Method, system and non-transitory computer readable medium for predicting properties of sample molecules - Patents.com

Abstract

To predict the properties of a chemical structure of a sample.SOLUTION: A method for predicting the properties of a sample molecule involves, for each of a number of reference molecules, obtaining a number of fingerprints and at least one property, and obtaining the number of fingerprints of the sample molecule. The method further involves, for each of the number of reference molecules, using each of the number of fingerprints, calculating the distance to the sample molecule, and for each of the number of reference molecules, determining a relative predictive dominance, based on the distance to the sample molecule. The method also involves, for each of the number of reference molecules, determining a fitness value based on the relative predictive dominance, and predicting the at least one property of the sample molecule based on the at least one property of the number of reference molecules and the fitness values obtained for the reference molecules.SELECTED DRAWING: Figure 1

Description

[Background technology]

[0001] [CROSS REFERENCE TO RELATED APPLICATIONS] This application claims priority under 35 U.S.C. § 119(e) to U.S. Provisional Application No. 62 / 730,913, entitled "METHOD AND SYSTEM FOR PREDICTING PROPERTIES OF CHEMICAL STRUCTURES," filed September 13, 2018, which has at least one of the same inventors as this application. U.S. Provisional Application No. 62 / 730,913 is incorporated herein by reference.

[0002] The relationship between chemical structure and the activity or properties of those chemical structures is known and described for many chemical structures. A newly discovered or synthesized chemical structure of a sample may have properties that are initially unknown. The similarity of the sample's chemical structure to known chemical structures can be used to predict the properties of the sample's chemical structure. Summary of the Invention

[0003] In general, in one aspect, one or more embodiments relate to a method for predicting a property of a sample molecule, the method including obtaining a plurality of fingerprints and at least one property for each of a plurality of reference molecules, obtaining a plurality of fingerprints of the sample molecule, calculating a distance to the sample molecule for each of the plurality of reference molecules using each of the plurality of fingerprints, determining a relative predictive dominance for each of the plurality of reference molecules based on the distance to the sample molecule, determining a fitness value for each of the plurality of reference molecules based on the relative predictive dominance, and predicting at least one property of the sample molecule based on the at least one property of the plurality of reference molecules and the fitness value obtained for the reference molecules.

[0004] In general, in one aspect, one or more embodiments relate to a system for predicting a property of a sample molecule, the system including: a reference molecule repository; and a molecular property prediction engine configured to: obtain a plurality of fingerprints and at least one property for each of a plurality of reference molecules from the reference molecule repository; obtain a plurality of fingerprints for the sample molecule; calculate a distance to the sample molecule for each of the plurality of reference molecules using each of the plurality of fingerprints; determine a relative prediction dominance for each of the plurality of reference molecules based on the distance to the sample molecule; determine a fitness value for each of the plurality of reference molecules based on the relative prediction dominance; and predict at least one property of the sample molecule based on the at least one property of the plurality of reference molecules and the fitness value obtained for the reference molecule.

[0005] In general, in one aspect, one or more embodiments relate to a non-transitory computer-readable medium including computer-readable program code for predicting a property of a sample molecule, the computer-readable program code causing a computer system to: obtain a plurality of fingerprints and at least one property for each of a plurality of reference molecules; obtain a plurality of fingerprints of the sample molecule; calculate a distance to the sample molecule for each of the plurality of reference molecules using each of the plurality of fingerprints; determine a relative predictive dominance for each of the plurality of reference molecules based on the distance to the sample molecule; determine a fitness value for each of the plurality of reference molecules based on the relative predictive dominance; and predict at least one property of the sample molecule based on the at least one property of the plurality of reference molecules and the fitness value obtained for the reference molecules.

[0006] Other aspects of the embodiments will become apparent from the following description and the appended claims. [Brief description of the drawings]

[0007] The present embodiments are illustrated by way of example and are not intended to be limited by the figures of the accompanying drawings. [Figure 1] FIG. 1 illustrates a block diagram of a system in accordance with one or more embodiments of the present invention. [Diagram 2] 1 illustrates a flowchart in accordance with one or more embodiments of the present invention. [Figure 3A] 1 illustrates pseudocode in accordance with one or more embodiments of the present invention. [Figure 3B] 1 illustrates pseudocode in accordance with one or more embodiments of the present invention. [Figure 4] 1 shows a performance comparison between an embodiment of the present invention and a conventional classifier. [Figure 5A] 1 shows cross-validation versus blinded study ROC AUC scores for embodiments of the present invention. [Figure 5B] Cross-validation versus blinded study ROC AUC scores of traditional quantitative structure-activity relationship (QSAR) approaches are shown. [Figure 6] 1 shows a summary of the performance of embodiments of the present invention versus the best performing combination of conventional classifiers and fingerprints. [Figure 7A] 1 illustrates a computing system in accordance with one or more embodiments of the present invention. [Figure 7B] 1 illustrates a computing system in accordance with one or more embodiments of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0008] Certain embodiments disclosed herein are described in detail below with reference to the accompanying drawings, in which like elements in the various figures may be designated by like reference numbers and / or like names for consistency.

[0009] The following detailed description is merely exemplary in nature and is not intended to limit the embodiments disclosed herein or the application and uses of the embodiments disclosed herein. Furthermore, there is no intention to be bound by any expressed or implied theory presented in the preceding technical field, background, brief summary or the following detailed description.

[0010] In the following detailed description of some embodiments disclosed herein, numerous specific details are set forth in order to provide a more thorough understanding of the various embodiments disclosed herein. However, it will be apparent to those skilled in the art that the embodiments may be practiced without these specific details. In other instances, well-known features have not been described in detail to avoid unnecessarily complicating the description.

[0011] Throughout this application, ordinal numbers (e.g., first, second, third, etc.) may be used as adjectives for elements (i.e., any noun in this application). The use of ordinal numbers does not imply or create any particular order of elements or limit any element to being only a single element, unless expressly disclosed, such as by use of terms such as "before," "after," "single," etc. Rather, the use of ordinal numbers is to distinguish elements. By way of example, a first element is different from a second element, a first element encompasses multiple elements, and may follow (or precede) a second element in the ordering of elements.

[0012] In one or more embodiments of the present invention, the properties of a sample's chemical structure are predicted based on the known properties of other chemical structures. This method can be used to predict the properties of small molecules, i.e., molecules with low molecular weight. These molecules can be, for example, pharmacological molecules. To evaluate whether a molecule is pharmacologically useful, it may be necessary to determine its properties. For example, it may be necessary to evaluate the distribution, metabolism, excretion, and toxicity (ADMET) properties of the molecule.

[0013] The relationship between chemical structures and the activities or properties of these chemical structures is known and described for many chemical structures. A newly discovered or synthesized chemical structure of a sample may have initially unknown properties. The similarity of the sample's chemical structure to known chemical structures can be used to predict the properties of the sample's chemical structure. In one or more embodiments of the present invention, the prediction is made using quantitative structure activity relationships (QSAR). The approach described below, according to one or more embodiments of the present invention, relies on multiple / many reference molecules, each of which may be represented by multiple fingerprints, to make a prediction of the sample molecule.

[0014] 1 illustrates a system (100) for predicting properties of chemical structures in accordance with one or more embodiments of the present invention. The system (100) includes a reference molecule repository (110) and a molecular property prediction engine (150). These components are described below.

[0015] The reference molecule repository (110) can be any type of storage unit and / or device for storing data (e.g., a file system, a database, a collection of tables, or any other storage mechanism). The reference molecule repository (110), according to one or more embodiments of the present invention, stores data related to reference molecules (112A-112N). More specifically, for each reference molecule, a molecular fingerprint and a molecular property are stored. Because the molecular fingerprint and molecular property are known, these reference molecules can be used to predict the molecular properties of sample molecules for which only the molecular fingerprint is known, but the molecular property is unknown. Data for any number of reference molecules can be stored in the reference molecule repository (110).

[0016] According to one or more embodiments of the present invention, the data stored about the reference molecules (112A-112N) includes molecular fingerprints (116A-116N) and molecular characteristics (114A-114N).

[0017] For a molecule, any number of molecular properties can be described. Molecular properties include, but are not limited to, physicochemical properties, ADMET properties, and biological activity. Examples of ADMET properties are AMES toxicity, androgen receptor binding, blood-brain barrier compatibility, Caco-2 permeability, hERG inhibition, and carcinogenicity.

[0018] A fingerprint (or other descriptor), according to one or more embodiments of the invention, captures the structure of a molecule in a descriptor format. A fingerprint is not necessarily unique to one molecule, but may allow the underlying molecule to be compared to other molecules to assess the similarity of these molecules. Fingerprints can be, for example, daylight (standard), atom_pair (standard), morgan2 (morgan with r=2), morgan4 (r=4), morgan2_feat (using features with r=2), morgan4_feat (etc.), pharm_base (pharm using default rdkit factory), pharm_gobbi (pharm using gobbi factory), layered (standard), pattern (standard). Fingerprints may be derived from a SMILES representation of the underlying molecule.

[0019] Each type of fingerprint associated with a molecule may represent the underlying information (i.e., the molecule itself) in a slightly different way, and therefore each type of fingerprint may establish an incomplete representation of the underlying molecule, with slightly different noise characteristics.

[0020] Continuing with the description of Figure 1, the reference molecule repository (110) is operatively connected to a molecular property prediction engine (150). The molecular property prediction engine (150) includes instructions in the form of computer readable program code for performing at least one of the steps described in Figure 2.

[0021] FIG. 2 illustrates a flow chart according to one or more embodiments. The flow chart of FIG. 2 illustrates a method for predicting properties of chemical structures, such as small molecules. One or more of the steps of FIG. 2 may be performed by components of the system (100) discussed above with reference to FIG. 1. In one or more embodiments, one or more of the steps shown in FIG. 2 may be omitted, repeated, and / or performed in a different order than that shown in FIG. 2. Additional steps may be further performed. Thus, the scope of the invention should not be considered limited to the particular arrangement of steps shown in FIG. 2.

[0022] In step 200, fingerprints and characteristics for the reference molecules are obtained. In one or more embodiments of the present invention, multiple fingerprints (e.g., 10 fingerprints) are obtained for each reference molecule. Each obtained fingerprint can be of a different type. Thus, the underlying information captured by the multiple fingerprints (i.e., the molecule itself) is always the same, but may be captured to different degrees by different types of fingerprints. The same type of fingerprint is obtained for each of the reference molecules. Depending on the analysis to be performed, one or more characteristics can be obtained for each of the reference molecules. For example, if the overall analysis is aimed at determining whether the sample molecule is capable of crossing the blood-brain barrier (BBB), a BBB crossing characteristic can be obtained for each of the reference molecules. If multiple characteristics are to be evaluated, multiple characteristics can be obtained.

[0023] In step 202, a fingerprint is obtained for the sample molecule. The same type of fingerprint is obtained as that obtained for the reference molecule.

[0024] In step 204, weights for types of fingerprints may be determined based on predicted characteristics. Broadly speaking, one type of fingerprint may be better at predicting a particular property than other types of fingerprints. Thus, a higher weight may be assigned to this type of fingerprint for subsequent operations. The weights of the fingerprints may be determined as follows:

[0025] For example, consider the ability of a molecule to cross the blood-brain barrier (BBB) ​​(BBB crossing property). For each of the reference molecules, the BBB crossing property is assumed to be known. Furthermore, based on the fingerprints of these reference molecules, the distance from each molecule to each other molecule can be calculated. Thus, for each possible pair of reference molecules, and using all available fingerprints, the distance can be calculated. To determine the weight, the correlation of the distance between the molecules based on the fingerprints, and the BBB crossing property can be analyzed. Generally speaking, for a pair of reference molecules, it can be expected that closer fingerprints are more likely to exhibit similar BBB crossing properties than fingerprints that are more distant. Thus, a fingerprint with a short distance between two reference molecules with identical BBB crossing properties (i.e., both reference molecules can cross the blood-brain barrier (BBB++) or both molecules cannot cross the blood-brain barrier (BBB--)) can be assigned a high weight. If this analysis is performed for many molecules (e.g., all reference molecules), the ability of the fingerprints to predict the BBB crossing property can be probabilistically evaluated. A high weight is assigned to fingerprints with a high probability of matching the characteristics P(BBB--) and P(BBB++), while a low weight is assigned to fingerprints with a low probability of matching the characteristics P(BBB--) and P(BBB++) (i.e., a high probability of mismatching the characteristics P(BBB+-) and P(BBB-+)). Thus, in the example of predicting BBB crossing properties, the fingerprint weights may include the posterior probabilities P(BBB+|reference molecule BBB+, d) and P(BBB-|reference molecule BBB-, d). That is, a high weight is assigned to fingerprints that are likely to correctly predict the BBB crossing property given the distance d, while a low weight is assigned to fingerprints that are unlikely to correctly predict the BBB crossing property given the distance d.

[0026] Therefore, fingerprint weights can be obtained for all available fingerprints. Fingerprint weights are, for example, in the numerical range of 0.0 to 1.0. Although the optimization of fingerprint weights is shown based on the example of BBB crossing properties, those skilled in the art will understand that fingerprint weights can be obtained for any other molecular properties.

[0027] Implementation of step 204 is optional, ie, the subsequent steps may alternatively be performed using uniform weighting for the different types of fingerprints.

[0028] In step 206, for each reference molecule and using each of the fingerprints, a distance to the sample molecule is calculated. The distance obtained between the reference molecule and the sample molecule may represent the inverse of the similarity of these two molecules based on the type of fingerprint used in the calculation. In other words, a smaller distance is obtained for more similar molecules and a larger distance is obtained for less similar molecules. In one embodiment of the present invention, the Tanimoto distance is used. The distance calculation can be performed separately for each type of fingerprint. Thus, if 10 types of fingerprints are available for the reference molecule and the sample molecule, 10 distances are obtained to represent the similarity of the reference molecule and the sample molecule. Furthermore, if a total of 50 reference molecules are available, a total of 500 distances (50 x 10) are obtained. The obtained distances represent the similarity of the reference molecule and the sample molecule in a high-dimensional space. The dimension of the space may be a result of the number of fingerprint types used. Specifically, in the above example, a 10-dimensional space (based on the use of 10 types of fingerprints) results, with 50 reference molecules being represented. The resulting distances may vary between different fingerprint types, since each type of fingerprint is associated with slightly different molecular features.

[0029] In step 208, the relative predictive dominance is determined for each of the reference molecules. Specifically, each reference molecule can be compared to any other reference molecule based on the distance obtained in step 206. For example, consider reference molecules A and B, both associated with three fingerprints. Thus, after performing step 206, three distances are available for each of the reference molecules A and B. Assume that the distance for molecule A is [0.6 0.4 0.7] and the distance for molecule B is [0.7 0.4 0.1]. When comparing A and B directly, for the first fingerprint, molecule A is closer than molecule B, for the second fingerprint, molecules A and B are equally close, and for the third fingerprint, molecule B is closer than molecule A. The scoring vector AvsB=[1 0.5 0] can be written, with "1" indicating that molecule A is closer than molecule B, 0 indicating that molecule B is closer than molecule A, and 0.5 indicating that molecules A and B are equally close. The weights obtained in step 204 can be used to adjust the area where distances are considered similar (resulting in a score of 0.5). A large weight may narrow the range where distances are considered similar, while a small weight may broaden the range where distances are considered similar. A weight of zero will always result in a score of 0.5, regardless of the actual distance. This step can be performed for any combination of reference molecules. Pseudocode providing an example for implementation is provided in FIG. 3A.

[0030] The predicted dominance for multiple reference molecules can be summarized in a dominance matrix that includes predicted dominance values ​​for the reference molecules.

[0031] In step 210, a fitness is determined for each reference molecule. In one or more embodiments of the invention, a scoring function is used to convert the dominance relationships obtained in step 208 into a single fitness value for each reference molecule. The fitness value may quantify how similar a particular reference molecule is to the sample molecule. Obtaining the fitness for each of the reference molecules is illustrated by the pseudocode provided in FIG. 3B. When assigning a fitness value, the scoring function may prioritize reference molecules that compare favorably on average, i.e., for all or most of the fingerprints. These may be reference molecules that are mostly "dominant" over other reference molecules (based on the predicted dominance obtained in step 208) and are less "dominant" over other reference molecules. In other words, a "best in class" reference molecule may be a reference molecule that is closest to the target molecule for more fingerprints than any other reference molecule.

[0032] In step 212, the properties of the sample molecule are predicted based on the properties of the reference molecules, scaled by the fitness values. The fitness values ​​can be scaled to a numerical range of 0 to 1. In one embodiment of the present invention, the fitness values ​​are scaled such that the sum of all fitness values ​​is "1". A "0" value, indicating low (or no) fitness, indicates that the associated reference molecule is a poor candidate for predicting the properties of the sample molecule. In contrast, a high value indicates that the associated reference molecule is a very good candidate to be used to predict the properties of the sample molecule. Thus, based on the fitness values ​​associated with the reference molecules, the properties of the reference molecules can be considered for prediction of the properties of the sample molecule.

[0033] For example, consider a scenario where a binary trait is predicted for a sample molecule. A "1" may be used to represent the trait being present and a "0" may be used to represent the trait being absent. Assume that a number of reference molecules are available for which this trait has previously been evaluated. One of these reference molecules has the trait and it has been determined that the fitness value associated with this reference molecule is 0.78. Thus, in step 212, this molecule will vote for "1" with a contribution of 0.78. The same analysis can be performed on all other reference molecules to obtain a majority vote. Each of the reference molecule's contributions to the majority vote can be scaled based on their associated fitness value. It is then determined based on the majority vote whether the binary trait associated with the sample molecule is "0" or "1".

[0034] Similar predictions based on majority voting can be made for properties that contain more than two classes. Furthermore, predictions for continuous properties can also be performed by scaling the contribution of each reference molecule based on its associated fitness value.

[0035] Various embodiments of the present invention have one or more of the following advantages: The embodiments of the present invention allow prediction of properties of a sample molecule. The prediction considers multiple types of fingerprints obtained for multiple reference molecules. Thus, there is no need to select a specific fingerprint before making the prediction. Distance-based metrics in a multidimensional space are used to identify molecular properties. There is no need to manually select hyperparameters, and new reference molecules can be considered in the prediction as soon as they become available. Based on the consideration of multiple fingerprints in the prediction, the risk of bias and overfitting of the prediction is reduced. Thus, the pitfalls of high-dimensional information are avoided.

[0036] FIG. 4 shows a performance comparison (400) of an embodiment of the present invention with a conventional classifier. In the comparison, the embodiment was tested in parallel with five other conventional classifiers to predict 17 ADMET properties. For each of the five conventional approaches, a five-fold cross-validation was performed using 80% of the dataset, leaving 20% ​​for blind testing. This process was repeated for eight types of chemical fingerprints, and a grid search strategy was used to tune the hyperparameters. In an embodiment of the present invention, an 80% / 20% test split was also applied to the direct comparison (denoted as "AP test set"), as well as a "leave one out" full cross-validation (denoted as "AP full set") to evaluate the stability of predictions against dataset size. Ten different fingerprints were used simultaneously to build one model, whereas the standard approach has only one fingerprint per conventional model. In the testing of the conventional classifiers, a model was built separately for each fingerprint, the best performing model in terms of cross-validation score was selected, and performance was measured on the test set. Predictive power for all models was assessed using the area under the receiver operating characteristic curve (ROC AUC) measure of predictive performance. The datasets used for the 17 different models were taken from curated published QSAR studies, each focused on building models optimized for a specific ADMET property.

[0037] As shown in Figure 4, embodiments of the present invention consistently outperform five conventional classifiers for all 17 ADMET traits based on ROC AUC scores. Because embodiments of the present invention have the ability to use all data for training without the need for an 80 / 20% split required by conventional approaches, high predictive power is obtained for most traits, as shown by the difference between "AP Full" and "AP Test" in Figure 4. Furthermore, even when using the same 80 / 20% data split that allows for a direct comparison between embodiments of the present invention and conventional classifiers, "AP Test" still outperforms each of the conventional classifiers.

[0038] Figures 5A and 5B show the ROC AUC scores of cross-validation versus blinded testing of an embodiment of the present invention (500A, Figure 5A) and a traditional quantitative structure-activity relationship (QSAR) approach (500B, Figure 5B) to assess overfitting. Almost all features represented by different points in the plot in Figure 5A (embodiments of the present invention) fall on the diagonal or are very close, whereas in Figure 5B (traditional classifiers), validation performance tends to be higher than test performance, indicating overfitting. The androgen receptor (AR) binding model appears to be overfitted for both tools, but the embodiment of the present invention still performs better (test: 0.62) than the best traditional classifier (test: 0.50).

[0039] Figure 6 shows a performance summary (600) of an embodiment of the present invention versus the best performing conventional classifier and fingerprint combination. The inability of conventional classifiers to use many fingerprints simultaneously results in inconsistencies in the type of fingerprint / classifier that produces the best performing model. This inconsistency is most evident with cytochrome P450 isoforms, where different fingerprints / classifiers output the best performing classifier for each isoform.

[0040] To address concerns related to the implementation of conventional classifiers, the performance of the present embodiment has also been compared with results previously reported in the literature. For example, Shen et al. (Shen J. et al., Estimation of ADME Properties with Substructure Pattern Recognition. J.Chem.Inf.Model.50,1034-1041(2010)) report the performance of AdmetSAR, a tool used in DrugBank, on blood-brain barrier (BBB) ​​and human intestinal absorption (HIA) properties. Based on their report, AdmetSAR ROC AUC scores (BBB:0.9517, HIA:0.9458) are better than the present embodiment (BBB:0.9738, HIA:0.9533).

[0041] In conclusion, embodiments of the present invention consistently outperform traditional machine learning methods.

[0042] Prediction can be performed on a wide range of molecules to predict a wide range of properties. One or more embodiments of the present invention can be particularly beneficial for the development of new drugs. Drug development is a long and expensive process. The ability to predict ADMET properties using embodiments of the present invention can accelerate this process and reduce costs. Furthermore, unlike typical machine learning based methods, there is no need to optimize hyperparameters, so embodiments of the present invention can be used by drug developers without a background in machine learning.

[0043] Unlike traditional modeling approaches that provide very specific tools designed to solve a single problem, embodiments of the present invention are applicable to a variety of fields beyond drug development. Any field that predicts quantitative or qualitative properties of molecules, molecular interactions, etc. can benefit. Embodiments of the present invention can be used when predicting molecular properties, in new molecular discovery / development cycles to iteratively obtain molecules with desired properties, and / or when identifying alternative uses for existing molecules.

[0044] The embodiments of the present disclosure may be implemented on a computing system. Any combination of mobile, desktop, server, router, switch, embedded device, or other types of hardware may be used. For example, as shown in FIG. 7A, a computing system (700) may include one or more computer processors (702), non-persistent storage (704) (e.g., volatile memory such as random access memory (RAM), cache memory), persistent storage (706) (e.g., hard disk, optical drive such as compact disk (CD) drive or digital versatile disk (DVD) drive, flash memory, etc.), communication interface (712) (e.g., Bluetooth interface, infrared interface, network interface, optical interface, etc.), and numerous other elements and functions.

[0045] The computer processor(s) (702) may be an integrated circuit for processing instructions. For example, the computer processor may be one or more cores or micro-cores of a processor. The computing system (700) may also include one or more input devices (710), such as a touch screen, a keyboard, a mouse, a microphone, a touch pad, an electronic pen, or any other type of input device.

[0046] The communications interface (712) may include integrated circuits for connecting the computing system (700) to a network (not shown) (e.g., a local area network (LAN), a wide area network (WAN) such as the Internet, a mobile network, or any other type of network) and / or to another device, such as another computing device.

[0047] Additionally, the computing system (700) may include one or more output devices (708), such as a screen (e.g., a liquid crystal display (LCD), plasma display, touch screen, cathode ray tube (CRT) monitor, projector, or other display device), a printer, external storage, or any other output device. The one or more output devices may be the same as or different from the input devices. The input and output device(s) may be locally or remotely connected to the computer processor (702), non-persistent storage (704), and persistent storage (706). Many different types of computing systems exist, and the aforementioned input and output devices may take other forms.

[0048] Software instructions in the form of computer readable program code for carrying out embodiments of the present disclosure may be stored, in whole or in part, temporarily or permanently, on a non-transitory computer readable medium, such as a CD, DVD, storage device, diskette, tape, flash memory, physical memory, or other computer readable storage medium. In particular, the software instructions may correspond to computer readable program code that, when executed by a processor, is configured to carry out one or more embodiments of the present disclosure.

[0049] The computing system (700) of FIG. 7A is connected to or may be part of a network. For example, as shown in FIG. 7B, the network (720) may include multiple nodes (e.g., node X (722), node Y (724)). Each node may correspond to a computing system such as the computing system shown in FIG. 7A, or a group of nodes combined may correspond to the computing system shown in FIG. 7A. As an example, an embodiment of the present disclosure may be implemented on a node of a distributed system that is connected to other nodes. As another example, an embodiment of the present disclosure may be implemented on a distributed computing system having multiple nodes, and each portion of the present disclosure may be located on a different node in the distributed computing system. Furthermore, one or more elements of the computing system (700) described above may be located at a remote location and connected to other elements over a network.

[0050] Although not shown in Figure 7B, a node may correspond to a blade in a server chassis that connects to other nodes via a backplane. As another example, a node may correspond to a server in a data center. As another example, a node may correspond to a computer processor or micro-core of a computer processor having shared memory and / or resources.

[0051] Nodes in network (720) (e.g., node X (722), node Y (724)) may be configured to provide services to client devices (726). For example, the nodes may be part of a cloud computing system. The nodes may include functionality to receive requests from and send responses to client devices (726). The client devices (726) may be computing systems, such as the computing system illustrated in FIG. 7A. Additionally, the client devices (726) may include and / or execute all or a portion of one or more embodiments of the present disclosure.

[0052] The computing system or groups of computing systems described in Figures 7A and 7B may include functionality to perform various operations disclosed herein. For example, the computing systems may perform communication between processes on the same or different systems. Various mechanisms using some form of active or passive communication may facilitate data exchange between processes on the same device. Examples representing these inter-process communications include, but are not limited to, files, signals, sockets, message queues, pipelines, semaphores, shared memory, message passing, and memory-mapped file implementations. Further details regarding some of these non-limiting examples are provided below.

[0053] Based on the client-server networking model, a socket may function as an interface or communication channel endpoint to enable bidirectional data transfer between processes on the same device. First and foremost, according to the client-server networking model, a server process (e.g., a process providing data) may create a first socket object. The server process then binds to the first socket object, thereby associating the first socket object with a unique name and / or address. After creating and binding the first socket object, the server process waits and listens for incoming connection requests from one or more client processes (e.g., processes seeking data). At this point, if a client process wishes to obtain data from the server process, the client process begins by creating a second socket object. The client process then proceeds to generate a connection request that includes at least the second socket object and the unique name and / or address associated with the first socket object. The client process then sends the connection request to the server process. Depending on availability, the server process may accept the connection request and establish a communication channel with the client process, or the server process may be busy processing other operations and queue the connection request in a buffer until the server process is ready. An established connection notifies the client process that communication may begin. In response, the client process may generate a data request that specifies the data the client process wishes to obtain. The data request is then sent to the server process. Upon receiving the data request, the server process analyzes the request and collects the requested data. Finally, the server process then generates a response that includes at least the requested data and sends the response to the client process. Data may more generally be transferred as datagrams or a stream of characters (e.g., bytes).

[0054] Shared memory refers to the allocation of virtual memory space to demonstrate a mechanism by which data may be communicated and / or accessed by multiple processes. In a shared memory implementation, an initialization process first creates a shareable segment in a persistent or non-persistent storage device. After creation, the initialization process mounts the shareable segment and then maps the shareable segment to an address space associated with the initialization process. Following mounting, the initialization process proceeds to identify and grant access permissions to one or more authorized processes that may write data to or read data from the shareable segment. Changes made to data in the shareable segment by one process may immediately affect other processes that are also linked to the shareable segment. Furthermore, when one of the authorized processes accesses the shareable segment, the shareable segment is mapped to the address space of that authorized process. Often, only one authorized process may mount the shareable segment at any given time, other than the initialization process.

[0055] Other techniques may be used to share data between processes, such as the various data described in this application, without departing from the scope of this disclosure. The processes may be part of the same or different applications and may run on the same or different computing systems.

[0056] Rather than or in addition to sharing data between processes, a computing system implementing one or more embodiments of the present disclosure may include functionality for receiving data from a user. For example, in one or more embodiments, a user may submit data via a graphical user interface (GUI) on a user device. Data may be submitted via the graphical user interface by a user selecting one or more graphical user interface widgets or inserting text and other data into the graphical user interface widgets using a touchpad, keyboard, mouse, or other input device. In response to a selection of a particular item, information regarding the particular item may be retrieved by the computer processor from persistent or non-persistent storage. Upon selection of an item by a user, the contents of the retrieved data regarding the particular item may be displayed on the user device in response to the user's selection.

[0057] As another example, a request to obtain data regarding a particular item may be sent to a server operatively connected to the user device over a network. For example, a user may select a Uniform Resource Locator (URL) link in a web client of the user device, thereby initiating a Hypertext Transfer Protocol (HTTP) or other protocol request that is sent to a network host associated with the URL. In response to the request, the server may extract data regarding the particular selected item and transmit the data to the device that initiated the request. Once the user device receives the data regarding the particular item, the contents of the received data regarding the particular item may be displayed on the user device in response to the user's selection. In addition to the above examples, data received from the server after selecting the URL link may provide a Hypertext Markup Language (HTML) web page that may be rendered by the web client and displayed on the user device.

[0058] Once data is retrieved, such as by using the techniques described above, or from a storage device, a computing system, in performing one or more embodiments of the present disclosure, may extract one or more data items from the retrieved data. For example, the extraction may be performed by the computing system of FIG. 7A as follows: First, an organizational pattern (e.g., grammar, schema, layout) of the data is determined, which may be based on one or more of position (e.g., bit or string position, Nth token in a data stream, etc.), attribute (where an attribute is associated with one or more values), or hierarchical / tree structure (composed of layers of nodes at various levels of detail, such as nested packet headers or nested document sections). Next, a stream of raw, unprocessed data symbols is parsed into a stream (or hierarchical structure) of tokens (each token may have an associated token "type") in the context of the organizational pattern.

[0059] The extraction criteria are then used to extract one or more data items from the token stream or structure, and the extraction criteria are processed according to an organizational pattern to extract one or more tokens (or nodes from a hierarchical structure). For position-based data, tokens at positions identified by the extraction criteria are extracted. For attribute / value-based data, tokens and / or nodes associated with attributes that satisfy the extraction criteria are extracted. For hierarchical / layered data, tokens associated with nodes that match the extraction criteria are extracted. The extraction criteria may be as simple as an identifier string, or may be a query provided to a structured data repository (which may be organized according to a database schema or data format such as XML).

[0060] The extracted data may be used for further processing by a computing system. For example, the computing system of FIG. 7A may perform data comparison while executing one or more embodiments of the present disclosure. Data comparison may be used to compare two or more data values (e.g., A, B). For example, one or more embodiments may determine whether A>B, A=B, A!=B, A<B, etc. This comparison may be performed by submitting an operation code that specifies an operation regarding the comparison to A, B, and an arithmetic logic unit (ALU) (i.e., a circuit that performs arithmetic operations and / or bitwise logical operations on two data values). The ALU outputs a numerical result of the operation and / or one or more status flags associated with the numerical result. For example, the status flag may indicate whether the numerical result is a positive number, a negative number, zero, etc. The comparison may be performed by selecting an appropriate operation code and then reading the numerical result and / or the status flag. For example, to determine whether A>B, B may be subtracted from A (i.e., A-B), and the status flag may be read to determine whether the result is positive (i.e., if A>B, then A-B>0). In one or more embodiments, B may be regarded as a threshold value, and A may be regarded as satisfying the threshold value if A=B or A>B as determined using the ALU. In one or more embodiments of the present disclosure, A and B may be vectors, and comparing A with B may require comparing the first element of vector A with the first element of vector B, comparing the second element of vector A with the second element of vector B, etc. In one or more embodiments, if A and B are character strings, the binary values of the character strings may be compared.

[0061] The computing system of Figure 7A may implement and / or be connected to a data repository. For example, one type of data repository is a database. A database is a collection of information structured to facilitate data retrieval, modification, reorganization, and deletion. A database management system (DBMS) is a software application that provides an interface for a user to define, create, query, update, or manage a database.

[0062] A user or software application may submit a statement or query to the DBMS. The DBMS then interprets the string. The string may be a select statement, update statement, create statement, delete statement, etc., that requests information. Additionally, the statement may include parameters that specify the data or data container (database, table, record, column, view, etc.), identifiers, conditions (comparison operators), functions (e.g., join, full join, count, average, etc.), sort (e.g., ascending, descending), etc. The DBMS may execute the statement. For example, the DBMS may access memory buffers, reference or index files to read, write, delete, or any combination thereof to respond to the statement. The DBMS may load data from persistent or non-persistent storage and perform calculations to respond to the query. The DBMS may return the results to the user or software application.

[0063] The computing system of FIG. 7A may include functionality to provide raw and / or processed data, such as the results of comparisons and other processing. For example, providing the data may be accomplished through various presentation methods. Specifically, the data may be provided through a user interface provided by the computing device. The user interface may include a GUI that displays information on a display device, such as a computer monitor or a touch screen on a handheld computing device. The GUI may include various GUI widgets that organize what data is shown and how the data is presented to the user. Additionally, the GUI may provide data directly to the user, for example, data provided as actual data values ​​through text, or data that is rendered by the computing device into a visual representation of the data, such as through visualization of a data model.

[0064] For example, the GUI may first obtain a notification from a software application that a particular data object is to be provided within the GUI. The GUI may then determine the data object type associated with the particular data object, for example, by obtaining data from a data attribute within the data object that identifies the data object type. The GUI may then determine any rules specified for displaying that data object type, for example rules specified by the software framework for the data object class or according to local parameters defined by the GUI for presenting that data object type. Finally, the GUI may obtain a data value from the particular data object and render a visual representation of the data value within a display device according to the rules specified for that data object type.

[0065] The data may also be provided in various audio manners, in particular the data may be rendered into an audio format and provided as sound via one or more speakers operatively connected to the computing device.

[0066] The data may also be provided to the user by tactile methods. For example, tactile methods may include vibrations or other physical signals generated by the computing system. For example, the data may be provided to the user using vibrations generated by a handheld computing device with a predetermined duration and vibration intensity to communicate the data.

[0067] The above functional descriptions provide only a few examples of functions performed by the computing systems of Figure 7A and the nodes and / or client devices of Figure 7B. Other functions may be performed using one or more embodiments of the present disclosure.

[0068] While this disclosure has been described with respect to a limited number of embodiments, those skilled in the art having the benefit of this disclosure will appreciate that other embodiments may be devised that do not depart from the scope of the disclosure disclosed herein. Accordingly, the scope of the disclosure should be limited only by the scope of the appended claims.

[0069] The embodiments and examples described herein are presented to best explain the invention and its particular applications, thereby enabling those skilled in the art to make and use the invention. However, those skilled in the art will recognize that the foregoing description and examples have been presented for the purposes of illustration and example only. The description set forth is not intended to be exhaustive or to limit the invention to the precise form disclosed.

[0070] While the present invention has been described with respect to a limited number of embodiments, those skilled in the art, having the benefit of this disclosure, will appreciate that they may devise other embodiments that do not depart from the scope of the invention disclosed herein. Accordingly, the scope of the present invention should be limited only by the appended claims.

Claims

1. 1. A computer-implemented method for predicting a property of a sample molecule, comprising: obtaining a plurality of fingerprints and at least one characteristic for each of a plurality of reference molecules; obtaining the plurality of fingerprints of the sample molecules; For each of the plurality of reference molecules, calculating a distance to the sample molecule using each of the plurality of fingerprints in a multidimensional feature space having a dimension corresponding to the number of fingerprints in the plurality of reference molecules; determining a relative prediction dominance for each of the plurality of reference molecules based on the distance to the sample molecule, the determining the relative prediction dominance including obtaining a scoring vector by comparing a first distance between a first reference molecule of the plurality of reference molecules and the sample molecule and a second distance between a second reference molecule of the plurality of reference molecules and the sample molecule in the multidimensional feature space; determining, for each of the plurality of reference molecules, an associated fitness value that quantifies a measure of similarity of each of the plurality of reference molecules to the sample molecule based on the relative predicted dominance; and predicting the at least one property of the sample molecule by obtaining a majority vote for the at least one property by each of the plurality of reference molecules based on the at least one property of each of the plurality of reference molecules and the associated fitness value of each of the plurality of reference molecules, and scaling a contribution by each of the plurality of reference molecules to the majority vote by the associated fitness value.

2. Calculating the distance to the sample molecule 2. The computer-implemented method of claim 1, comprising calculating a Tanimoto distance to the sample molecule for each of the plurality of reference molecules.

3. Determining the relative predictive dominance 2. The computer-implemented method of claim 1, comprising obtaining an additional scoring vector by comparing a third distance between a third reference molecule of the plurality of reference molecules and the sample molecule to the first distance between a first reference molecule of the plurality of reference molecules and the sample molecule.

4. 2. The computer-implemented method of claim 1, wherein the scoring vector indicates whether the first reference molecule or the second reference molecule is close to the sample molecule.

5. Determining the relative predictive dominance 4. The computer implemented method of claim 3, further comprising combining the scoring vector with the additional scoring vector to form a dominance matrix.

6. determining the associated fitness value based on the relative predictive dominance; 2. The computer-implemented method of claim 1, comprising prioritizing one reference molecule of the plurality of reference molecules over other reference molecules based on the one reference molecule being closer to the sample molecule than any of the other reference molecules for a greater number of fingerprints of the plurality of fingerprints.

7. The at least one characteristic is Binary characteristics, and The computer implemented method of claim 1 , wherein the at least one characteristic is selected from the group consisting of:

8. 1. A system for predicting a property of a sample molecule, comprising: a reference molecule repository; 1. A molecular property prediction engine, comprising: obtaining a plurality of fingerprints and at least one characteristic for each of a plurality of reference molecules from the reference molecule repository; obtaining the plurality of fingerprints of the sample molecules; For each of a plurality of reference molecules, calculating a distance to the sample molecule using each of the plurality of fingerprints in a multidimensional feature space having a dimension corresponding to the number of fingerprints in the plurality of reference molecules; determining a relative prediction dominance for each of the plurality of reference molecules based on the distance to the sample molecule, the determining the relative prediction dominance including obtaining a scoring vector by comparing a first distance between a first reference molecule of the plurality of reference molecules and the sample molecule and a second distance between a second reference molecule of the plurality of reference molecules and the sample molecule in the multidimensional feature space; determining, for each of the plurality of reference molecules, an associated fitness value that quantifies a measure of similarity of each of the plurality of reference molecules to the sample molecule based on the relative predicted dominance; and a molecular property prediction engine configured to: obtain a majority vote for the at least one property by each of the plurality of reference molecules based on at least one property of each of the plurality of reference molecules and the associated fitness value of each of the plurality of reference molecules; and predict the at least one property of the sample molecule by scaling a contribution by each of the plurality of reference molecules to the majority vote by the associated fitness value.

9. Determining the relative predictive dominance 9. The system of claim 8, further comprising: obtaining an additional scoring vector by comparing a third distance between a third reference molecule of the plurality of reference molecules and the sample molecule with the first distance between a first reference molecule of the plurality of reference molecules and the sample molecule, wherein the additional scoring vector indicates whether the first reference molecule or the third reference molecule is closer to the sample molecule.

10. determining the associated fitness value based on the relative predictive dominance; 9. The system of claim 8, comprising prioritizing one reference molecule of the plurality of reference molecules over other reference molecules based on the one reference molecule being closer to the sample molecule than any of the other reference molecules for a greater number of fingerprints of the plurality of fingerprints.

11. 1. A non-transitory computer readable medium comprising computer readable program code for predicting a property of a sample molecule, the computer readable program code being configured to execute, on a computer system, obtaining a plurality of fingerprints and at least one characteristic for each of a plurality of reference molecules; obtaining the plurality of fingerprints of the sample molecules; For each of a plurality of reference molecules, calculating a distance to the sample molecule using each of the plurality of fingerprints in a multidimensional feature space having a dimension corresponding to the number of fingerprints in the plurality of reference molecules; determining a relative prediction dominance for each of the plurality of reference molecules based on the distance to the sample molecule, the determining the relative prediction dominance including obtaining a scoring vector by comparing a first distance between a first reference molecule of the plurality of reference molecules and the sample molecule and a second distance between a second reference molecule of the plurality of reference molecules and the sample molecule in the multidimensional feature space; determining, for each of the plurality of reference molecules, an associated fitness value that quantifies a measure of similarity of each of the plurality of reference molecules to the sample molecule based on the relative predicted dominance; and predicting the at least one property of the sample molecule by obtaining a majority vote for the at least one property by each of the plurality of reference molecules based on the at least one property of each of the plurality of reference molecules and the associated fitness value of each of the plurality of reference molecules, and scaling a contribution by each of the plurality of reference molecules to the majority vote by the associated fitness value.

12. Determining the relative predictive dominance 12. The non-transitory computer-readable medium of claim 11, comprising obtaining an additional scoring vector by comparing a third distance between a third reference molecule of the plurality of reference molecules and the sample molecule with the first distance between a first reference molecule of the plurality of reference molecules and the sample molecule.

13. 13. The non-transitory computer readable medium of claim 12, wherein the scoring vector indicates whether the first reference molecule or the second reference molecule is close to the sample molecule.

14. Determining the relative predictive dominance 13. The non-transitory computer readable medium of claim 12, further comprising combining the scoring vector with the additional scoring vector to form a dominance matrix.

15. determining the associated fitness value based on the relative predictive dominance; 12. The non-transitory computer readable medium of claim 11, comprising prioritizing one reference molecule of the plurality of reference molecules over other reference molecules based on the one reference molecule being closer to the sample molecule than any of the other reference molecules for a greater number of fingerprints of the plurality of fingerprints.

16. The at least one characteristic is 12. The non-transitory computer readable medium of claim 11, wherein the at least one selected from a binary characteristic and a continuous characteristic.

Images