Medical information processing apparatus and information processing method

Abstract

To provide a medical information processing apparatus and an information processing method for visualizing a region which is influenced by multiple medications on a protein network.SOLUTION: In a medical information processing apparatus 10, a memory stores protein-protein interaction data representing the presence / absence of interaction between two or more proteins out of multiple proteins. Processing circuitry 22 is configured to: receive a designation of a first drug associated with a first protein set comprising a first set of proteins, and a second drug associated with a second protein set comprising a second set of proteins; estimate, using the protein-protein interaction data stored in the memory and the first drug and the second drug for which the designation has been received by a receiving unit, a first set of proteins based on the first drug and a second set of proteins based on the second drug; and determine, based on the first set of proteins and the second set of proteins, if there is at least one protein which is influenced by both the first drug and the second drug.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] Specifically, the present invention relates to a medical information processing device and an information processing method for processing protein interaction data when applied to a protein-protein interaction network. [Background technology]

[0002] Biological networks are naturally occurring networks that consist of a scale-free topology with nodes of various degrees. Biological networks can be used to represent interactions of molecules such as proteins that do not act alone but exert effects on other molecules. Protein-protein interaction (PPI) networks represent individual proteins (or each protein-coding gene) as nodes that are connected by edges to other proteins that share functional significance. Edges between nodes can represent canonical physical interactions, direct protein fusions, or correlations in bioavailability (co-expression). The likelihood of a protein having n interactions drops exponentially with n+1. The nodes around a particular protein in the network can be referred to as the local neighbourhood or local interactome of that protein.

[0003] Polypharmacy is a rapidly growing area of ​​healthcare due to an ageing and increasing comorbid population. This is particularly relevant in high-income countries (HICs) where individuals continue to have access to high-quality healthcare but accumulate morbidity over their lifetime. Polypharmacy is loosely defined as the administration of three or more drugs to a patient, with some definitions as >5 drugs. Typically, polypharmacy is applied to target completely different disease states, but it may also be used to treat the same disease state in resistant individuals, for example, three or more different drug families for resistant hypertension or diabetes. Polypharmacy can also be applied in situations where the etiology of a first disease is causing a second disease, for example, when renal impairment causes secondary hypertension. Treatment of the secondary disease may be treated as distinct from treatment of the primary disease and may be associated with polypharmacy.

[0004] It is beneficial for clinicians to understand polypharmacy when evaluating patients who have a poor response to treatment or who have experienced a suspected adverse drug reaction (ADR). It is also beneficial for clinicians to consider the effects of polypharmacy when dealing with heavily medicated patients. Also, during the drug R&D process, it is important to have a deeper understanding of polypharmacy so that the effect of a new drug on patient biology, including potential biological effects, drug-to-gene interactions, and contraindications can be evaluated.

[0005] Current attempts to understand polypharmacy focus on predicting new links in knowledge graphs. These analyses include the generation of large knowledge graphs that represent drugs and their side effects. However, studying the effects of drug combinations based on known side effects ignores the role of proteins and genes that mediate these effects. Proteins and genes are fundamental in explaining the occurrence of these side effects. Homeostatic feedback mechanisms allow cells to respond to perturbations of drugs to proteins (a process that demonstrates drug resistance acquisition over time). Thus, local neighborhoods around drug targets can be affected. Overlapping local neighborhoods in the context of polypharmacy can be crucial in demonstrating potential resistance or ADRs. Understanding the interactions between local neighborhoods is also important for interpreting responses in terms of pharmacodynamics and pharmacokinetics. Thus, polypharmacy needs to be studied in ways that consider the interactions between these intermediate molecules. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] US Patent Application Publication No. 2016 / 246919 Summary of the Invention [Problem to be solved by the invention]

[0007] One of the problems to be solved by the embodiments disclosed in this specification and the drawings is to visualize regions in a protein network that are affected by the use of multiple drugs in combination. However, the problems to be solved by the embodiments disclosed in this specification and the drawings are not limited to the above problem. Problems corresponding to the effects of each configuration shown in the embodiments described later can also be positioned as other problems. [Means for solving the problem]

[0008] The medical information processing device according to the embodiment includes a storage unit, a receiving unit, an estimating unit, and a determining unit. The storage unit stores protein-protein interaction data representing the presence or absence of an interaction between at least two or more proteins among the plurality of proteins for a plurality of proteins. The receiving unit receives designation of a first drug associated with a first protein set consisting of a set of a first protein, and a second drug associated with a second protein set consisting of a set of a second protein. The estimating unit estimates a first protein set based on the first drug and a second protein set based on the second drug, using the protein-protein interaction data stored in the storage unit and the first and second drugs designated by the receiving unit. The determining unit determines whether there is at least one protein that is affected by both the first drug and the second drug, based on the first protein set and the second protein set. [Brief description of the drawings]

[0009] Embodiments will now be described, by way of non-limiting example, and illustrated in the following figures: [Figure 1] FIG. 1 is a schematic diagram of an example of an apparatus according to an embodiment. [Diagram 2] FIG. 2 is a schematic diagram of an example of a method according to an embodiment. [Diagram 3] FIG. 3 is a flow chart outlining an example of a method according to an embodiment. [Figure 4] FIG. 4 is a graph showing an example of threshold calculation according to the embodiment. [Diagram 5] FIG. 5 is an example of a display image of a local interactome of a drug target according to an embodiment. [Figure 6] FIG. 6 is an example display image of the local interactome of three drug targets according to an embodiment. [Figure 7] FIG. 7 is an example of the display image of FIG. 6 further annotated according to an embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0010] One embodiment provides a medical information processing device including a processing circuit, which receives protein-protein interaction data, estimates a first protein set based on a first drug and a second protein set based on a second drug using the protein-protein interaction data, and determines whether there is at least one protein that is affected by both the first drug and the second drug based on the first protein set and the second protein set.

[0011] One embodiment provides a method including receiving protein-protein interaction data; using the protein-protein interaction data to infer a first set of proteins based on a first drug and a second set of proteins based on a second drug; and inferring whether there is at least one protein that is affected by both the first drug and the second drug based on the first set of proteins and the second set of proteins.

[0012] An apparatus 10 according to an embodiment is shown diagrammatically in Fig. 1. The apparatus 10 may also be referred to as a medical information processing apparatus. The apparatus 10 processes protein-protein interaction data. The apparatus 10 further displays an image based on the protein-protein interaction data.

[0013] Protein-protein interaction data can be any type of data, including protein sets and representations of interactions between individual proteins. Interactions between pairs of proteins can correspond to canonical physical interactions, direct protein fusions, or correlations in bioavailability (co-expression). In addition, interactions between pairs of proteins can be predicted. The predicted interactions between pairs of proteins can be based on consideration of the location in the genome of the genes encoding each protein and consideration of the topology of the genome.

[0014] Protein-protein interaction data is a type of omics data. Omics data relates to information obtained by high-throughput experiments to characterize and quantify biomolecules. In other embodiments, the device 10 may process other types of omics data, such as genomics or metabolomics data. In other embodiments, the device 10 may process any suitable data that can be represented using a graph.

[0015] The apparatus 10 comprises a computing device 12, which in this example is a personal computer (PC) or workstation. The computing device 12 is connected to a display screen 16 or other display device and one or more input devices 18, such as a computer keyboard and a mouse. The computing device 12 receives data from a memory 37, which may also be referred to as a data store or storage. In alternative embodiments, the computing device 12 may receive data from one or more further data stores (not shown) instead of or in addition to the memory 37. For example, the computing device 12 may receive data from one or more remote data stores (not shown), which may include cloud-based storage.

[0016] The memory 37 stores the protein-protein interaction data. The memory 37 is an example of a storage unit. The protein-protein interaction data may be stored in any suitable file format for storing text-based data or graphical representations, such as TSV, CSV, XLS, XML, etc. In other embodiments, the protein-protein interaction data may be stored in another suitable memory, such as, for example, another device or a cloud-based memory.

[0017] The computing device 12 includes a processing circuit 22 for processing data. The processing circuit 22 includes a central processing unit (CPU) and a graphical processing unit (GPU). The processing circuit 22 provides processing resources for automatically or semi-automatically processing the protein-protein interaction data.

[0018] The processing circuit 22 includes a graph circuit 24 that receives protein-protein interaction data and converts the protein-protein interaction data into a graph representation if the protein-protein interaction data is not structured as a graph representation, a random walk circuit 26 that performs a random walk on the graph based on the protein-protein interaction data and outputs a list of count values ​​of proteins corresponding to visited nodes, a filtering circuit 28 that filters the proteins based on the count values, an interaction score circuit 30 that calculates the interaction scores of the proteins, a display circuit 32 that displays the graph, and an analysis circuit 34 that further analyzes the list of proteins.

[0019] In this embodiment, the circuits 24, 26, 28, 30, 32, and 34 are each implemented in a CPU and / or GPU by a computer program having computer readable instructions executable to perform the method of the embodiment, while in other embodiments the circuits may be implemented as one or more Application Specific Integrated Circuits (ASICs) or Field Programmable Gate Arrays (FPGAs).

[0020] Computing device 12 also has a hard drive and other components of a PC, including RAM, ROM, a data bus, an operating system including various device drivers, and hardware devices including a graphics card, although such components are not shown in FIG.

[0021] The processing circuitry 22 of Figure 1 performs a method according to Figures 2 and 3. Figure 2 is a schematic diagram of a method for identifying proteins affected by a first drug and a second drug. Figure 3 is a flow chart outlining the method.

[0022] As shown in FIG. 2, in one embodiment, at stage 100, the user receives a list of n types of drugs 40 prescribed for the subject 80, where n is 2 or more. In this case, it can be said that the user receives the designation of the n types of drugs via the processing circuit 22. That is, the processing circuit 22 in this case is an example of a receiving unit. In this embodiment, the drugs are respectively drugs for treating depression, anxiety, and hypertension. In other embodiments, any other combination of drugs may be selected. The user then identifies n proteins 44 each associated with the n types of drugs 30. In this case, it can be said that the user estimates the n proteins 44 associated with each of the n types of drugs 30 via the processing circuit 22. That is, the processing circuit 22 in this case is an example of an estimating unit. In this embodiment, each of the n proteins is a drug target for each drug, and is identified by the user by searching a drug target database. Of course, in some embodiments, preferably, a drug may have multiple targets, and in such a case, each protein 44 associated with a drug may be one of the multiple targets of the drug.

[0023] At stage 120, the graph circuitry 24 receives n proteins 44 and a graph data structure 43 that includes a protein-protein interaction (PPI) network. The network, sometimes referred to as a graph, includes a set of nodes connected by edges. The nodes of the graph data structure 43 correspond to proteins and the edges correspond to interactions between the proteins. The degree of a node, defined as the number of edges going into a node, varies depending on the amount of interactions the proteins have. Nodes within a small number of edges, for example 3, 4, or 5 edges, from a node are defined to be local to the node. By identifying proteins that correspond to nodes that are local to a node representing a protein of interest, a local interactome or local neighborhood of the protein of interest can be found.

[0024] The random walk circuit 26 performs a series of random walks on the graph based on the graph data structure 43. A random walk is the generation of a series of random discrete steps in a mathematical space. When performed on a graph, a random walk consists of a series of random steps, each of which has length 1 and traverses from node to node through the graph. If one lands on a given node, there is an equal probability of traversing any edge that connects to that node. Performing a random walk on a graph allows one to explore the graph in an unbiased manner in a local neighborhood around the starting node.

[0025] For each of the n nodes corresponding to each protein, a random walk of length j is performed i times, resulting in a total of n×i random walks. The data resulting from the random walks is stored as raw count data 45. For each of the n starting proteins 44, the filtering circuit 28 summarizes and then normalizes the raw count data 45 to generate normalized count data 47. The filtering circuit 28 in this case is an example of a normalizer. The filtering circuit 28 then filters the normalized count data 47 based on the degree of each of the n starting nodes to generate filtered count data 48. The filtering circuit 28 in this case is an example of a filtering unit. For each of the n proteins 44, the set of proteins remaining in the filtered count data 48 can be considered to belong to the local neighborhood of the protein. At stage 120, the graph circuit 24, the random walk circuit 26, and the filtering circuit 28 can be said to have determined that for each of the n proteins 44, the set of proteins remaining in the filtered count data 48 are proteins that are affected by all of the n drugs 30. Therefore, in this case, the graph circuit 24, the random walk circuit 26, and the filtering circuit 28 can be said to be an example of a judgment section.

[0026] At stage 150, the display circuitry 32 displays the protein-protein interaction data stored in the graph data structure 43 on the display screen 16. The display circuitry 32 is an example of a display control unit. The nodes of the graph are colored based on the filtered count data 48 so that a user can visualize the local neighborhood of each of the n proteins 44. Furthermore, the nodes are colored so that a user can see which proteins belong to the local neighborhood of at least one of the n proteins 44.

[0027] In FIG. 3, the method is examined in more detail.

[0028] At stage 200, a user receives a list of n drugs 40 and information stored in a drug target database 41. In one embodiment, the n drugs 40 are selected based on medications that the patient is or may be prescribed. In this embodiment, the n drugs 40 include selective serotonin reuptake inhibitors (SSRIs), beta blockers, and calcium channel blockers (CCBs). In another embodiment, at least one of the drugs is a drug in clinical trials.

[0029] A user specifies a list of n proteins 44 each associated with one of the n drugs 40. In this embodiment, the n proteins 44 are drug targets of individual drugs. In this embodiment, the n proteins 44 are the depression serotonin receptor (HTR3A), the anxiety beta adrenergic receptor (ADRB1), and the hypertension L-type calcium channel (RYR1). In other embodiments, at least one of the n proteins 44 may be associated with one of the n drugs 40 because its activity is modulated by the drug, even if it is not a direct target of the drug.

[0030] The drug targets may be identified based on information contained in the drug target database 41 or from a pharmaceutical manufacturer. In this embodiment, the n proteins 44 are determined by a user. In other embodiments, the graph circuitry 24 may automatically determine the n proteins based on the list of n drugs 40 and the drug target database 41. The drug target database 41 may be stored in the memory 37 or any other suitable data storage device.

[0031] At stage 210, the graph circuitry 24 receives protein-protein interaction data 42 from memory 37 or any suitable data storage device. The protein-protein interaction data 42 includes data representing a set of proteins (or a set of genes encoding those proteins) and a representation of the interactions between each protein. The protein-protein interaction data may include a subset of the known proteins in an organism, or may include all of the known proteins in an organism. In the case of humans, the protein-protein interaction data may include all of the known human protein-coding genes (~20,000).

[0032] The graph circuitry 24 converts the protein-protein interaction data 42 into a graph data structure 43 if the protein-protein interaction data 42 is not in a graph data structure format. The graph data structure 43 is a simplified data structure suitable for representing a graph. Examples of graph data structures include an edge list and an adjacency matrix. An edge list specifies the start and end nodes of each edge in the graph and may be represented as a two-column matrix. An adjacency matrix includes rows and columns both labeled with nodes, with a 1 or 0 entered in each cell depending on whether an edge exists connecting to that particular node. In other embodiments, any suitable graph data structure may be used.

[0033] In this embodiment, the protein-protein interaction data 42 is received by the graph circuitry 24 in a format other than the graph data structure 43 and is converted by the graph circuitry 24 to the graph data structure 43. In other embodiments, the graph circuitry 24 may receive the protein-protein interaction data directly in the graph data structure 43 format.

[0034] In this embodiment, the graph data structure 43 corresponds to an unweighted graph, i.e., each edge of the graph has a weight of 1. In other embodiments, if the protein-protein interaction data includes weighted interactions, such as data obtained from the STRING database, the graph circuitry 24 may first apply a threshold to the protein-protein interaction data to obtain unweighted protein-protein interaction data. The unweighted protein-protein interaction data is then converted into the graph data structure 43.

[0035] Alternatively, in some embodiments it is envisioned that the weighted protein-protein interaction data is directly incorporated into the graph data structure 43 such that the graph data structure 43 includes weighted edges.

[0036] At the end of stage 210, graph circuitry 24 outputs a graph data structure 43.

[0037] At stage 220, the random walk circuit 26 receives a graph data structure 43. The random walk circuit also receives a list of n proteins 44. Each of the n proteins is used as a starting node to initiate at least one random walk through a graph based on the graph data structure 43.

[0038] The random walk circuit 26 performs i random walks for each of the n starting nodes, for a total of n×i random walks, where i is an integer greater than or equal to 1. Each walk has length j, where j is an integer greater than 1. Desirably, the length j of each random walk should be long enough to capture the local neighborhood around the starting node n and allow for effective visualization. This could be, for example, a length j of 4 or 5 steps. Longer walks can capture interactions further away from the node origin.

[0039] It is contemplated that any suitable random walk algorithm may be used to perform a random walk on the graph based on the graph data structure 43. If the graph data structure 43 is weighted, then a weighted random walk algorithm may be used to determine the probability of transitions between each node according to the weights of the edges connecting those nodes.

[0040] The random walk circuit 26 records the scores of the visited nodes in the i random walks of each of the n starting nodes, which are stored as raw count data 45. At the end of stage 220, the random walk circuit 26 outputs the raw count data 45.

[0041] At stage 230, filtering circuit 28 receives raw count data 45. Filtering circuit 28 summarizes raw count data 45 to produce summarized count data 46. The summarized count data 46 includes, for each of the n starting nodes, a list of proteins corresponding to visited nodes in the i random walks and associated count values ​​corresponding to the number of visits of the node associated with each protein. The summarized count data 46 is stored as a single table or list indexed for each of the n starting nodes, or as a set of n tables or lists for each of the n starting nodes.

[0042] The count values ​​in the summarized count data 46 are normalized by i, the number of random walks taken for each n starting nodes, resulting in normalized count data 47.

[0043] The filtering circuit 28 then applies a threshold to each of the n sets of proteins in the normalized count data 47 to filter out proteins with low normalized count values. The threshold is chosen to be a function of the degree of the starting node, which is obtained from the graph data structure 43. A threshold value T is chosen according to the function, as shown in FIG.

[0044] T=0.1×0.95 x +0.1 x is the degree of the starting node

[0045] Thus, the higher the degree of the node, the lower the threshold. Such threshold adjustment takes into account the topology of the graph representing the protein-protein interaction data. A random walk starting at a high-dimensional node tends to visit many nodes in fewer visits. A random walk starting at a low-dimensional node tends to visit fewer nodes in more visits. The above function for calculating threshold T adjusts the threshold according to the degree of the node, such that the higher the degree of the starting node, the lower the threshold.

[0046] If each of the n starting nodes has a different dimension, then a different filtering threshold will be applied to each of the n sets of proteins.

[0047] At the end of stage 230, the filtering circuit 28 outputs filtered count data 48. The remaining proteins in the filtered count data 48 of each of the n starting nodes 44 may be referred to as the local neighborhood or local interactome of that protein / associated drug.

[0048] At stage 240 , the interaction score circuit 30 receives the filtered count data 48 .

[0049] The interaction score circuit 30 calculates an interaction score 48 for each unique protein in the filtered count data 48. The interaction score circuit 30 is an example of a calculation unit. The interaction score of a protein is calculated by finding the average count value of the protein to the n drugs 40. The interaction score is sometimes referred to as gene neighborhood overlap. If a protein is not present in one of the n proteins in the filtered count data, the set is considered to have a count value of 0. As a result, a protein that is present in all of the n proteins in the filtered count data 48 will tend to have a high interaction score. Of course, a protein that is disproportionately visited in one or a few of the n proteins may also have a high interaction score. This may occur for a protein that is strongly related to one or more of the n drugs 40, but weakly linked to the remaining drugs.

[0050] At the end of stage 240, the interaction score circuit 30 outputs an interaction score 49 for each protein.

[0051] At the end of stage 250, the display circuitry 32 receives the interaction scores 49, the filtered count data 48, and the graph data structure 43. Based on these inputs, the display circuitry 32 generates a graph display 50 showing which proteins are affected by the n drugs 40.

[0052] In the present embodiment shown in FIG. 5, the display circuitry 32 highlights nodes corresponding to proteins associated with only one of the n drugs 40 based on the filtered count data 48. In other words, the display circuitry 32 shows the local neighborhood of one of the n drugs 40. The starting node is shown by coloring the starting node differently from the other nodes (shown in grayscale in FIG. 5). In other embodiments, other methods may be used to show the starting node, such as the starting node having a different size or shape compared to the other nodes. The display circuitry 32 also provides a display interface that allows the user to select one or a combination of the n sets of proteins to display from the filtered count data 48. This allows the user to explore the local interactome of each of the n drugs 40.

[0053] 6, display circuitry 32 simultaneously displays all of the sets of proteins from the filtered count data 48. Each node belonging to each set of proteins is shaded a different color, and each starting node is shaded the same color. For nodes that correspond to proteins that are part of more than one set of proteins, the node is further colored or modified, or a graphical feature is added, to indicate the overlap of that protein between the protein sets.

[0054] In the present embodiment shown in Figure 7, there is a node representing protein DLG4 in each set of proteins associated with selective serotonin reuptake inhibitors (SSRIs), beta blockers, and calcium channel blockers (CCBs). The node representing protein DLG4 is shaded in a different color than the other nodes in the graph. In an alternative embodiment, nodes corresponding to overlapping proteins are shaded in a combination of colors that indicate all of the drugs with which the protein corresponding to that node is associated. Figure 7 further shows circular regions 52 that indicate nodes that are present in more than one set of proteins. Although not shown, interaction scores 49 may be indicated on the image 50, such as by coloring some of the nodes according to the interaction score or by providing a mouse-over function that displays the interaction score.

[0055] At stage 260, analysis circuitry 34 receives the filtered count data 48 and the graph data structure 43. Some or all of the proteins from the filtered count data 48 are then selected for further analysis, either by a user or automatically by analysis circuitry 34. The proteins selected for analysis may include proteins that belong to the local interactomes of at least two of the n drugs 40.

[0056] The analysis circuit 34 then performs Gene Set Enrichment Analysis (GSEA), pathway analysis, or Uniprot analysis on all or a portion of the genes in the filtered count data to investigate potential biological mechanisms of side effects and adverse reactions when the n drugs 40 are administered to a patient. The analysis circuit 34 is an example of an analysis unit. GSEA is a tool that associates biological pathways stored in the MSigDB database with gene / protein groups. Uniprot is a tool for associating functional information with protein groups. Based on the identified pathways, clinicians and researchers can determine potential side effects and adverse reactions when the n drugs are administered.

[0057] At stage 270, analysis circuitry 34 performs further decision steps based on the determined biological mechanisms, side effects or adverse reactions, or the decision steps may be performed by a clinician or researcher.

[0058] In one embodiment, a first drug is being administered in a clinical trial and a second drug is prescribed to a potential subject in the clinical trial. Stages 200, 210, 230, 240, 250, 260 are evaluated based on the first drug and the second drug. In stage 270, the analysis circuitry determines whether the subject should be removed from the clinical trial and / or whether the subject should be classified into a subgroup of clinical trial participants based on a biological mechanism, side effects, or adverse reactions. In another embodiment, the first drug and the second drug are part of a treatment plan for a patient. Stages 200, 210, 230, 240, 250, 260 are evaluated based on the first drug and the second drug. In stage 270, the analysis circuitry determines whether one of the first drug or the second drug should be removed from the treatment plan and / or whether the dosage of the first drug or the second drug should be changed based on a biological mechanism, side effects, or adverse reactions. In another embodiment, the first drug is a drug candidate and the second drug is a commonly prescribed drug. Stages 200, 210, 230, 240, 250, and 260 are evaluated based on the first drug and the second drug. In stage 270, analysis circuit 34 determines whether the drug candidate should proceed to a drug development stage based on biological mechanisms, side effects, or adverse reactions.

[0059] A random walk on the PPI network may explore the local neighborhood of proteins in an unbiased manner. Filtering the visited nodes takes into account the topology of the graph. This may provide an improvement over naive methods of graph exploration that simply relate to the number of hops from the drug target and do not consider the density of the neighborhood. Comparing the local neighborhood with the unfiltered one is less meaningful since the density of the neighborhood is not considered. The above method allows for the effective discovery and display of the local interactome, independent of the neighborhood density. The output of the method is a list of genes / proteins that represent potential hubs for threshold polydrug interactions. This list can be utilized in biological / clinical sense.

[0060] The above method can be used by clinicians to screen patients with poor response to drug therapy or to screen individuals with suspected adverse drug reactions (ADRs). In difficult patient cases, clinicians can input a list of patient-specific drug protein targets and obtain a visual representation of the drug's effect at the protein level. The visual representation can also be used in a teaching context to help students visualize the polypharmacy network.

[0061] The above method may also be a useful tool in the drug research and development process. For example, the method may be used by researchers who want to study the effect of a new drug on patient biology, including potential biological effects, drug-protein / gene interactions, and potential contraindications in polypharmacy patients. This information can be used to help decide whether a new drug candidate should be passed through screening. The method may be used for patient selection in clinical trial design. Patients may be excluded from clinical trials if strong adverse effects are predicted.

[0062] In other embodiments, different devices may be used to perform different or part of the above processes. For example, a first device may be used to convert protein-protein interaction data, a second device may be used to perform random walk, a third device may be used to perform filtering, a fourth device may be used to perform scoring, a fifth device may be used to display graph data, and a sixth device may be used to perform further analysis. Any suitable combination of devices may be used.

[0063] Although particular circuits are described herein, in alternative embodiments, the functionality of one or more of those circuits may be provided by a single processing resource or other component, or the functionality provided by a single circuit may be provided by a combination of two or more processing resources or other components. Reference to a single circuit encompasses multiple components that provide the functionality of that circuit, whether or not such components are separate from one another. Reference to multiple circuits encompasses single components that provide the functionality of those circuits.

[0064] An embodiment provides a medical information processing device including a processing circuit, which receives knowledge about protein / gene interactions, estimates a first range based on a first drug and a second range based on a second drug based on the knowledge, and determines target proteins / genes that are affected by both the first drug and the second drug based on the first range and the second range.

[0065] The knowledge may be a knowledge graph, which uses nodes and edges to represent protein / gene interactions.

[0066] The processing circuit may further determine a first node and a second node corresponding to the first drug and the second drug based on the knowledge graph, and perform a first search and a second search from the first node and the second node to estimate the first range and the second range.

[0067] One embodiment provides a method to display local areas of drug target effect in gene or protein networks where nodes affected by multiple drug therapies are identified, colored by the strength of the interaction, and output as a list of nodes representing genes / proteins that can be used to characterize the potential biological mechanisms of polypharmacy.

[0068] The drug target may be treated as the starting node of a random walk that is used to explore the local neighborhood.

[0069] A function of the starting node dimension may be used to determine a threshold for filtering the count of visited nodes used to identify the drug's local neighborhood.

[0070] The interactions of sets of visited nodes generated (after filtering) from random walks with different starting nodes may be used to identify genes / proteins affected by multiple drug treatments.

[0071] The counts of visited nodes generated by the initiating node from different random walks (after filtering) may be averaged to generate an interaction score.

[0072] The colors of intersecting nodes in multiple neighborhoods may be scaled by the interaction score.

[0073] The list of genes / proteins identified as enriched in the vicinity of overlapping drugs may be used as input for downstream analyses (e.g., gene set enrichment analysis, Uniplot analysis) to infer biological mechanisms, side effects or adverse reactions.

[0074] The method of identifying local neighborhoods and overlaps between neighborhoods may not be limited to the biological domain, but can be extended to graphs from other domains / domains.

[0075] The drug combination analysis and visualization methods may be applicable to life forms other than Homo sapiens.

[0076] While certain embodiments have been described, these embodiments are presented for illustrative purposes only and are not intended to limit the scope of the invention. In fact, the novel methods and systems described herein may be embodied in a variety of other forms. Furthermore, various omissions, substitutions, and changes in the form of the methods and systems described herein may be made without departing from the spirit of the invention. The appended claims and their equivalents are intended to cover such forms and modifications as fall within the scope of the invention. [Explanation of symbols]

[0077] 10 equipment 12 Computing Equipment 16 Display Screens 18 Input Devices 22 Processing circuit 24 Graph Circuits 26 Random Walk Circuit 28 Filtering Circuit 30 Interaction Score Circuit 32 Display circuit 34 Analysis circuit

Claims

1. a storage unit that stores protein-protein interaction data representing the presence or absence of interactions between at least two or more proteins among a plurality of proteins; a receiver that receives designations of a first drug associated with a first set of proteins comprising a first collection of proteins and a second drug associated with a second set of proteins comprising a second collection of proteins; an estimation unit that estimates a first protein set based on a first drug and a second protein set based on a second drug, using the protein-protein interaction data stored in the storage unit and the first and second drugs designated by the receiving unit; A determination unit that determines whether there is at least one protein that is affected by both the first drug and the second drug based on the first protein set and the second protein set; A medical information processing device comprising:

2. The protein-protein interaction data includes a protein-protein interaction (PPI) network, which is a graph in which the proteins are nodes and the presence or absence of an interaction between at least two or more proteins among the plurality of proteins is represented by a connection between the nodes. The medical information processing device according to claim 1 .

3. interactions between the first set of proteins and the second set of proteins are used to predict whether there is at least one protein that is affected by both the first drug and the second drug; The medical information processing device according to claim 2 .

4. the first set of proteins includes a plurality of proteins corresponding to nodes visited during at least one random walk through the PPI network starting from a node corresponding to a protein related to the first drug; The second set of proteins includes a plurality of proteins corresponding to nodes visited during at least one random walk through the PPI network starting from a node corresponding to a protein related to the second drug. The medical information processing device according to claim 3 .

5. a node corresponding to the protein associated with the first drug is a target of the first drug; A node corresponding to the protein associated with the second drug is a target of the second drug. The medical information processing device according to claim 4 .

6. each of the nodes corresponding to the plurality of proteins in the first set of proteins has an associated count value equal to the number of times the node corresponding to that protein is visited during the at least one random walk for the first drug; each of the nodes corresponding to the plurality of proteins in the second set of proteins has an associated count value equal to the number of times the node corresponding to that protein is visited during the at least one random walk for the second drug. The medical information processing device according to claim 5 .

7. After estimating the first set of proteins and the second set of proteins, and before determining a node corresponding to the at least one protein that is affected by both the first drug and the second drug, and a normalization unit that normalizes the association count values ​​of each of the nodes corresponding to the plurality of proteins in the first set of proteins and the second set of proteins based on the number of random walks for the first drug and the second drug, respectively. The medical information processing device according to claim 6 .

8. After normalizing the association count values ​​and prior to determining a node corresponding to the at least one protein that is affected by both the first drug and the second drug, a filtering unit configured to filter the first protein set and the second protein set based on the normalized count value using a first threshold for the first protein set and a second threshold for the second protein set; The medical information processing device according to claim 7 .

9. the first and second thresholds are calculated based on a dimension of a node corresponding to the protein associated with the first drug and the protein associated with the second drug, respectively, the dimension of the node being equal to the number of edges entering the node; The medical information processing device according to claim 8 .

10. The first threshold and the second threshold are expressed by the formula 0.1×0.95 ｘ +0.1, where x is the dimension of the node corresponding to the protein associated with the first drug and the second drug, respectively. The medical information processing device according to claim 9 .

11. A calculation unit is further provided to calculate an interaction score for each of the nodes corresponding to the plurality of proteins in the first protein set and the second protein set based on an average value of the normalized association count values ​​of the nodes corresponding to the proteins in the first protein set and the second protein set. The medical information processing device according to claim 10.

12. a display control unit that displays a subgraph of the PPI network, the subgraph including the first protein set and the second protein set; the display control unit visually indicates whether a protein in the subgraph belongs to the first protein set, the second protein set, or both the first protein set and the second protein set. The medical information processing device according to claim 11.

13. The display control unit visually indicates the interaction score of at least one protein in the subgraph. The medical information processing device according to claim 12.

14. The interaction score is represented by a color scale. The medical information processing device according to claim 13.

15. The method further includes an analysis unit that analyzes the at least one protein affected by both the first drug and the second drug to infer at least one of a biological mechanism, a side effect, or an adverse reaction related to the treatment using the first drug and the second drug; The medical information processing device according to claim 14.

16. The biological mechanism is inferred by performing gene set enrichment analysis, pathway analysis, or Uniprot analysis; The medical information processing device according to claim 15.

17. the first drug is being administered in a clinical trial and the second drug is prescribed to a potential subject in the clinical trial; The analysis unit further determines whether the subject candidate should be excluded from the clinical trial and / or whether the subject candidate should be classified into a subgroup within a cohort of clinical trial participants based on at least one of the biological mechanism, side effects, or adverse reactions. The medical information processing device according to claim 15.

18. the first drug and the second drug relate to a treatment regimen for the patient; The analysis unit further determines whether one of the first drug or the second drug should be removed from the treatment plan and / or whether a dosage of the first drug or the second drug should be changed based on at least one of the biological mechanism, the side effect, or the adverse reaction. The medical information processing device according to claim 15.

19. the first drug is a new drug candidate and the second drug is a commonly prescribed drug; The analysis unit further determines whether the new drug candidate should proceed to a drug development stage based on at least one of the biological mechanism, the side effect, or the adverse reaction. The medical information processing device according to claim 15.

20. An information processing method by a medical information processing device, comprising: A receiving step of receiving protein-protein interaction data; an estimation step of estimating a first protein set based on a first drug and a second protein set based on a second drug using the protein-protein interaction data; determining whether there is at least one protein that is affected by both the first drug and the second drug based on the first set of proteins and the second set of proteins; An information processing method comprising:

Images