Identification of one or more compounds for targeting a gene

Abstract

A computer-implemented method for identifying tool compounds is provided. The method includes: searching a database for first candidate compounds each targeting one or more first target genes; generating a first fingerprint for each first candidate compound by searching the database for genes associated with the first candidate compound and predicting genes associated with the first candidate compound; and filtering the first candidate compounds using the first fingerprint to identify a first optimized compound targeting the one or more first target genes.

Description

[0001] The present application relates to systems and methods for identifying tool compounds. Tool compounds are compounds that can be used to target a gene to test whether the gene is associated with a disease under study. Background Art

[0002] In the field of drug discovery, a disease target hypothesis is a hypothesis that a disease is associated with a target gene. To test the disease target hypothesis, drug discovery scientists are interested in understanding which are the optimal tool compounds that can be used to target the gene.

[0003] However, the process of identifying the most effective and commercially viable tool compounds for testing genes is time-consuming, introducing significant delays and costs to drug discovery programs.

[0004] A technology that more efficiently identifies tool compounds for targeting genes is needed to help enable rapid, high-volume validation of disease target hypotheses.

[0005] The embodiments described below are not limited to implementations that solve any or all disadvantages of the known methods described above. Summary of the invention

[0006] The purpose of providing this summary is to introduce a selection of concepts in a simplified form that are further described below in the detailed description. This summary is not intended to identify key features or essential features of the claimed subject matter, nor is it intended to be used to determine the scope of the claimed subject matter.

[0007] In a first aspect, the present disclosure provides a computer-implemented method for identifying tool compounds, the method comprising: searching a database for first candidate compounds each of which targets one or more first target genes; generating a first fingerprint for each first candidate compound by: searching the database for genes associated with the first candidate compound and predicting genes associated with the first candidate compound; and filtering the first candidate compounds using the first fingerprint to identify a first optimized compound targeting the one or more first target genes.

[0008] Optionally, predicting genes associated with the first candidate compound may include using a trained machine learning model to predict gene interaction profiles with a range of compounds.

[0009] Optionally, the model may include a neural network.

[0010] Optionally, the method may include predicting genes associated with the first candidate compound only when no association data is available in the database.

[0011] Optionally, filtering the first candidate compounds may include comparing each first fingerprint to an ideal fingerprint of a theoretical tool compound.

[0012] Optionally, the comparison may include calculating a similarity score.

[0013] Optionally, the method includes identifying a first candidate compound that is most similar to the theoretical tool compound as a first optimized compound.

[0014] Optionally, filtering the first candidate compounds may include generating a metric using the first fingerprint and filtering the first candidate compounds using the metric.

[0015] Optionally, generating the first fingerprint may include obtaining metadata about the one or more first candidate compounds.

[0016] Optionally, metadata may include clinical trial phase data, drug name or identity, or information from a compound supplier.

[0017] Optionally, the method may include using the library evaluation framework to retrieve an indication of how many targets each first candidate compound has.

[0018] Optionally, the method may include: searching a database for second candidate compounds each targeting one or more second target genes; generating a second fingerprint for each second candidate compound by: searching a database for genes associated with the second candidate compound and predicting genes associated with the second candidate compound; and filtering a group containing the first candidate compound and the second candidate compound using the first fingerprint and the second fingerprint to identify the first optimized compound and identify the second optimized compound targeting the one or more second target genes.

[0019] In a second aspect, the present disclosure provides a system for identifying tool compounds, the system comprising: a compound search module, which is configured to search a database for first candidate compounds that each target one or more first target genes; a fingerprint module, which is configured to generate a first fingerprint for each first candidate compound, the fingerprint module comprising: a gene search module, which is configured to search a database for genes associated with the first candidate compound, and a prediction module, which is configured to predict genes associated with the first candidate compound; and a filter module, which is configured to filter the first candidate compounds using the first fingerprint to identify a first optimized compound for targeting the one or more first target genes.

[0020] Optionally, the prediction module can be configured to use a model trained to predict gene interaction profiles with a range of compounds.

[0021] Optionally, the model may include a neural network.

[0022] Optionally, the prediction module may be configured to predict genes associated with the first candidate compound only when no associated data is available in the database.

[0023] Optionally, the filter module may be configured to filter the first candidate compounds by comparing each first fingerprint to an ideal fingerprint of a theoretical tool compound.

[0024] Optionally, the comparison may include calculating a similarity score.

[0025] Optionally, the filter module can be configured to identify one or more first candidate compounds that are most similar to an ideal tool compound.

[0026] Optionally, the filter module can be configured to select a first candidate compound that is most similar to the ideal tool compound as the first optimized compound.

[0027] Optionally, the fingerprint module may be configured to obtain metadata about the one or more first candidate compounds.

[0028] Optionally, metadata may include clinical trial phase data, drug name or identity, or information from a compound supplier.

[0029] Optionally, the fingerprint module can be configured to use the library evaluation framework to retrieve an indication of how many targets each first candidate compound has.

[0030] Optionally, the compound search module can be used to search the database for second candidate compounds each targeting one or more second target genes; the fingerprint module can be configured to generate a second fingerprint for each second candidate compound; the gene search module can be configured to search the database for genes associated with the second candidate compounds; the prediction module can be configured to predict genes associated with the second candidate compounds; and the filter module can be configured to use the first fingerprint and the second fingerprint to filter the group containing the first candidate compound and the second candidate compound to identify the first optimized compound and identify the second optimized compound for targeting the one or more second target genes.

[0031] In a third aspect, the present disclosure provides a computer readable medium storing code, which, when executed by a computer, causes the computer to perform the method of the first aspect.

[0032] The methods described herein may be performed by software in machine-readable form on a tangible storage medium, such as in the form of a computer program, the computer program comprising computer program code means adapted to perform all steps of any method described herein when the program is run on a computer and when the computer program may be embodied on a computer-readable medium. Examples of tangible (or non-transitory) storage media include disks, thumb drives, memory cards, etc., and do not include propagating signals. The software may be adapted to be executed on a parallel processor or a serial processor, such that the method steps may be performed in any suitable order or simultaneously.

[0033] This application recognizes that firmware and software can be valuable, separately tradable commodities. It is intended to cover software that runs on or controls "dumb" or standard hardware to achieve desired functions. It is also intended to cover software that "describes" or defines the configuration of hardware, such as HDL (Hardware Description Language) software, such as used to design silicon chips or for configuring general-purpose configurable chips to achieve desired functions.

[0034] As will be apparent to the skilled person, the preferred features may be combined as appropriate and may be combined with any aspect of the invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Embodiments of the invention will now be described by way of example and with reference to the accompanying drawings, in which:

[0036] Figure 1 is a schematic diagram showing an embodiment of the present invention for identifying optimized tool compounds targeting a target gene;

[0037] Figure 2 is a flow chart of a method for identifying an optimized tool compound that targets a target gene;

[0038] Figure 3 is a schematic diagram of the polypharmacological fingerprint of the candidate compound;

[0039] Figure 4 is a schematic diagram representing an embodiment of the present invention for identifying optimized tool compounds for targeted gene sets;

[0040] Figure 5 is a schematic diagram representing an embodiment of the present invention for identifying corresponding optimized tool compounds targeting corresponding target genes;

[0041] Figure 6 is a schematic diagram representing an embodiment of the present invention for identifying corresponding optimized tool compounds targeting corresponding gene sets;

[0042] Figure 7 is a block diagram of a system according to an embodiment of the present invention; and

[0043] Figure 8 is a block diagram of a computer suitable for implementing embodiments of the present invention.

[0044] Common reference numbers are used throughout the drawings to indicate similar features. DETAILED DESCRIPTION

[0045] Embodiments of the present invention are described below by way of example only. These examples represent the best methods of practicing the present invention currently known to applicants, although these examples are not the only methods that can implement the present invention. This description sets forth the functions of the examples and the order of steps for constructing and operating the examples. However, the same or equivalent functions and sequences can be accomplished by different examples.

[0046] The present invention provides an automated way to generate candidate compounds for targeting genes and filter the candidate compounds to identify an optimized tool compound or a shortlist of optimized tool compounds. This enables drug discovery scientists to quickly identify one or more optimized tool compounds for targeting genes to test disease target hypotheses.

[0047] refer to Figure 1 and Figure 2 , a drug discovery scientist might want to identify tool compounds for targeting gene G100. Figure 2 As shown, a method 200 for identifying a tool compound according to an embodiment of the present invention includes searching 202 a database for a candidate compound 102 that targets a gene G 100. Figure 1 As shown, this search generated n candidate compounds C1-C n 102, which may be suitable for targeting gene G 100. In the present disclosure, the term 'database' is used to refer to one or more databases. Each of the one or more databases may include a distributed database. Searching the one or more databases may include using an application program interface (API) to search.

[0048] The database may include a compound database that can be searched to find compounds associated with gene G. In a suitable example, the compound database may store structured data from a series of public sources, including but not limited to chemical databases, patents, and compound pairs and predictions between target genes. The database may additionally or alternatively include unstructured data such as patents or articles, and may also include processed unstructured data. Therefore, the association extracted from the database is usually experimentally verified. Any compound identified as being associated with gene G in the search is a candidate compound for targeting gene G.

[0049] This is followed by an analysis phase, in which candidate compounds C1-C n102 is characterized and filtered to identify one or more tool compounds with optimized characteristics for targeting gene G 100. In order to characterize candidate compounds C1-C n 102, generate 204 fingerprints 104 for each, which allows the candidate compounds C1-C n The manner in which 102 can be evaluated describes the features and properties of the corresponding candidate compound.

[0050] Identify candidate compounds C1-C n One useful factor in the suitability of candidate compounds 102 is which genes they are associated with. As a result, the fingerprint 104 for each candidate compound 102 includes a polypharmacological fingerprint that describes which genes the candidate compound 102 is associated with. For example, the polypharmacological fingerprint 300 of the candidate compound 102 is shown in Figure 3 For a series of genes G1-G m For each of 304, the polypharmacology fingerprint 300 includes an indication of whether each corresponding gene 304 is associated with the candidate compound 102. Figure 3 In the example of , a representation of the degree of association 302 is provided, which can, for example, represent the degree of upregulation of each corresponding gene 304 by the candidate compound 102. In other examples, the polypharmacological fingerprint can also display the genes inhibited by the candidate compound 102 and the degree to which they are inhibited. In any case, the polypharmacological fingerprint describes the activity of the compound relative to a preferably large number of genes and can be used to filter the candidate compounds 102 to find a single optimized compound or multiple optimized compounds for targeting gene G.

[0051] In order to establish a polypharmacological fingerprint of candidate compound 102, data related to candidate compound 102 and a series of genes are required. Genes associated with candidate compound 102 are identified in two ways, such as by upregulating or inhibiting it.

[0052] First, the above database is searched for genes associated with candidate compound 102. Data related to the nature of the associations, such as whether they are upregulated or repressed and by how much, can be retrieved in this search.

[0053] In addition, metadata about the compound from the supplier can be extracted in this search. The example of metadata from the supplier includes the real-time availability of inventory and price information. Metadata from other suppliers or other sources can include the clinical trial stage of the molecule, and the name of the molecule (if it is a drug) (such as celecoxib). Additionally or alternatively, suitable tools such as library evaluation frameworks can be used to retrieve information related to how many targets have been identified with a group of candidate compounds. This tool provides a fast, simple and explainable method, which can be used to quantitatively evaluate it before purchasing or using the library in biological experiments. This is advantageous because when purchased from a supplier, the available information of the molecular mass provided as a part of the library is usually limited.

[0054] Second, to make the polypharmacology fingerprint more extensive, or if no association data for the candidate compound 102 is available, the model is used to predict which genes are associated with the candidate compound 102. The association data includes experimental data, such as experimental data from biological assays, which are reported in the literature and can be retrieved from databases. The association data represents the association between the candidate compound 102 and the gene, and may, for example, include binding data of the candidate compound 102 to the target gene, or alternatively may include drug metabolism and pharmacokinetics (DMPK) and / or absorption, distribution, metabolism and excretion (ADMET) properties of the candidate compound 102, such as solubility, metabolic stability, etc.

[0055] Associations can be predicted using a trained machine learning algorithm such as a neural network or any other suitable machine learning model. The choice of machine learning model may be affected by the size of the dataset available for training. For example, for large datasets, a random forest algorithm may be appropriate, while for small datasets, a transfer learning algorithm may be preferred.

[0056] Any data source describing the interaction between genes and compounds can be used. In suitable instances, the machine learning model predicts the association between compounds and genes based on the known association between the same compound and similar or related genes (e.g., genes with similar binding sites). In other suitable instances, the machine learning model uses three-dimensional interaction data to predict the interaction between compounds and gene binding sites.] The three-dimensional interaction data may include data related to the conformation of a molecule or compound in three spatial dimensions, or may include data related to the structure of at least a portion of a gene in three spatial dimensions. By virtue of the prediction, the machine learning model determines which compounds and genes are associated with each other.

[0057] This process is repeated to generate each candidate compound C1-C n 102 fingerprints.

[0058] Once candidate compound C1-C n 102 A full set of fingerprints was generated and the fingerprints were used to filter 206 candidate compounds C1-C n 102 to obtain a list of optimization tool compounds or a single optimization tool compound 106 for targeting gene G 100.

[0059] There are a variety of methods for filtering candidate compounds 102. The fingerprints can be compared to the ideal fingerprint of a theoretical tool compound to identify the fingerprint that is most similar to the ideal fingerprint. The comparison can include calculating a similarity score between each fingerprint and the ideal fingerprint of the theoretical tool compound. The candidate compound with the highest similarity score is selected as the optimized tool compound, or, if multiple tool compounds are needed, the candidate compound with the highest similarity score is selected as the tool compound.

[0060] Alternatively, a metric can be generated from the fingerprint and used to filter candidate compounds. For example, the metric can include, but is not limited to, default scoring metrics, such as those related to: physical or chemical properties, such as molar weight (MW), logarithm of the partition coefficient (logP), number of hydrogen bond acceptors or donors, etc., or enzyme activity, such as the value of the half-maximal inhibitory concentration (IC50) of the molecule in the assay or the half-maximal effective concentration (EC50) of the molecule, the selectivity of the compound to the target gene, the number of off-targets (i.e., other unwanted genes affected by the compound), the effectiveness of the compound to the gene, solubility, cell data indicating the activity of the compound in the cell assay, and commercial availability. The metric used can be user-selected and can be weighted by the user in terms of importance additionally or alternatively. A combination of metrics can be used to generate a total score for each candidate compound.

[0061] Other methods may include a combination of filtering candidate compounds by comparing fingerprints to ideal fingerprints and filtering candidate compounds by generating metrics from the fingerprints.

[0062] The present invention can be used to identify tool compounds that are different from each other. If two compounds are identified that target the same gene but have different off-targets, this can be used to increase the confidence that the target gene is associated with the disease treatment mechanism if the two compounds have a beneficial effect in treating the disease.

[0063] In the above embodiments, the present invention is used to find one or more optimized tool compounds targeting a single gene. However, in some cases, drug discovery scientists may wish to find a single compound targeting multiple genes, such as for the effective treatment of diseases with more complex disease mechanisms. In this case, alternative embodiments may be used to find one or more optimized compounds for targeting gene sets.

[0064] refer to Figure 4, a gene set G 400 containing multiple genes is used to search the database for compounds associated with one or more genes of the gene set G 400. The compounds 402 returned in the search are candidate compounds 402 for targeting the genes of the gene set G 400, and all members of the gene set G 400 can be targeted simultaneously. A fingerprint 404 is generated for each candidate compound 402 and used to filter the candidate compounds 402. In this case, the ideal fingerprint of the theoretical compound will describe the ideal interaction of the tool compound with all genes in the gene set G 400. This enables the identification of one or more optimized tool compounds 406 for targeting the genes of the gene set G 400.

[0065] There may be something drug discovery scientists hope to use more than once Figure 1 For example, a drug discovery scientist may wish to identify a first optimization tool compound for targeting a first target gene and a second optimization tool compound for targeting a second target gene. In this case, Figure 1 Embodiments of the invention can be run in parallel to simultaneously identify corresponding optimization tool compounds.

[0066] In one example of this approach, tool compounds are needed to target multiple genes G1 500, G2 502, G3 504, and G m Each of the 506. Reference Figure 5 , the database was searched to identify genes G1 500, G2 502, G3 504, and G m One or more related compounds in 506 are identified in the search. Compounds 508 identified in the search are candidate compounds 508 for targeting the corresponding gene. Fingerprints 510 are generated for each candidate compound 508 and used to filter candidate compounds 508. This enables the generation of fingerprints for genes G1 500, G2 502, G3 504, and G m 506 to identify a corresponding optimized tool compound 512, 514, 516. If multiple tool compounds are required for each gene, this method can also be used to identify the corresponding optimized tool compounds 512, 514, 516 for genes G1 500, G2 502, G3 504, and G4 505. m Each of 506 identifies a corresponding plurality of optimization tool compounds.

[0067] Similarly, there may be drugs that drug discovery scientists wish to use more than once. Figure 4 For example, a drug discovery scientist may wish to identify a first optimization tool compound for targeting a first gene set and a second optimization tool compound for targeting a second gene set. In this case, Figure 4Embodiments of the invention can be run in parallel to simultaneously identify corresponding optimization tool compounds.

[0068] In one example of this approach, tool compounds are needed to target multiple gene sets G1 600, G2 602, G3 604, and G m Each of 606. Reference Figure 6 , the database was searched to identify genes corresponding to gene sets G1 600, G2 602, G3 604, and G m Compounds 608 identified in the search are compounds that target the corresponding gene sets G1 600, G2 602, G3 604, and G m 606 candidate compounds 608. A fingerprint 610 is generated for each candidate compound 608 and used to filter the candidate compounds 608. This enables the generation of a fingerprint for gene sets G1 600, G2 602, G3 604, and G m 606 to identify a corresponding optimized tool compound 612, 614, 616. If multiple tool compounds are required for each gene set, this method can also be used to identify tool compounds for gene sets G1 600, G2 602, G3 604, and G m Each of 606 identifies a corresponding plurality of optimization tool compounds.

[0069] The system 700 for identifying tool compounds according to the present invention is Figure 7 . The system includes a compound search module 702, which is configured to search for candidate compounds that each target one or more target genes in a database 704. The system 700 also includes a fingerprint module 706, which is configured to generate a fingerprint for each candidate compound. The fingerprint module 706 includes a gene search module 708 configured to search for genes associated with each candidate compound in the database 704 and a prediction module 710 configured to predict genes associated with each candidate compound. Finally, the system 700 also includes a filter module 712, which is configured to filter candidate compounds using fingerprints to identify optimized compounds for targeting one or more target genes.

[0070] A computer device 800 suitable for implementing the method according to the present invention is as follows: Figure 8 The apparatus 800 includes a processor 802, an input-output device 804, a communication port 806, and a computer memory 808. For example, the memory 808 may store instructions that, when executed by the processor 802, cause the apparatus 800 to execute Figure 2 The code of method 200 is shown.

[0071] In the embodiment described above, the server can include a single server or a server network. In some instances, the functionality of the server can be provided by a server network (such as a server global distributed network) distributed across geographic regions, and the user can be connected to a suitable one in the server network based on the user's location.

[0072] For clarity, the above description discusses embodiments of the present invention with reference to a single user. It should be understood that, in practice, the system may be shared by multiple users and potentially by a very large number of users simultaneously.

[0073] The embodiments described above are fully automatic. In some instances, a user or operator of the system may manually direct the execution of some steps of the method.

[0074] In the described embodiments of the present invention, the system may be implemented as any form of computing and / or electronic device. Such a device may include one or more processors, which may be a microprocessor, a controller, or any other suitable type of processor for processing computer executable instructions to control the operation of the device to collect and record routing information. In some instances, such as when using a system-on-chip architecture, the processor may include one or more fixed functional blocks (also referred to as accelerators) that implement a portion of the method in hardware (rather than software or firmware). Platform software including an operating system or any other suitable platform software may be provided at a computing-based device to enable application software to be executed on the device.

[0075] The various functions described herein can be implemented in hardware, software or any combination thereof. If implemented in software, the function can be stored as one or more instructions or codes on a computer-readable medium or transmitted by a computer-readable medium. Computer-readable media may include, for example, computer-readable storage media. Computer-readable storage media may include volatile or non-volatile media, removable or non-removable media for storing information such as computer-readable instructions, data structures, program modules, or other data implemented in any method or technology. Computer-readable storage media can be any available storage medium that can be accessed by a computer. By way of example and not limitation, such computer-readable storage media may include RAM, ROM, EEPROM, flash memory or other memory devices, CD-ROM or other optical disk storage devices, magnetic disk storage devices or other magnetic storage devices or any other medium that can be used to carry or store the desired program code in the form of an instruction or data structure and can be accessed by a computer. As used herein, disks and discs include compact discs (CDs), laser discs, optical discs, digital versatile discs (DVDs), floppy disks and blue-ray discs (BDs). Further, the propagated signal is not included in the scope of computer-readable storage media. Computer-readable media also include communication media, which include any media that facilitate the transfer of a computer program from one place to another. For example, a connection can be a communication medium. For example, if a coaxial cable, optical cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwaves are used to transmit software from a website, server, or other remote source, then the coaxial cable, optical cable, twisted pair, DSL, or wireless technologies such as infrared, radio, and microwaves are included in the definition of communication media. Combinations of the foregoing should also be included within the scope of computer-readable media.

[0076] Alternatively or additionally, the functionality described herein may be at least partially performed by one or more hardware logic components. For example, and not limitation, the hardware logic components that may be used may include a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), an application specific standard product (ASSP), a system on a chip (SOC), a complex programmable logic device (CPLD), etc.

[0077] Although described as a single system, it should be understood that the computing device may be a distributed system. Thus, for example, several devices may communicate via a network connection and may jointly perform tasks described as being performed by the computing device.

[0078] Although shown as a local device, it will be appreciated that the computing device may be located remotely and accessed over a network or other communications link (eg, using a communications interface).

[0079] The term "computer" is used herein to refer to any device having processing capabilities so that it can execute instructions. Those skilled in the art will recognize that such processing capabilities are incorporated into many different devices, and therefore the term "computer" includes PCs, servers, mobile phones, personal digital assistants, and many other devices.

[0080] Those skilled in the art will recognize that the storage device for storing program instructions can be distributed across a network. For example, a remote computer can store a processing instance described as software. A local or terminal computer can access the remote computer and download a part or all of the software to run the program. Alternatively, a local computer can download a software on demand or execute some software instructions at a local terminal and some software instructions at a remote computer (or computer network). Those skilled in the art will also recognize that all or part of the software instructions can be executed by a dedicated circuit (such as a DSP, a programmable logic array, etc.) by using conventional techniques known to those skilled in the art.

[0081] It should be understood that the benefits and advantages described above may relate to one embodiment or may relate to several embodiments. The embodiments are not limited to those that solve any or all of the stated problems or those that have any or all of the stated benefits and advantages.

[0082] Any reference to 'an' item refers to one or more of those items.The term "comprising" is used herein to mean including the identified method steps or elements, but such steps or elements do not comprise an exclusive list and the method or apparatus may contain additional steps or elements.

[0083] As used herein, the terms "component" and "system" are intended to encompass a computer-readable data storage device configured with computer-executable instructions that enable certain functionality to be performed when executed by a processor. Computer-executable instructions may include routines, functions, etc. It should also be understood that a component or system may be localized on a single device or distributed across several devices.

[0084] Further, as used herein, the term "exemplary" is intended to mean "serving as an illustration or example of something."

[0085] Further, to the extent that the term "comprising" is used in either the detailed description or the claims, such term is intended to be open ended in a manner similar to the term "including," in that "including," when used, is to be interpreted as a transitional word in a claim term.

[0086] The accompanying drawings show exemplary methods. Although the method is shown and described as a series of actions performed in a specific sequence, it should be understood and appreciated that the method is not limited by the order of the sequence. For example, some actions may occur in a sequence different from the order described herein. In addition, an action may occur simultaneously with another action. Further, in some instances, all actions are not required to implement the method described herein.

[0087] In addition, the actions described herein may include computer-executable instructions that may be implemented by one or more processors and / or stored on one or more computer-readable media. Computer-executable instructions may include routines, subroutines, programs, execution threads, and / or the like. Still further, the results of the actions of the methods may be stored in a computer-readable medium, displayed on a display device, and / or the like.

[0088] The order of the steps of the methods described herein is exemplary, but the steps may be performed in any suitable order or simultaneously where appropriate. Additionally, steps may be added or replaced or individual steps may be deleted from any method without departing from the scope of the subject matter described herein. Aspects of any example described herein may be combined with aspects of any other example described to form additional examples without loss of effect.

[0089] It should be understood that the above description of the preferred embodiment is given only as an example and that various modifications may be made by those skilled in the art. The content described above includes examples of one or more embodiments. Of course, it is impossible to describe every conceivable modification and change to the above-mentioned device or method for the purpose of describing the aspects mentioned above, but it will be appreciated by those skilled in the art that many additional modifications and arrangements of various aspects are possible. Therefore, the various aspects described are intended to cover all such changes, modifications and variations that fall within the scope of the appended claims.

Claims

1. A computer-implemented method for identifying a compound for targeting one or more target genes, the method comprising: searching the database for first candidate compounds that each target one or more first target genes; A first fingerprint is generated for each first candidate compound by: Searching the database for genes associated with the first candidate compound, predicting genes associated with the first candidate compound using a machine learning model trained to predict gene interaction profiles for a range of compounds, and constructing the first fingerprint using the genes returned by the search and the predicted genes; and Using the first fingerprint to filter the first candidate compounds to identify a first optimized compound for targeting the one or more first target genes, wherein filtering the first candidate compounds comprises: combining filtering by comparing each of said first fingerprints to an ideal fingerprint of a theoretical tool compound and filtering by using a metric generated from said first fingerprint, Each fingerprint includes a polypharmacological fingerprint that describes which genes the compound is associated with.

2. The computer-implemented method of claim 1, wherein the model comprises a neural network. The computer-implemented method of claim 1 , wherein the comparing comprises calculating a similarity score.

4. The computer-implemented method of any one of claims 1 to 3, comprising identifying a first candidate compound that is most similar to the theoretical tool compound as a first optimized compound. 5 . The computer-implemented method of claim 1 , wherein generating a first fingerprint comprises obtaining metadata about one or more first candidate compounds. 6 . The computer-implemented method of claim 5 , wherein the metadata includes clinical trial phase data, drug name or identity, or information from a compound supplier.

7. The computer-implemented method of any one of claims 1 to 3, comprising using a library evaluation framework to retrieve an indication of how many targets each first candidate compound has.

8. The computer-implemented method according to any one of claims 1 to 3, comprising: searching the database for second candidate compounds that each target one or more second target genes; A second fingerprint is generated for each second candidate compound by: Searching the database for genes associated with the second candidate compound, predicting genes associated with a second candidate compound using a machine learning model trained to predict gene interaction profiles for a range of compounds, and constructing the second fingerprint using the genes returned by the search and the predicted genes; and The first fingerprint and the second fingerprint are used to filter the group comprising the first candidate compound and the second candidate compound to identify the first optimized compound and to identify a second optimized compound for targeting the one or more second target genes, wherein filtering the group of first candidate compounds and second candidate compounds comprises comparing each of the first fingerprint and the second fingerprint to an ideal fingerprint of a theoretical tool compound, wherein each fingerprint comprises a polypharmacology fingerprint describing which genes the compound is associated with.

9. A system for identifying a compound for targeting one or more target genes, the system comprising: a compound search module configured to search a database for first candidate compounds each targeting one or more first target genes; A fingerprint module is configured to generate a first fingerprint for each first candidate compound, the fingerprint module comprising: a gene search module configured to search the database for genes associated with the first candidate compound, and a prediction module configured to predict genes associated with a first candidate compound using a machine learning model trained to predict gene interaction profiles of a series of compounds, a fingerprint module configured to construct the first fingerprint using the genes returned by the gene search module and the prediction module; and a filter module configured to filter the first candidate compounds using the first fingerprint to identify a first optimized compound for targeting the one or more first target genes, wherein filtering the first candidate compounds comprises: combining filtering by comparing each of said first fingerprints to an ideal fingerprint of a theoretical tool compound and filtering by using a metric generated from said first fingerprint, Each fingerprint includes a polypharmacological fingerprint that describes which genes the compound is associated with.

10. The system of claim 9, wherein the model comprises a neural network. The system of claim 9 , wherein the comparing comprises calculating a similarity score.

12. The system of any one of claims 9-11, wherein the filter module is configured to identify one or more first candidate compounds that are most similar to the theoretical tool compound.

13. The system according to any one of claims 9-11, wherein the filter module is configured to select a first candidate compound that is most similar to the theoretical tool compound as the first optimized compound.

14. The system according to any one of claims 9-11, wherein the fingerprint module is configured to obtain metadata about one or more first candidate compounds.

15. The system of claim 14, wherein the metadata includes clinical trial phase data, drug name or identity, or information from a compound supplier.

16. The system of any one of claims 9-11, wherein the fingerprint module is configured to use a library evaluation framework to retrieve an indication of how many targets each first candidate compound has.

17. The system according to any one of claims 9 to 11, wherein: The compound search module is configured to search the database for second candidate compounds each targeting one or more second target genes; The fingerprint module is configured to generate a second fingerprint for each second candidate compound; The gene search module is configured to search the database for genes related to the second candidate compound; The prediction module is configured to predict genes associated with the second candidate compound using a machine learning model, the machine learning model being trained to predict gene interaction profiles of a series of compounds, and the fingerprint module is configured to construct the second fingerprint using the genes returned by the gene search module and the prediction module; and The filter module is configured to filter the group comprising the first candidate compound and the second candidate compound using the first fingerprint and the second fingerprint to identify the first optimized compound and to identify the second optimized compound for targeting the one or more second target genes, wherein filtering the group of first candidate compounds and second candidate compounds comprises comparing each of the first fingerprint and the second fingerprint to an ideal fingerprint of a theoretical tool compound, wherein each fingerprint comprises a polypharmacological fingerprint describing which genes the compound is associated with.

18. A computer-readable medium storing code which, when executed by a computer, causes the computer to perform the method of any one of claims 1-8.

Images