a sequence feature matrix of the input data. The sequence features include data about the distribution of possible 6-base k-mers in the genomic sequence of the input data 214.

[0042] At 230, the method of the genetic sequence classification program 175 optionally utilizes epigenetic data 218 to reduce the number of entries in the feature matrix from 220. The method removes features associated with known heavily methylated TFBS sites from the matrix or reduces the associated matrix entry values to zero.

[0043] At 240, the method of the genetic sequence classification program 175 classifies or predicts a classification for the input genetic sequence feature set from 220 or modified with epigenetic information from 230. The method utilizes a machine learning model trained to classify genetic sequences using a training data set of labeled genetic sequence data related to the desired classification. For example, a machine learning model trained using labeled genetic sequences associated with each of circadian and non-circadian genetic sequences provides a prediction of circadian or non-circadian for the provided input feature set.

[0044] At 250, the method of the genetic sequence classification program 175 uses the classified classification model explanation to generate a candidate causal feature set. The set includes those sequence features of the input genetic sequence that are most likely to have caused the model classification of the input sequence. In embodiments, the method ranks the members of the candidate feature set from most likely to least likely.

[0045] At 260, the method of the genetic sequence classification program 175 selectively edits the input genetic sequence and associated input sequence feature set from 220 or 230. For each member of the candidate causal feature set, the method removes the feature from the input genetic sequence and associated input sequence feature set.

[0046] At 270, the method of the gene sequence classification program 175 uses the trained machine learning model to make a prediction or classification on the edited input feature set. The method passes its input features that remove the classification change to the target feature set, 280. The method returns to 260 and edits each candidate causal feature, with each iteration, the input sequence and associated feature set is edited by only a single candidate causal feature.

[0047] In an embodiment, the method is a universal candidate causal feature set for each possible classification of the machine learning model. In this embodiment, at 260, the method removes candidate causal features from the input sequence and removes input features from the universal candidate causal feature set for the classification of the input sequence, or adds candidate causal features from the universal candidate causal feature set for a different classification. For example, for an input sequence classified as circadian, the method adds candidate causal features from the universal candidate causal feature set for non-circadian sequences, or removes candidate casual features from the candidate causal feature set of the input sequence and input feature set. In this embodiment, the method refines the target feature set for each possible classification of the machine learning classification model. (Features added from the universal causal feature set that cause a change in classification are added to the associated target feature set for that classification; for example, the method adds features from the universal candidate causal feature set that are added to a circadian sequence that causes the sequence to be reclassified as non-circadian to the target feature set for non-circadian sequences.)

[0048] The method provides the target feature set from 280 to the user via the user interface 210. The user can use the target features to selectively edit actual genetic sequences for genetic therapies associated with changing gene expression patterns, or to change genetic expression of plant species to enhance agricultural production.

[0049] In an embodiment, the execution of the disclosed methods requires computing resources that exceed those available locally to the user. In this embodiment, the user connects to networked resources including edge cloud and cloud resources to enable the timely execution of these methods.

[0050] It should be appreciated that, although the present disclosure includes detailed descriptions regarding cloud computing, implementation of the teachings referenced herein are not limited to cloud computing environments. Rather, embodiments of the present invention are capable of implementation in conjunction with any other type of computing environment now known or later developed.

[0051] Cloud computing is a service delivery model for enabling convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, network bandwidth, servers, processing, storage, applications, virtual machines, and services) that can be rapidly configured and released with minimal management effort or interaction with service providers. This cloud model may include at least five features, at least three service models, and at least four deployment models.

[0052] The features are as follows:

[0053] On-demand self-service: Cloud consumers can automatically and unilaterally provision computing power, such as server time and network storage, on demand, without human interaction with the service provider.

[0054] Extensive network access: Capabilities are available on the network and accessed through standard mechanisms that facilitate the use of heterogeneous thin or thick client platforms (e.g., mobile phones, laptops, and PDAs).

[0055] Resource pooling: A provider's computing resources are pooled to serve multiple consumers using a multi-tenant model, where different physical and virtual resources are dynamically allocated and reallocated as needed. Location independence is significant because consumers typically do not have control or knowledge of the exact location of the resources provided, but can specify the location at a higher level of abstraction (e.g., country, state, or data center).

[0056] Rapid elasticity: Capacity can be provided quickly and flexibly (in some cases, automatically) to shrink rapidly and expand rapidly. For consumers, the capacity available for supply often appears unlimited and can be purchased in any quantity at any time.

[0057] Measuring services: Cloud systems automatically control and optimize resource usage by leveraging metering capabilities at some level of abstraction appropriate to the service type (e.g., storage, processing, bandwidth, and active user accounts). Resource usage can be monitored, controlled, and reported, providing transparency for both the providers and consumers of the services being utilized.

[0058] The business model is as follows:

[0059] Software as a Service (SaaS): The capability offered to consumers is the ability to use the provider's applications running on cloud infrastructure. Applications can be accessed from different client devices through thin client interfaces such as web browsers (e.g., web-based email). Consumers do not manage or control the underlying cloud infrastructure, including the network, servers, operating system, storage, or even the individual application capabilities, with possible exceptions of limited user-specific application configuration settings.

[0060] Platform as a Service (PaaS): the capability provided to the consumer is to deploy onto the cloud infrastructure consumer-created or acquired applications created using programming languages and tools supported by the provider. The consumer does not manage or control the underlying cloud infrastructure including networks, servers, operating systems, or storage, but has control over the deployed applications and possibly application hosting environment configurations.

[0061] Infrastructure as a Service (laaS): the capability provided to the consumer is to provision processing, storage, networks, and other fundamental computing resources where the consumer is able to deploy and run arbitrary software, which can include operating systems and applications. The consumer does not manage or control the underlying cloud infrastructure but has control over operating systems, storage, deployed applications, and possibly limited control of select networking components (e.g., host firewalls).

[0062] Deployment models are as follows:

[0063] Private cloud: the cloud infrastructure is operated solely for an organization. It can be managed by the organization or a third party and can exist on-premises or off-premises.

[0064] Community cloud: the cloud infrastructure is shared by several organizations and supports mission-oriented business

[0065] Public cloud: the cloud infrastructure is made available to the general public or a large industry group and is owned by an organization selling cloud services.

[0066] Hybrid cloud: the cloud infrastructure is a composition of two or more clouds (private, community, or public) that remain unique entities but are bound together, creating the hybrid cloud.

[0067] A cloud computing environment is service-oriented, focusing on statelessness, loose coupling, modularity, and semantic interoperability. The core of cloud computing is an infrastructure that includes a network of interconnected nodes.

[0068] Referring now to the drawings Figure 3, an illustrative cloud computing environment 50 is depicted. As shown, cloud computing environment 50 includes one or more cloud computing nodes 10 with which a cloud consumer can engage with to use the inventive embodiments. Cloud computing nodes 10 can communicate with one another. They can be Figure 3 The types of computing devices 54A-N shown are intended to be illustrative only and computing nodes 10 and cloud computing environment 50 can communicate with any type of computerized devices over any type of network and / or network addressable connection (e.g., using a web browser).

[0069] Referring now to Figure 4 , a set of functional abstraction layers provided by cloud computing environment 50 (( Figure 3 ) is shown. It should be understood that Figure 4 The components, layers, and functions shown in FIG. 6 are intended to be illustrative only and embodiments of the present application are not limited in their application to this. As depicted, the following layers and corresponding functionality are provided:

[0070] Hardware and software layer 60 includes hardware and software components. Examples of hardware components include: mainframes 61; RISC (Reduced Instruction Set Computer) architecture based servers 62; servers 63; blade servers 64; storage devices 65; and networks and networking components 66. In some embodiments, software components include network application server software 67 and database software 68.

[0071] Virtualization layer 70 provides an abstraction layer from which the following examples of virtual entities can be provided: virtual servers 71; virtual storage 72; virtual networks 73, including virtual private networks; virtual applications and operating systems 74; and virtual clients 75.

[0072] In one example, management layer 80 can provide the functions described below. Resource provisioning 81 provides dynamic procurement of computing resources and other resources that are utilized to perform tasks within the cloud computing environment. Metering and Pricing 82 provide cost tracking as resources are utilized within the cloud computing environment, and billing or invoicing for consumption of these resources. In one example, these resources can include application software licenses. Security provides identity verification and protection for the cloud consumers and tasks that utilize the cloud computing environment. User portal 83 provides access to the cloud computing environment for consumers and system administrators. Service level management 84 provides cloud computing resource allocation and management such that required service levels are met. Service Level Agreement (SLA) planning and fulfillment 85 provide pre-arrangement for, and procurement of, cloud computing resources for which future requirements are anticipated in accordance with an SLA.

[0073] Workloads layer 90 provides examples of functionality for which the cloud computing environment can be utilized. Examples of workloads and functions which can be provided from this layer include: mapping and navigation 91; software development and lifecycle management 92; virtual classroom education delivery 93; data analytics processing 94; transaction processing 95; and gene sequence classification programs 175.

[0074] The present application can be a system, a method, and / or a computer program product at any possible technical detail level of integration. The application can advantageously implement in any system (or combination thereof) that processes a stream of instructions. The computer program product can include a computer readable storage medium (or media) having computer readable program instructions thereon for causing a processor to carry out aspects of the present application.

[0075] The computer readable storage medium can be a tangible device that can retain and store instructions for use by an instruction execution device. The computer readable storage medium can be, for example, but is not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. A non-exhaustive list of more specific examples of the computer readable storage medium includes the following: a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanically encoded device such as punch-cards or raised structures in a groove having instructions recorded thereon, and any suitable combination of the foregoing. A computer readable storage medium, as used herein, is not to be construed as being transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission media (e.g., light pulses passing through a fiber-optic cable), or electrical signals transmitted through a wire.

[0076] Computer readable program instructions described herein can be downloaded to respective computing / processing devices from a computer readable storage medium or to an external computer or external storage device via a network, for example, the Internet, a local area network, a wide area network and / or a wireless network. The network can comprise copper transmission cables, optical transmission fibers, wireless transmission, routers, firewalls, switches, gateway computers and / or edge servers. A network adapter card or network interface in each computing / processing device receives computer readable program instructions from the network and forwards the computer readable program instructions to a computer readable storage medium within the respective computing / processing device for storage and / or execution.

[0077] Computer readable program instructions for carrying out operations of the present application can be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state setting data, configuration data for an integrated circuit, or source code or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, C++ or the like, and procedural programming languages, such as the "C" programming language or similar programming languages. The computer readable program instructions can execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider). In some embodiments, electronic circuitry including, for example, programmable logic circuitry, field-programmable gate arrays (FPGA), or programmable logic arrays (PLA) can execute the computer readable program instructions by utilizing state information of the computer readable program instructions to personalize the electronic circuitry, in order to perform aspects of the present application.

[0078] The computer readable program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other device to produce a computer implemented process, such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0079] These computer readable program instructions can be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks. These computer readable program instructions can also be stored in a computer readable storage medium that can direct a computer, a programmable data processing apparatus, and / or other devices to function in a particular manner, such that the computer readable storage medium having instructions stored therein comprises an article of manufacture including

[0080] The computer readable program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other device to produce a computer implemented process, such that the instructions which execute on the computer, other programmable apparatus, or other device implement the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0081] The flow diagrams and the block diagrams in the drawings are meant as possible implementations of systems, methods, and computer program products according to the application. It will be understood that each block of the flowchart and / or block diagram, and combinations of blocks in the flowchart and / or block diagram, can be implemented by special purpose hardware-based systems that perform the specified functions or acts or combinations of special purpose hardware and computer instructions.

[0082] Reference throughout this specification to "one embodiment", "an embodiment", "exemplary embodiment", etc., means that a particular feature, structure, or characteristic described in connection with the embodiment is included in at least one embodiment. The appearances of the phrase "in one embodiment" in various places in the specification are not necessarily all referring to the same embodiment.

[0083] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the present application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0084] The description of the different embodiments of the application has been presented for purposes of illustration and is not intended to be exhaustive or limited to the embodiments disclosed. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the application. The terminology used herein was chosen to best explain the principles of the embodiment, the practical application or technical improvement over technology found in the marketplace, or to enable others skilled in the art to understand the embodiment disclosed in the present disclosure.

Claims

1. A computer-implemented method for classifying a gene sequence according to sequence features associated with gene expression, the method comprising: receiving, by one or more computer processors, gene sequence data; determining, by the one or more computer processors, a set of k-mer based gene sequence features; determining, by the one or more computer processors, a first classification for the set of k-mer based gene sequence features according to a machine learning model; defining, by the one or more computer processors, a causal feature set associated with the first classification for the gene sequence according to the machine learning model; altering, by the one or more computer processors, the causal feature set for the gene sequence resulting in an altered causal feature set; determining, by the one or more computer processors, a second classification for the altered causal feature set according to the machine learning model, wherein the second classification is different from the first classification; and defining, by the one or more computer processors, a target feature set, wherein the target features comprise causal features from the altered causal feature set, wherein the altering, by the one or more computer processors, of the causal feature set for the gene sequence resulting in an altered causal feature set comprises, for each feature in the causal feature set, selectively editing the gene sequence, removing a candidate feature from the gene sequence and the causal feature set; wherein the determining, by the one or more computer processors, of the second classification for the altered causal feature set comprises classifying the edited causal feature set; and wherein the defining, by the one or more computer processors, of the target feature set comprises classifying the edited features in the candidate features that result in a change in classification as members of the target feature set.

2. The computer-implemented method of claim 1, wherein determining the set of k-mer based gene sequence features comprises determining the set of k-mer based gene sequence features from epigenetic data.

3. The computer-implemented method of claim 1, wherein determining the set of k-mer based gene sequence features comprises: defining a set of possible k-mer based gene sequence features; and determining a distribution of each possible gene feature within the gene sequence.

4. The computer-implemented method of claim 1, wherein determining the first classification for the set of k-mer based gene sequence features according to a machine learning model comprises determining a circadian / non-circadian classification for the gene sequence.

5. The computer-implemented method of claim 1, further comprising identifying, by the one or more computer processors, a gene homolog for the gene sequence in a related species according to the target feature set.

6. The computer-implemented method of claim 1, further comprising identifying, by the one or more computer processors, editing candidates within the gene sequence associated with altering expression of the gene sequence according to the target feature set.

7. The computer-implemented method of claim 1, further comprising ranking the target feature set according to gene sequence expression prediction.

8. A computer program product for classifying a gene sequence according to gene sequence features associated with gene expression, the computer program product comprising program instructions, the program instructions comprising: program instructions for receiving gene sequence data; program instructions for determining a set of k-mer-based gene sequence features; program instructions for determining a first classification for the set of k-mer-based gene sequence features according to a machine learning model; program instructions for defining a causal feature set associated with the first classification for the gene sequence according to the machine learning model; program instructions for altering the causal feature set to produce an altered causal feature set; program instructions for determining a second classification for the altered causal feature set according to the machine learning model, wherein the second classification is different from the first classification; and program instructions for defining a target feature set, wherein the target features comprise causal features from the altered causal feature set, wherein the altering the causal feature set to produce an altered causal feature set comprises, for each feature in the causal feature set, selectively editing the gene sequence, removing a candidate feature from the gene sequence and the causal feature set; wherein the determining a second classification for the altered causal feature set according to the machine learning model comprises classifying the edited causal feature set; and wherein the defining a target feature set comprises classifying the candidate features that result in a change in classification as members of the target feature set.

9. The computer program product of claim 8, wherein the program instructions for determining the set of k-mer-based gene sequence features comprise program instructions for determining the set of k-mer-based gene sequence features according to epigenetic data.

10. The computer program product of claim 8, wherein the program instructions for determining the set of k-mer-based gene sequence features comprise: program instructions for defining a set of possible k-mer-based gene sequence features; and program instructions for determining a distribution of each possible gene feature within the gene sequence.

11. The computer program product of claim 8, wherein the program instructions for determining a first classification for the set of k-mer-based gene sequence features according to a machine learning model comprise program instructions for determining a circadian / non-circadian classification for the gene sequence.

12. The computer program product of claim 8, the program instructions further comprising program instructions for identifying gene homologs for the gene sequence in related species according to the target feature set. ​ 13. The computer program product of claim 8, the program instructions further comprising program instructions for identifying candidate editing sites within the genetic sequence according to the target feature set, the candidate editing sites being associated with altering expression of the genetic sequence.

14. The computer program product of claim 8, the program instructions further comprising program instructions for ranking the target feature set according to genetic sequence expression prediction.

15. A computer system for classifying genetic sequences according to genetic sequence features associated with genetic expression, the computer system comprising: one or more computer processors; one or more computer-readable storage devices; and program instructions stored on the one or more computer-readable storage devices for execution by the one or more computer processors, the stored program instructions comprising: program instructions for receiving genetic sequence data; program instructions for determining a set of k-mer-based genetic sequence features; program instructions for determining a first classification for the set of k-mer-based genetic sequence features according to a machine learning model; program instructions for defining a causal feature set associated with the first classification for the genetic sequence according to the machine learning model; program instructions for altering the causal feature set for the genetic sequence to produce an altered causal feature set; program instructions for determining a second classification for the altered causal feature set according to the machine learning model, wherein the second classification is different from the first classification; and program instructions for defining a target feature set, wherein the target feature comprises a causal feature from the altered causal feature set, wherein the altering the causal feature set for the genetic sequence to produce an altered causal feature set comprises, for each feature in the causal feature set, selectively editing the genetic sequence, removing a candidate feature from the genetic sequence and the causal feature set; wherein the determining a second classification for the altered causal feature set according to the machine learning model comprises classifying the edited causal feature set; and wherein the defining a target feature set comprises classifying the candidate features that result in a change in classification as members of the target feature set.

16. The computer system of claim 15, wherein the program instructions for determining the set of k-mer-based genetic sequence features comprise program instructions for determining the set of k-mer-based genetic sequence features according to epigenetic data.

17. The computer system of claim 15, wherein the program instructions for determining the set of k-mer-based genetic sequence features comprise: program instructions for defining a set of possible k-mer-based genetic sequence features; and program instructions for determining a distribution of each possible genetic feature within the genetic sequence. ​ ​ 18. The computer system of claim 15, wherein the program instructions for determining, from a machine learning model, a first classification for the set of k-mer based genetic sequence features comprises program instructions for determining a circadian / circadian-like classification for the genetic sequence.

19. The computer system of claim 15, the stored program instructions further comprising program instructions for identifying, from the target feature set, a genetic homolog for the genetic sequence in a related species.

20. The computer system of claim 15, the stored program instructions further comprising program instructions for identifying, from the target feature set, a candidate editing site within the genetic sequence, the candidate editing site being associated with altering expression of the genetic sequence.

21. A computer system comprising modules for performing the steps of the method of any one of claims 1-7, respectively.

Images