Joshua F. McMichael
Research Subjects
Joshua F. McMichael has a total of 350 co-authors. Here are the top co-authors by number of shared publications:
- Malachi Griffith
Washington University
37 shared publications - Obi L. Griffith
Washington University School of Medicine
35 shared publications - Kilannin Krysiak
Department of Pathology & Immunology, Washington University School of Medicine , Saint Louis, MO 63110
23 shared publications - Christopher A. Miller
Division of Oncology, Department of Medicine, Washington University School of Medicine, St. Louis, Missouri
19 shared publications - Cody A. Ramirez
Department of Developmental Neurobiology, St. Jude Children’s Research Hospital, Memphis, USA
19 shared publications - Todd A. Fehniger
Division of Oncology, Department of Medicine, Washington University School of Medicine, St. Louis, Missouri
19 shared publications
Latest Publications
- Open Access
- 20 November 2026
- Published by Oxford University Press (OUP) in Nucleic Acids Research
CIViC 2026: expanding cancer variant curation with AI, genome-wide biomarkers, and genomic regions
The CIViC (Clinical Interpretations of Variants in Cancer) knowledgebase (https://civicdb.org) remains one of the few free, unrestricted knowledgebases for the distribution of somatic cancer variant interpretations. CIViC leverages public curation and expert moderation to address the bottleneck problem of cancer variant interpretation for patient care. As the curation platform for ClinGen Somatic Cancer, the CIViC knowledge model is informed by multiple experts across disparate fields in cancer and, in turn, acts as a model for cancer variant classification and representation. In order to capture a wider range of genomic variations, several updates to CIViC’s data and curation model have been made with a focus on maintaining familiar workflows and user interface layouts. A new top-level entity called the Feature represents Genes, Fusions, Regions for large-scale structural variations, and Factors for genome-wide biomarkers. Accompanying Variants capture information unique to each genomic event. Molecular Profiles enable the curation of complex combinations of these Variants. CIViC has also developed a system for approval and submission of Assertions to ClinVar, guidance for the use of AI in curation and moderation of content, and additional capabilities such as source retraction monitoring and an advanced query builder.
- 20 November 2026
- Published by Oxford University Press (OUP) in Nucleic Acids Research
Open AccessCIViC 2026: expanding cancer variant curation with AI, genome-wide biomarkers, and genomic regions
The CIViC (Clinical Interpretations of Variants in Cancer) knowledgebase (https://civicdb.org) remains one of the few free, unrestricted knowledgebases for the distribution of somatic cancer variant interpretations. CIViC leverages public curation and expert moderation to address the bottleneck problem of cancer variant interpretation for patient care. As the curation platform for ClinGen Somatic Cancer, the CIViC knowledge model is informed by multiple experts across disparate fields in cancer and, in turn, acts as a model for cancer variant classification and representation. In order to capture a wider range of genomic variations, several updates to CIViC’s data and curation model have been made with a focus on maintaining familiar workflows and user interface layouts. A new top-level entity called the Feature represents Genes, Fusions, Regions for large-scale structural variations, and Factors for genome-wide biomarkers. Accompanying Variants capture information unique to each genomic event. Molecular Profiles enable the curation of complex combinations of these Variants. CIViC has also developed a system for approval and submission of Assertions to ClinVar, guidance for the use of AI in curation and moderation of content, and additional capabilities such as source retraction monitoring and an advanced query builder.
- Open Access
- 19 November 2026
- Published by Oxford University Press (OUP) in Nucleic Acids Research
DGIdb in 2026: an interoperable resource for the druggable genome
The Drug–Gene Interaction Database (DGIdb; https://dgidb.org) is an open-access aggregate knowledgebase that normalizes and centralizes drug–gene interaction and gene annotation data from diverse therapeutic and biomedical sources. Since its previous update, DGIdb has expanded its data content and added new tooling to better support programmatic access and local deployment. This release updates existing sources and incorporates three new sources of interaction data, including annotation data from the Molecular Oncology Almanac, the PRISM Repurposing Dataset, and the Broad Institute Drug Repurposing Hub. To streamline data access for computational workflows, we have developed dgiPy and dgiR, Python and R client libraries that provide standardized access to the DGIdb API. We additionally now provide versioned Docker images and a public data repository containing current and historical DGIdb data releases in TSV and SQL formats. Together with continued improvements to the web interface, these new resources expand FAIR, transparent, and reproducible access to drug-gene interaction data in the bioinformatics community.
- 19 November 2026
- Published by Oxford University Press (OUP) in Nucleic Acids Research
Open AccessDGIdb in 2026: an interoperable resource for the druggable genome
The Drug–Gene Interaction Database (DGIdb; https://dgidb.org) is an open-access aggregate knowledgebase that normalizes and centralizes drug–gene interaction and gene annotation data from diverse therapeutic and biomedical sources. Since its previous update, DGIdb has expanded its data content and added new tooling to better support programmatic access and local deployment. This release updates existing sources and incorporates three new sources of interaction data, including annotation data from the Molecular Oncology Almanac, the PRISM Repurposing Dataset, and the Broad Institute Drug Repurposing Hub. To streamline data access for computational workflows, we have developed dgiPy and dgiR, Python and R client libraries that provide standardized access to the DGIdb API. We additionally now provide versioned Docker images and a public data repository containing current and historical DGIdb data releases in TSV and SQL formats. Together with continued improvements to the web interface, these new resources expand FAIR, transparent, and reproducible access to drug-gene interaction data in the bioinformatics community.
- Open Access
- 24 August 2026
- Published by Oxford University Press (OUP) in Nucleic Acids Research
- Vol. 54 (16)
CompBio and MIRaS—a multi-omic analysis platform built on a memory-based intelligence engine
As molecular and cellular technologies have advanced, the need to analyze and interpret the resulting vast and multi-modal data into actionable intelligence has become a rate-limiting factor in scientific advancement. While traditional knowledgebase-pathway tools and LLM-based AI tools provide a degree of support in data interpretation, both exhibit known limitations. Presented here is the CompBio multi-omic analysis platform built upon Memory-based Intelligent Reasoning System (MIRaS), an associative reasoning engine grounded in episodic and semantic memory, capable of inferring biological relationships directly from stored knowledge rather than from model training. The MIRaS system is not dependent on human-curated pathway knowledgebases, provides statistical rigor, does not suffer from hallucination, and produces fully traceable results, representing a number of significant advancements. Furthermore, substantial effort has been placed in the computer–human knowledge transfer components of the system to enable efficient interaction, interpretation, and learning. The system has undergone years of extensive testing, validation, and evolution in omics-based analysis with many associated publications in translational and human subjects research. These publications have covered numerous areas such as nutrition, immunology, infectious disease, cardiovascular disease, and degenerative diseases with many experimentally verified findings. The CompBio-MIRaS system described and assessed here is freely available to all academic and non-profit institutions.
- 24 August 2026
- Published by Oxford University Press (OUP) in Nucleic Acids Research
- Vol. 54 (16)
Open AccessCompBio and MIRaS—a multi-omic analysis platform built on a memory-based intelligence engine
As molecular and cellular technologies have advanced, the need to analyze and interpret the resulting vast and multi-modal data into actionable intelligence has become a rate-limiting factor in scientific advancement. While traditional knowledgebase-pathway tools and LLM-based AI tools provide a degree of support in data interpretation, both exhibit known limitations. Presented here is the CompBio multi-omic analysis platform built upon Memory-based Intelligent Reasoning System (MIRaS), an associative reasoning engine grounded in episodic and semantic memory, capable of inferring biological relationships directly from stored knowledge rather than from model training. The MIRaS system is not dependent on human-curated pathway knowledgebases, provides statistical rigor, does not suffer from hallucination, and produces fully traceable results, representing a number of significant advancements. Furthermore, substantial effort has been placed in the computer–human knowledge transfer components of the system to enable efficient interaction, interpretation, and learning. The system has undergone years of extensive testing, validation, and evolution in omics-based analysis with many associated publications in translational and human subjects research. These publications have covered numerous areas such as nutrition, immunology, infectious disease, cardiovascular disease, and degenerative diseases with many experimentally verified findings. The CompBio-MIRaS system described and assessed here is freely available to all academic and non-profit institutions.
- Open Access
- 22 August 2026
- Published by Oxford University Press (OUP) in Bioinformatics
- Vol. 42 (9)
Searching the druggable genome using large language models
Summary: The druggable genome encompasses the genes that are known or predicted to interact with drugs. The Drug-Gene Interaction Database (DGIdb) provides an integrated resource for discovering and contextualizing these interactions, supporting a broad range of research and clinical applications. DGIdb is currently accessed through structured web interfaces and API calls, requiring users to translate natural-language questions into database-specific query patterns. To allow for the use of DGIdb through natural language, we developed the DGIdb Model Context Protocol (MCP) server, which allows large language models (LLMs) access to up-to-date information through the DGIdb API. We demonstrate that the MCP server improves an LLM’s ability to answer questions requiring accurate, up-to-date biomedical knowledge drawn from structured external resources. Availability and implementation: The DGIdb MCP server is detailed at https://github.com/dgidb/dgidb-mcp-server and includes instructions for accessing the server through the Claude desktop app.
- 22 August 2026
- Published by Oxford University Press (OUP) in Bioinformatics
- Vol. 42 (9)
Open AccessSearching the druggable genome using large language models
Summary: The druggable genome encompasses the genes that are known or predicted to interact with drugs. The Drug-Gene Interaction Database (DGIdb) provides an integrated resource for discovering and contextualizing these interactions, supporting a broad range of research and clinical applications. DGIdb is currently accessed through structured web interfaces and API calls, requiring users to translate natural-language questions into database-specific query patterns. To allow for the use of DGIdb through natural language, we developed the DGIdb Model Context Protocol (MCP) server, which allows large language models (LLMs) access to up-to-date information through the DGIdb API. We demonstrate that the MCP server improves an LLM’s ability to answer questions requiring accurate, up-to-date biomedical knowledge drawn from structured external resources. Availability and implementation: The DGIdb MCP server is detailed at https://github.com/dgidb/dgidb-mcp-server and includes instructions for accessing the server through the Claude desktop app.
- Open Access
- 7 January 2026
- Published by Oxford University Press (OUP) in Bioinformatics Advances
- Vol. 6 (1)
CIViC MCP: integrating large language models with the Clinical Interpretations of Variants in Cancer
Motivation: The Clinical Interpretations of Variants in Cancer (CIViC) knowledgebase provides a community-driven, open-source platform for discussing the biological and clinical significance of molecular variants in cancer. Large language models (LLMs) offer a natural-language interface to CIViC, enabling users to discover and summarize curated oncology knowledge, but LLMs cannot guarantee coverage of specialized, rapidly updated resources like CIViC from pretraining alone. Results: To enable users to make complex connections between CIViC information, we developed the CIViC Model Context Protocol (MCP) server. It allows users to interface with the CIViC API through natural language via LLMs. This facilitates the rapid summarization of expertly curated cancer variant interpretations. Our benchmark shows the server improves the accuracy of clinical significance retrieval while avoiding the far greater latency of GPT-5’s Agent Mode, which navigates the web by simulating human inputs. Availability and implementation: The CIViC MCP server is detailed at https://github.com/griffithlab/civic-mcp-server. The repository includes instructions for accessing the server through the Claude desktop app (our recommended approach) and hosting it locally with GPT-5, as well as a Python script for directly querying the MCP server. We also provide an MCP-supported chatbot for CIViC users at https://civicdb.org/mcp-chat.
- 7 January 2026
- Published by Oxford University Press (OUP) in Bioinformatics Advances
- Vol. 6 (1)
Open AccessCIViC MCP: integrating large language models with the Clinical Interpretations of Variants in Cancer
Motivation: The Clinical Interpretations of Variants in Cancer (CIViC) knowledgebase provides a community-driven, open-source platform for discussing the biological and clinical significance of molecular variants in cancer. Large language models (LLMs) offer a natural-language interface to CIViC, enabling users to discover and summarize curated oncology knowledge, but LLMs cannot guarantee coverage of specialized, rapidly updated resources like CIViC from pretraining alone. Results: To enable users to make complex connections between CIViC information, we developed the CIViC Model Context Protocol (MCP) server. It allows users to interface with the CIViC API through natural language via LLMs. This facilitates the rapid summarization of expertly curated cancer variant interpretations. Our benchmark shows the server improves the accuracy of clinical significance retrieval while avoiding the far greater latency of GPT-5’s Agent Mode, which navigates the web by simulating human inputs. Availability and implementation: The CIViC MCP server is detailed at https://github.com/griffithlab/civic-mcp-server. The repository includes instructions for accessing the server through the Claude desktop app (our recommended approach) and hosting it locally with GPT-5, as well as a Python script for directly querying the MCP server. We also provide an MCP-supported chatbot for CIViC users at https://civicdb.org/mcp-chat.
- Open Access
Evaluating Language Models for Biomedical Fact-Checking: A Benchmark Dataset for Cancer Variant Interpretation Verification
Accurate interpretation of genomic variants is critical for precision oncology but remains slow and dependent on specialized expertise. Public knowledgebases such as the Clinical Interpretation of Variants in Cancer (CIViC) help by curating literature-backed variant interpretations in a structured form, yet verification and review have become major bottlenecks. To address this, we developed CIViC-Fact, a benchmark dataset and pipeline for testing automated systems that verify the accuracy of cancer variant claims. CIViC-Fact links structured claims to sentence-level supporting or refuting evidence from full-text articles, and includes expert annotations and explanations. We evaluated multiple language models. Proprietary models performed well without training, but a smaller open-source model, fine-tuned on CIViC-Fact, achieved the highest accuracy (89%). Applying our fact-checking pipeline to real CIViC entries showed that reviewing less than 20% of content, focusing on flagged entries, would be sufficient to catch over half of all errors. This AI-assisted triage greatly accelerates the review process without replacing or reducing expert insight, ensuring that existing careful oversight remains in place while curators can work more efficiently. CIViC-Fact provides a realistic, high-consequence framework for biomedical fact-checking and a path toward more rigorous and efficient knowledgebase curation.
Open AccessEvaluating Language Models for Biomedical Fact-Checking: A Benchmark Dataset for Cancer Variant Interpretation Verification
Accurate interpretation of genomic variants is critical for precision oncology but remains slow and dependent on specialized expertise. Public knowledgebases such as the Clinical Interpretation of Variants in Cancer (CIViC) help by curating literature-backed variant interpretations in a structured form, yet verification and review have become major bottlenecks. To address this, we developed CIViC-Fact, a benchmark dataset and pipeline for testing automated systems that verify the accuracy of cancer variant claims. CIViC-Fact links structured claims to sentence-level supporting or refuting evidence from full-text articles, and includes expert annotations and explanations. We evaluated multiple language models. Proprietary models performed well without training, but a smaller open-source model, fine-tuned on CIViC-Fact, achieved the highest accuracy (89%). Applying our fact-checking pipeline to real CIViC entries showed that reviewing less than 20% of content, focusing on flagged entries, would be sufficient to catch over half of all errors. This AI-assisted triage greatly accelerates the review process without replacing or reducing expert insight, ensuring that existing careful oversight remains in place while curators can work more efficiently. CIViC-Fact provides a realistic, high-consequence framework for biomedical fact-checking and a path toward more rigorous and efficient knowledgebase curation.
- Open Access
- 13 February 2025
- Published by American Association for Cancer Research (AACR)
FIGURE 6 from Ultra-Deep Sequencing Reveals the Mutational Landscape of Classical Hodgkin Lymphoma
Regulating Hippo. A, Diagram of pathways that regulate Hippo signaling. Identified SMGs are shown in red; genes mutated in at least one non-hypermutated sample are shown in yellow; genes only mutated in the hypermutated sample (HL-513) are shown in orange. The gene mutation frequency across the cohort is shown as a percent. The pathways labeled in green indicate larger pathways not shown in this diagram. B, The total number and type of mutation observed are shown in the inset bar chart. C, Lolliplot of CDH5; mutations identified in the current study are shown on the top and COSMIC mutations found in lymphoid tissue on the bottom. D, Lolliplot of mutations identified at GNA13; mutations identified in the current study are shown on the top and COSMIC mutations found in lymphoid tissue on the bottom.
- 13 February 2025
- Published by American Association for Cancer Research (AACR)
Open AccessFIGURE 6 from Ultra-Deep Sequencing Reveals the Mutational Landscape of Classical Hodgkin Lymphoma
Regulating Hippo. A, Diagram of pathways that regulate Hippo signaling. Identified SMGs are shown in red; genes mutated in at least one non-hypermutated sample are shown in yellow; genes only mutated in the hypermutated sample (HL-513) are shown in orange. The gene mutation frequency across the cohort is shown as a percent. The pathways labeled in green indicate larger pathways not shown in this diagram. B, The total number and type of mutation observed are shown in the inset bar chart. C, Lolliplot of CDH5; mutations identified in the current study are shown on the top and COSMIC mutations found in lymphoid tissue on the bottom. D, Lolliplot of mutations identified at GNA13; mutations identified in the current study are shown on the top and COSMIC mutations found in lymphoid tissue on the bottom.
- Open Access
- 13 February 2025
- Published by American Association for Cancer Research (AACR)
Supplementary Tables 1-14 from Ultra-Deep Sequencing Reveals the Mutational Landscape of Classical Hodgkin Lymphoma
- 13 February 2025
- Published by American Association for Cancer Research (AACR)
Open AccessSupplementary Tables 1-14 from Ultra-Deep Sequencing Reveals the Mutational Landscape of Classical Hodgkin Lymphoma
- Open Access
- 13 February 2025
- Published by American Association for Cancer Research (AACR)
Supplementary Figure 6 from Ultra-Deep Sequencing Reveals the Mutational Landscape of Classical Hodgkin Lymphoma
Coverage Summary Across Sites that Passed and Failed Validation
- 13 February 2025
- Published by American Association for Cancer Research (AACR)
Open AccessSupplementary Figure 6 from Ultra-Deep Sequencing Reveals the Mutational Landscape of Classical Hodgkin Lymphoma
Coverage Summary Across Sites that Passed and Failed Validation
- Open Access
- 13 February 2025
- Published by American Association for Cancer Research (AACR)
Supplementary Figure 11 from Ultra-Deep Sequencing Reveals the Mutational Landscape of Classical Hodgkin Lymphoma
Summary of Recurrently Mutated Genes for Relapsed Samples
- 13 February 2025
- Published by American Association for Cancer Research (AACR)
Open AccessSupplementary Figure 11 from Ultra-Deep Sequencing Reveals the Mutational Landscape of Classical Hodgkin Lymphoma
Summary of Recurrently Mutated Genes for Relapsed Samples