INTERACTIVE PRECISION MEDICINE EXPLORER FOR GENOMIC ABBERATIONS AND TREATMENT OPTIONS

Abstract

A data-driven integrative visualization system and method for summarizing and presenting genomic aberrations, their drug responses and multi-omic data of a patient, is disclosed. Specifically, a method for displaying genomic aberrations and multi-omic data of a patient in an interactive tool which allows the medical practitioner to access underlying supporting biologic and scientific evidence from relevant knowledge bases through a set of graphical interactions, is described. The method comprises the steps of obtaining and inputting multi-omic data of a patient or cohorts, identifying genomic aberrations and their drug responses, and displaying this information in a first level interactive classical/circular ideogram in one or multiple layers on a GUI, from which the user can access and view further information on the gene and molecular levels. The system provides an improved process of integrative analysis on a patient's multi-omic data for effective treatment planning.

Claims

1. A computer-implemented method for summarizing and presenting patient-specific multi-omic data in a multilayered format, the method comprising: a computing device with a graphical user interface, determining a dataset of files containing patient information by obtaining genomic aberration and other omics data from a patient and storing said data on a non-transitory computer readable storage medium; determining selection criteria based on the patient dataset; inputting patient-specific data, by a user interface, onto a processor configured to receive said patient-specific data, selecting a cohort of samples based on user-defined demographic and phenotypic criteria from a repository of patient or healthy samples, and inputting said demographic and phenotype criteria into said computing device through said graphical user interface; extracting said cohort genomic aberration and omics data for comparison with the patient of interest based on said demographic and phenotype criteria and inputting said cohort genomic aberration and omics data, by a user interface, onto a processor configured to receive said cohort genomic aberration and omics data; annotating said patient-specific genomic aberration and omics data in a first layer of said multilayered format, using internal/external knowledge bases, which include information such as mutation impact, population allele frequency, disease association with model of inheritance and drug response; filtering said patient-specific genomic aberrations and omics data based on user-defined criteria, such as chromosome regions, genes and variant type/function/impact/population allele frequency; and displaying said patient-specific genomic aberration and omics data in said interactive multi-level format, wherein said multilayer format comprises; said first layer, said first layer comprising an interactive chromosomal view that summarizes all the clinically relevant or actionable genomic aberrations of said patient by marking them on the genome coordinates, including known drug responses associated with a particular mutation/gene marked next to the mutation/gene accordingly, said first layer further comprising; a first sub-layer comprising an ideogram view where chromosomes are arranged in a circular format; a second sub-layer comprising an ideogram view where each chromosome in said first sub-layer is separately displayed in a schematic; a second layer comprising an interactive intergenic genomic scale where multiple genes are displayed with their expression levels indicated by color; a third level comprising an interactive genic scale, depicting the structure and functional blocks within a gene, omics data such as methylation levels and gene/exon expression, the 3D protein structure (ribbon plot), with mutations marked and including general information about said gene; and a fourth level, comprising a molecular scale displaying the molecular sequence and its detailed annotations, such as the nucleotide sequence of the reference genome, the corresponding amino acid sequence in the protein-coding regions, nucleotide/amino acid changes caused by the mutations, exon/gene expression and methylation levels of CpG sites, ChIP-Seq data for histone modification.

2. The method of claim 1, wherein the mulilayered format is a circular or linear multilayered format.

3. The method of claim 1, wherein said obtaining genomic aberration and other omics data from a patient comprises the collection of tissue and blood samples from said patient, performing next-generation sample preparation and DNA/RNA seqeuncing, read alignment and culling of variants and gene expressions.

4. The method of claim 1, wherein said second layer further comprises additional data tracks to add more details, such as methylation, chromatin immunoprecipitation sequencing and assay data which may improve the functional view of genomic aberrations.

5. A non-transitory computer readable storage medium tangibly encoded with computer-executable instructions, that when executed by a processor associated with computing device having a graphical user interface, cause the device to carry out the steps of the method as defined in claim 1.

6. A computer program product, comprising a computer-readable code to be executed by one or more processors when retrieved from a non-transitory computer-readable medium, the computer-readable program code including instructions to: determine a dataset of files containing patient information by obtaining genomic aberration and other omics data from a patient and storing said data on a non-transitory computer readable storage medium; receive selection criteria by a user through a graphical user interface, said selection criteria determined by said user based on said patient dataset, and input said patient-specific data, onto a processor configured to receive said patient-specific data, select a cohort of samples based on user-defined demographic and phenotypic criteria from a repository of patient or healthy samples, and input said demographic and phenotype criteria into said computing device through said graphical user interface; extract said cohort genomic aberration and omics data for comparison with the patient of interest based on said demographic and phenotype criteria and inputting said cohort genomic aberration and omics data, by a user interface, onto a processor configured to receive said cohort genomic aberration and omics data; annotate said patient-specific genomic aberration and omics data, using internal/external knowledge bases, which include information such as mutation impact, population allele frequency, disease association with model of inheritance and drug response; filter said patient-specific genomic aberrations and omics data based on user-defined criteria, such as chromosome regions, genes and variant type/function/impact/population allele frequency; and display said patient-specific genomic aberration and omics data in said interactive multi-level format, wherein said multilayer format comprises; a first layer comprising an interactive chromosomal view that summarizes all the clinically relevant or actionable genomic aberrations of said patient by marking them on the genome coordinates, including known drug responses associated with a particular mutation/gene marked next to the mutation/gene accordingly, said first layer further comprising; a first sub-layer comprising an ideogram view where chromosomes are arranged in a circular format; a second sub-layer comprising an ideogram view where each chromosome in said first sub-layer is separately displayed in a schematic; a second layer comprising an interactive intergenic genomic scale where multiple genes are displayed with their expression levels indicated by color; a third level comprising an interactive genic scale, depicting the structure and functional blocks within a gene, omics data such as methylation levels and gene/exon expression, the 3D protein structure (ribbon plot), with mutations marked and including general information about said gene; and a fourth level, comprising a molecular scale displaying the molecular sequence and its detailed annotations, such as the nucleotide sequence of the reference genome, the corresponding amino acid sequence in the protein-coding regions, nucleotide/amino acid changes caused by the mutations, exon/gene expression and methylation levels of CpG sites, ChIP-Seq data for histone modification.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0023] The methods according to the invention will now be described in more detail with regard to the accompanying figures. The figures showing ways of implementing the present invention and are not to be construed as being limiting to other possible embodiments falling within the scope of the attached claims.

[0024] The methods according to the invention will now be described in more detail with regard to the accompanying figures. The figures showing ways of implementing the present invention and are not to be construed as being limiting to other possible embodiments falling within the scope of the attached claims.

[0025] FIG. 1 is a high-level flow diagram that gives an overview of the computational steps and data sources involved in the processing and presentation of multi-omics data in our Precision Medicine Explorer;

[0026] FIG. 2 is a flow diagram that shows the detailed steps and components for the two main functionalities of the Precision Medicine Explorer: (a) filtering and searching of variant and omics data, and (b) data visualization and exploration;

[0027] FIG. 3 is a circular ideogram view of Level 1, displaying the genomic aberrations of a patient and their associated drug responses;

[0028] FIG. 4 is a classical ideogram view of Level 1, displaying the genomic aberrations of a patient and their associated drug responses;

[0029] FIG. 5 is a view of Level 2, an intergenic genomic scale where multiple genes are displayed with their expression levels indicated by color;

[0030] FIG. 6 is a view of Level 3, a genic scale where the methylation and gene/exon expression levels are indicated by color;

[0031] FIG. 7 is a view of Level 4, showing the nucleotide sequence, amino acid sequence and methylation level;

[0032] FIG. 8 is a schematic view of multiple layers for the comparison of genomic aberrations and treatment options across different patients and cohorts;

[0033] FIG. 9 illustrates a circular ideogram showing genes with associated keywords for searching purposes; and

[0034] FIG. 10 is a 3D view of our Precision Medicine Explorer.

DETAILED DESCRIPTION OF THE INVENTION

[0035] The present invention provides a system and method for summarizing and presenting genomic aberrations, their drug responses and multi-omic data of a patient, by displaying genomic aberrations and multi-omic data of the patient in an interactive classical/circular ideogram format which allows the medical practitioner to access underlying supporting biologic and scientific evidence from relevant knowledge bases through a set of graphical interactions. The present invention is described in further detail below with reference made to FIGS. 1-10. Referring now to the figures, FIG. 1 is a flow diagram that shows an overview of the computational steps and data sources involved in the processing and presentation of multi-omics data in the Precision Medicine Explorer. Similarly, FIG. 2 is a flow diagram showing the steps and components for two main functionalities of the Precision Medicine Explorer: (a) filtering and searching of variant and omics data, and (b) data visualization and exploration. FIGS. 1 and 2 illustrate an embodiment of the invention which provides a system and a method for obtaining and organizing relevant patient-specific genomic information, presenting such information on a visual display that is a circular or linear multilayered interactive plot, usually displayed on a graphical user interface. The method entails obtaining genomic aberration and other omics data from a patient and storing that data on a non-transitory computer readable storage medium. One of the common processses for data generation involves the collection of tissue and blood samples from the patient, performing next-generation sample preparation and DNA/RNA seqeuncing, read alignment and culling of variants and gene expressions, etc. Optionally, a user could select a cohort of samples based on demographic and phenotypic criteria, defined by the user, from a repository of patient or healthy samples, and extracting their genomic aberration and omics data for comparison with the patient of interest. The genomic aberration and omics data are annotated the using internal/external knowledge bases (FIG. 1), which include information such as mutation impact, population allele frequency, disease association with model of inheritance, drug response, etc. The genomic aberrations and omics data are then filtered based the on user-defined criteria (FIG. 2), such as chromosome regions, genes, variant type/function/impact/population allele frequency, etc.

[0036] With a computing device having a graphical user interface, the genomic aberration and omics data are then displayed in an interactive multi-level format. At Level 1 of the method and system for displaying patient-specific genomic data and genomic aberrations, all the clinically relevant or actionable aberrations of a patient are summarized by marking them on the genome coordinates (see FIGS. 3 and 4). If there are any drug responses associated with a mutation/gene, they are marked next to the mutation/gene accordingly. Within this level, there are at least three possibilities: Level 1Acircular ideogram view, where chromosomes are arranged in a circular layout, and Level 1Bclassical ideogram view, where each chromosome is separately displayed in a schematic that uses the familiar karyogram representation. According to an embodiment of the present invention, FIG. 3 is an interactive circular ideogram view at Level 1A and FIG. 4 is an interactive classical ideogram view at Level 1B. Both views are displayed by the computer on a graphical user interface (GUI). Users are able to switch from one view to the other by interacting with the GUI. A third representation would be linear horizontal representation which contains the same layers on a horizontal axis stacked on top of each other. The user accesses the chromosonal sub-levels by clicking on or selecting a mutation or gene in the GUI at Level 1, and by similarly selecting a region on the chromosome, the user can zoom in to view and explore data at different levels.

[0037] FIG. 5 illustrates the second level, Level 2 of the embodiment of FIGS. 3 and 4. Level 2 is an interactive intergenic genomic scale where multiple genes are labeled by their gene symbols and displayed with their expression levels indicated by color, along with any relevant/targetable mutations and their corresponding drug options. The user may add data tracks, such as methylation, ChIP-Seq, NET-Seq and ATAC-Seq data to incorporate more details to complete the functional picture of the genomic aberrations (or the lack thereof).

[0038] By selecting a specific gene at Level 2, the user is directed to Level 3 of this embodiment, as shown in FIG. 6. Level 3 is a genic scale where the methylation, gene/exon expression levels and other omics data of the gene selected at Level 2 are indicated by color or other attributes, along with any relevant/targetable mutations and their corresponding drug options. Further data tracks as already mentioned can be added to incorporate more details. The reason for this multi-track representation is to be able to make inferences about the functional impact of the genomic aberrations. With the multi-track representation, we want to support event-based querying where multiple events for the SAME gene may affect the ability of the gene to drive a tumor. We need to enable better association of drugs (e.g., ALK fusions to a targeted drug named Crizotinibthat may inhibit a gene with an activating mutation or avoid therapies that may be directed at inactivated genes). Level 3 also includes general information, included at the top for reference, about the gene selected, its functional blocks (promoter, transcription start/stop site, exon, intron, etc.) and 3D structure (ribbon plot) with mutations being marked.

[0039] Similarly, the user accesses Level 4, as seen in FIG. 7, by selecting a specific gene at Level 3. Level 4 comprises information about the gene at the molecular level where the nucleotide sequence, amino acid sequence and methylation level are displayed. As aforementioned, data tracks can be added to incorporate more details, such as the nucleotide and amino acid changes caused by the mutations and create impression about the functional impact of the genomic aberrations. The important information that the user needs to visualize is if there is activating effect of the genomic aberrations: mutations/fusions on gene expression and downstream targets of that gene, or inactivating effect. By bringing this information together within the single visual framework, we bring the evidence so that the clinician is able to make decisions.

Mutations and Drug Response

[0040] To enhance data presentation, the invention employs different symbols to represent different types of aberrations and drug/clinical trial associations with their levels of significance indicated by properties such as color and size, as can be seen in FIG. 3. An example of a scheme of data representation is as follows: [0041] 1. Single nucleotide variant (SNV)A circle symbol for missense; A square symbol for nonsense [0042] 2. InsertionA downward facing triangle symbol [0043] 3. DeletionA cross or X symbol [0044] 4. Fusionan arc joining the donor and acceptor genes [0045] 5. Copy number variationa plus sign with the number of copies at the top right as a superscript (.sup.3) [0046] 6. Over- or under-expression: An up arrow for over-expression and a down arrow for under-expression and the differential expression in log 2 fold change can be labelled at the top right [0047] 7. Variant classification, such as pathogenic, likely pathogenic, unknown significance (VUC), likely benign and benign, can be represented by different colors of the mutation symbol [0048] 8. A combined pathogenicity score based on multiple algorithms can be marked at the top right of the mutation symbol, e.g. a square symbol with a 0.9 superscript denotes a nonsense SNV with a combined pathogenicity score of 0.9 [0049] 9. Additional annotations, such as frame shift (FS), splice site (SS), nonsense mediated decay (NMD), etc., can be labelled at the top right of the mutation symbol, e.g., a downward facing triangle with a FS superscript denotes a frameshift insertion [0050] 10. Each mutation is precisely labelled by using HGSV nomenclature (hgvs.org/mutnomen/). Additional nomenclatures may be used. [0051] 11. Explicit reference to activating or inactivating genomic aberration is made in the UX. This information can be inferred based on 1) the pathogenicity score, or 2) manually curated information that is assembled based on previous experimental and published findings. [0052] 12. Drugs option is represented by a pill [0053] (a) Drug options with increased response is denoted by a pill with an up-arrow in green [0054] (b) Drug options with decreased response is denoted by a pill with a down-arrow in blue [0055] (c) Drug options with severe side effects is denoted by a pill with an exclamation mark in red [0056] (d) The best level of evidence among the drug options is indicated by the fill level [0057] (e) Number of drug options belonging to the same category is labelled next to the symbol [0058] (f) For example, a green pill with a superscript including a chevron and 4 means there are four drugs with increased response associated with a mutation, whereas a red pill with a superscript including an exclamation point and 2 indicates there are two drugs with severe side effects, or if there is orange color, that means that the genomic aberration is associated with a resistant marker. [0059] 13. Clinical trial is represented by a test tube, with the number of trials stated at the top right and the level of evidence, if available, indicated by the fill level, e.g., a test tube with a 2 superscript indicates there are two clinical trials associated with a mutation [0060] 14. Symbols of the genes carrying clinically relevant mutations are marked at their genomic positions, with their mutations grouped and listed next to them. [0061] 15. The strand of a gene can be indicated by an arrow: .fwdarw.right or clockwise for forward strand, left or anti-clockwise for reverse strand [0062] The choice of symbols is not restricted to those illustrated in the above examples.

Interactions

[0063] To enable the seamless navigation to the patient's multi-omic data at different levels of details and quick access to relevant information from different knowledge bases, the Precision Medicine tool of this invention is highly interactive and user friendly. The set of supported user interactions include, but are not limited to, the following: [0064] 1. Toggle between the classical ideogram, circos and horizontal (linear) views of the genome [0065] 2. Zoom in/out to different data levels by using a zoom slider, selecting a region on the genomic scale or directly specifying a gene, locus or the start and end chromosome positions [0066] 3. Rearrange the layout of chromosomes in the ideogram, rotate the circular ideogram or navigate to nearby regions by swiping [0067] 4. Select the inclusion/exclusion criteria for aberrations to be displayed, e.g., by specifying the types of mutations and the chromosome regions or gene subsets [0068] 5. Import and display additional tracks of data and annotation, e.g., mutational density [0069] 6. Select and display the omic data of one or more individual patients and cohorts in multiple layers [0070] 7. Hover on any color-scaled data, such as gene expression and methylation levels, and display the actual numerical values [0071] 8. Select a nucleotide, amino acid or mutation and their locations will be marked on the corresponding gene and 3D protein structure (see FIG. 7) [0072] 9. Rotate and zoom in/out the 3D protein structure [0073] 10. Select and display genes, mutations or other data associated with a concept or keywords [0074] 11. Access to more detailed information related to an object or component by clicking/hovering on it, or right-clicking and then selecting from a pop-up menu: [0075] (a) Mutationschromosome/transcript/protein positions, amino acid changes, genotypes for germline mutations or variant allele fraction for somatic mutations, allelic balance, number of reads (for sequencing data), call quality (e.g., phred score), function (nonsense, missense, frame-shift, splice-site, NMD, etc.), variant classification, population allele frequencies, pathogenicity scores, related publications, etc. [0076] (b) Drugs optionsa list of drug names, their levels of evidence, supporting publications, etc. [0077] (c) Clinical trialsa list of clinical trials, conducting institutions, short description, etc. [0078] (d) Gene-level detailsfull name of the gene, brief description, genomic size, number of exons, pathway/disease/drug associations, summary of patient-specific data such as gene expression and list of mutations, etc. [0079] (e) Information on the functional impact of the genomic aberrations [0080] 12. Include information on activating or deactivating effect of the genomic aberration [0081] 13. Include hyperlinks for terms, such as gene symbols and drug names, whenever necessary for further information

Comparison of Multiple Samples and Cohorts

[0082] In a further embodiment, users can choose to display the omic data of a patient or a cohort of patients in multiple layers of the visual representation in the Precision Medicine Explorer for side-by-side comparison. See FIG. 8. The genome coordinates of each layer of ideogram should be coherently aligned with other layers. Users are able to add/remove/combine/change the order of multiple layers and explore any one of them in detail through all interactions applicable to a single layer. For example, FIG. 8 schematically illustrates a stack of circular layers for the comparison of genomic aberrations and treatment options across different patients and cohorts. Each layer presents the data of one patient or a cohort consisting of many patients. In this example, the genomic aberrations of the current patient are summarized in the top circle, and compared against individuals (the genomic profile of the patient's mother and sister), cohorts that have prognostic information (Luminal A, Luminal B, HER2+, Basal) and BRCA mutations from ClinVar.

Presentation Filters for Genomics Aberrations

[0083] In genomics, it is customary to offer multiple filtering options to the user for each of the types of genomic aberrations. Within this embodiment, the goal is to associate the genomic aberrations to key evidence for treatment planning. In any embodiment of this invention, users can determine what data is to be presented in one or multiple layers of ideogram by applying a combination of filters that include but are not limited to the following: [0084] 1. Chromosome regions, e.g., chr1:1000000-5000000, chrX, etc. [0085] 2. Genes [0086] (a) List of specific genes [0087] (b) Biological concepts or terms that are associated with gene subsets, e.g., oncogene, suppressor, transcription factor, signaling pathways such as ER, PR, Wnt, PI3K, MAPK, etc. [0088] (c) Significantly mutated genes (SMGs)users can select the methods for computing the SMGs and their parameters [0089] (d) Mutation burdenusers can specify the number and types of mutations that a gene needs to carry to be included for display [0090] (e) Genes which have associated drug response information: [0091] 3. Variant Type: single nucleotide variants (SNVs), short insertions/deletions (indels), copy number variations (CNVs), gene fusions, over expression, under expression, etc. [0092] 4. Variant Function: synonymous, missense, nonsense, nonsense mediated decay (NMD), frameshift, splice site, promoter, etc. [0093] 5. Variant Impact [0094] (a) Therapeutic/Pharmacogeneticvariants with available drug options. Genomic aberrations have associated drug response information: 1) resistance association that depicts that the mutation is associated with resistance within a certain indication and 2) response association that depicts that the mutation is associated with likely response to the drug within a certain indication (e.g., response to First generation Tyrosine kinase inhibitor) [0095] (b) Classificationcan be based on the ACMG guidelines, i.e., Classes 1-5 for somatic mutations, and for germline mutations pathogenic, likely pathogenic, uncertain significance, likely benign or benign [0096] (c) Pathogenicity predictionusers can choose a combination of algorithms and their thresholds, which are joined together by and/or operators [0097] 6. Variant Frequency in Ethnic Groupsminor allele frequency thresholds in one or more ethnic groups (white/black/Asian/all), with the conditions joined by and/or operators [0098] 7. Variant Frequency in Samples/Cohortsfor each sample/cohort, users can specify the range of the number/frequency of a variant or their carriers, with the conditions joined by and/or operators
Depending on the purpose of the application, e.g. diagnostic, therapy selection or research, different default filter settings can be applied so that only the relevant information is shown.
Search by Keywords with Autocomplete Suggestions

[0099] Users can show the genes or other information associated with a keyword on the ideogram by typing the keyword in a search box with autocomplete functionality. The search term can be a gene symbol, signaling pathway, disease, drug, or biological concept such as oncogene/suppressor, etc. Users can also search for a combination of these terms concatenated by logical operators, such as ,/OR, &/AND, etc. Once the data related to the search term(s) are retrieved from the databases, they are displayed on the same or a separate ideogram (see FIG. 9). The search results can be highlighted and presented in such a way that they are distinguishable from the patient's primary data. Search history is tracked to let users select the results of one or more searches for quick viewing and comparison.

[0100] Referring to FIG. 9, a keyword search allows genes associated with a term to be looked up and displayed in the ideogram. In this example, all genes in the ER Pathway are shown.

[0101] To make the zoom-in or zoom-out transition look continuous and smooth, and enhance the navigation and user experience, our Precision Medicine Explorer includes a 3D option that enables users to view the chromosome layouts from different visual perspectives (see FIG. 10).

Association with Evidence for Key Findings

[0102] One essential functionality of our Precision Medicine Explorer is to display the drugs/treatments with their known predicted/experimental/clinical responses (increased/decreased) or clinical trial options associated with patient-specific data, such as genomic aberrations, up/down-regulated gene expressions, abnormal methylation levels or other omics anomalies with supporting evidence, which can be further explored through user interactions. For example, the gene mutation BRAF V600E is known for increased sensitivity to Vemurafenib in Melanoma, and the gene mutation EGFR T790M for resistance to tyrosine kinase inhibitors. Such associations can be looked up from local/external knowledge bases such as the Catalogue Of Somatic Mutations In Cancer (COSMIC) Database, the Mutations and Drugs Portal (MDP), the Cancer Drug Resistance Database (CancerDR), the Drug Gene Interaction Database (DGIdb) and ClinicalTrials.gov. Additional information on the drugs, such as the side effects, toxicity, mechanism of action, interactions with other drugs and the supporting scientific evidence can be accessed for display. Gathering, summarizing and presenting such information in one single tool can facilitate the design of combinatorial therapy and caution the potential threats of certain drug combinations that should be avoided.

Example

[0103] As a use case example, our Precision Medicine Explorer is used for examining the omic data of an ER+ breast cancer patient. From the top-level view, the oncologist gets a genomic overview of the clinically relevant mutations carried by the patient and the available drug options. As expected, an overexpression of the ESR1 gene was reported with a list of drug options consisting of ER inhibitors. If the oncologist wants to further examine the expression levels of the genes in the ER pathway, she would then add a track for gene expression and filter for a pre-defined panel of ER pathway genes. After inspecting the expression values, she confirmed whether the patient has a hyperactive ER pathway, which could be effectively suppressed by ER inhibitors. She also noticed that the patient carries a known pathogenic mutation in the PIK3CA gene. She clicks on the mutation and checks the allele frequency, function, pathogenicity, call quality, related publications, among other details, and confirmed that the mutation served as a good prognostic biomarker for favorable therapeutic response of PIK3CA inhibitors. After comparing the clinical evidence and possible side effects of the drug options, she decided to administer the two inhibitors with the strongest clinical evidence respectively for suppressing the activities of ER and PIK3CA in combination for treating the patient. Our Precision Medicine Explorer significantly improved the workflow of an oncologist in performing integrative analysis on a patient's omic data for treatment planning.