About

Isaac Kohane, MD, PhD, is the inaugural Chair of the Department of Biomedical Informatics and the Marion V. Nelson Professor of Biomedical Informatics at Harvard Medical School. He developed and applies computational techniques to address disease at multiple scales, including healthcare systems as 'living laboratories' and the functional genomics of neurodevelopment with a focus on autism. Over the last 30 years, Kohane's research has been driven by the vision of transforming biomedical data into knowledge and translating that knowledge into practice to find new cures, improve diagnoses, and deliver optimal care. He has designed and led multiple internationally adopted efforts to instrument the healthcare enterprise for discovery and to enable innovative decision-making tools at the point of care. His work on 'omic-scale molecular analyses' has contributed to re-characterizing and reclassifying diseases such as autism, rheumatoid arthritis, and cancers, often utilizing developmental trajectories of genes to unravel complex diseases. Kohane earned his MD/PhD from Boston University in 1987, completed post-doctoral work at Boston Children’s Hospital, and has been a faculty member at Harvard Medical School since 1992. He served as Director of Countway Library and Co-Director of the Center for Biomedical Informatics before it became the Department of Biomedical Informatics in 2015. He is a member of the Institute of Medicine and the American Society for Clinical Investigation, has published several hundred papers, authored notable books including 'Microarrays for an Integrative Genomics' and 'The AI Revolution in Medicine: GPT-4 and Beyond,' and is the Editor-in-Chief of NEJM AI. His work continues to focus on leveraging data and computational methods to advance biomedical science and medicine.

Research topics

• Medicine
• Biology
• Genetics
• Internal medicine
• Psychology
• Computer Science
• Psychiatry
• Artificial Intelligence
• Political Science
• Pathology
• Nursing
• Machine Learning
• Oncology
• Clinical psychology
• Business
• Environmental health
• Biochemistry
• Law
• Bioinformatics
• Statistics
• Public relations
• Mathematics
• Algorithm
• Data science

Selected publications

• Rare Germline Variants in Immune and Drug Target Genes Among Cancer Exceptional Responders

  medRxiv · 2026-05-19

  articleSenior author

  Abstract Background Cancer treatment response is highly variable, even among patients with the same tumor type and treatment. Exceptional responders (ERs), who are individuals who experience unusually favorable outcomes, provide critical insights into the biological factors driving treatment success. While prior studies have highlighted the role of somatic changes, the contribution of germline rare variants remains underexplored. This study aimed to uncover the genetic underpinnings of exceptional responses by identifying rare, non-silent and predicted deleterious germline mutations enriched among ERs compared to typical cancer patients. Methods The Network of Enigmatic Exceptional Responders (NEER) project collected clinical and germline whole-genome sequencing (WGS) data from 53 ERs. After quality control procedures and ancestry background checks, 51 ERs were left for final analysis. While non-silent mutations were identified based on allele frequencies and mutation types, multiple pathogenicity predictors were applied for predicted deleterious variants. These were compared to a harmonized and comparable subset from the Pan-Cancer Analysis of Whole Genomes (PCAWG) cohort (n=414) using Fisher’s exact tests. Kaplan-Meier survival analysis applied to evaluate prognostic associations in PCAWG patients. Additionally, Fisher’s exact tests were conducted stratified by cancer type and treatment regimen to identify potential associations between rare germline variants and therapeutic responses. Results Variants in immune-related genes such as CCL26 and GPRC5D were prevalent, suggesting enhanced immune regulation among ERs. Fourteen genes with non-silent and eight with predicted deleterious mutations showed significantly different frequencies between NEER and PCAWG cohorts (FDR < 0.05). IRX3 emerged as a protective gene enriched in ERs, whereas OR6B2 was associated with poor survival in PCAWG lung cancer patients. Moreover, rare non-silent germline variants in drug target genes were enriched among ERs treated with cisplatin and doxorubicin, implicating altered DNA repair and drug-binding mechanisms in their remarkable outcomes. Conclusions This study reveals a distinctive germline mutation landscape in exceptional cancer responders, marked by immune-related and drug-target-associated variants that may enhance therapy response and prolong survival. The findings highlight potential novel prognostic biomarkers, such as IRX3 and OR6B2, providing a foundation for developing personalized cancer treatments informed by rare genetic variation.

  PublisherDOI

• One Patient, Many Contexts: Scaling Medical AI with Contextual Intelligence

  ArXiv.org · 2025-06-11 · 1 citations

  preprintOpen access

  Medical AI, including clinical language models, vision-language models, and multimodal health record models, already summarizes notes, answers questions, and supports decisions. Their adaptation to new populations, specialties, or care settings often relies on fine-tuning, prompting, or retrieval from external knowledge bases. These strategies can scale poorly and risk contextual errors: outputs that appear plausible but miss critical patient or situational information. We envision context switching as a solution. Context switching adjusts model reasoning at inference without retraining. Generative models can tailor outputs to patient biology, care setting, or disease. Multimodal models can reason on notes, laboratory results, imaging, and genomics, even when some data are missing or delayed. Agent models can coordinate tools and roles based on tasks and users. In each case, context switching enables medical AI to adapt across specialties, populations, and geographies. It requires advances in data design, model architectures, and evaluation frameworks, and establishes a foundation for medical AI that scales to infinitely many contexts while remaining reliable and suited to real-world care.

  PublisherOA PDFDOI

• Genome-wide association analyses identify 44 risk variants and refine the genetic architecture of major depression.

  UNC Libraries · 2025-10-17

  articleOpen access

  Major depressive disorder (MDD) is a common illness accompanied by considerable morbidity, mortality, costs, and heightened risk of suicide. We conducted a genome-wide association meta-analysis based in 135,458 cases and 344,901 controls and identified 44 independent and significant loci. The genetic findings were associated with clinical features of major depression and implicated brain regions exhibiting anatomical differences in cases. Targets of antidepressant medications and genes involved in gene splicing were enriched for smaller association signal. We found important relationships of genetic risk for major depression with educational attainment, body mass, and schizophrenia: lower educational attainment and higher body mass were putatively causal, whereas major depression and schizophrenia reflected a partly shared biological etiology. All humans carry lesser or greater numbers of genetic risk factors for major depression. These findings help refine the basis of major depression and imply that a continuous measure of risk underlies the clinical phenotype.

  PublisherDOI

• International electronic health record-derived COVID-19 clinical course profiles: the 4CE consortium

  UNC Libraries · 2025-06-26

  articleOpen access
  PublisherDOI

• Accelerating Science with Human+AI Review

  NEJM AI · 2025-11-26 · 10 citations

  articleSenior author
  PublisherDOI

• GENIE: Generative Note Information Extraction model for structuring EHR data

  ArXiv.org · 2025-01-30 · 2 citations

  preprintOpen access

  Electronic Health Records (EHRs) hold immense potential for advancing healthcare, offering rich, longitudinal data that combines structured information with valuable insights from unstructured clinical notes. However, the unstructured nature of clinical text poses significant challenges for secondary applications. Traditional methods for structuring EHR free-text data, such as rule-based systems and multi-stage pipelines, are often limited by their time-consuming configurations and inability to adapt across clinical notes from diverse healthcare settings. Few systems provide a comprehensive attribute extraction for terminologies. While giant large language models (LLMs) like GPT-4 and LLaMA 405B excel at structuring tasks, they are slow, costly, and impractical for large-scale use. To overcome these limitations, we introduce GENIE, a Generative Note Information Extraction system that leverages LLMs to streamline the structuring of unstructured clinical text into usable data with standardized format. GENIE processes entire paragraphs in a single pass, extracting entities, assertion statuses, locations, modifiers, values, and purposes with high accuracy. Its unified, end-to-end approach simplifies workflows, reduces errors, and eliminates the need for extensive manual intervention. Using a robust data preparation pipeline and fine-tuned small scale LLMs, GENIE achieves competitive performance across multiple information extraction tasks, outperforming traditional tools like cTAKES and MetaMap and can handle extra attributes to be extracted. GENIE strongly enhances real-world applicability and scalability in healthcare systems. By open-sourcing the model and test data, we aim to encourage collaboration and drive further advancements in EHR structurization.

  PublisherOA PDFDOI

• Heterogeneous Effect of Automated Alerts on Mortality

  medRxiv · 2025-08-12

  preprintOpen access

  We analyzed data from 13,483 hospitalized patients with acute kidney injury (AKI) from three randomized controlled trials to assess the heterogeneous effects of automated electronic alerts on 14-day mortality. We modeled and predicted individualized alert effects on a subset of the ELAIA-1 patients and validated it internally on ELAIA-1 holdout patients and externally on ELAIA-2 and UPenn trial patients. Patients predicted to benefit from alerts had significantly lower mortality compared to those predicted to be harmed (p-interaction<0.05). In external cohorts, 43 deaths may have been preventable if alerts were restricted to likely beneficiaries. Machine-learning based meta-analysis identified reduced mortality with alerts among patients with higher blood pressures and lower predicted risk, but increased mortality in non-urban and non-teaching hospitals. Provider responses to alerts varied across subgroups. These findings suggest that tailoring alerts to patient phenotypes may improve outcomes and support the need for a prospective trial of individualized alert strategies. Trial Registration: https://clinicaltrials.gov/ct2/show/NCT02753751 and https://clinicaltrials.gov/ct2/show/NCT02771977.

  PublisherOA PDFDOI

• A case study on using a large language model to analyze continuous glucose monitoring data

  Scientific Reports · 2025-01-07 · 11 citations

  articleOpen accessSenior author

  Continuous glucose monitors (CGM) provide valuable insights about glycemic control that aid in diabetes management. However, interpreting metrics and charts and synthesizing them into linguistic summaries is often non-trivial for patients and providers. The advent of large language models (LLMs) has enabled real-time text generation and summarization of medical data. The objective of this study was to assess the strengths and limitations of using an LLM to analyze raw CGM data and produce summaries of 14 days of data for patients with type 1 diabetes. We first evaluated the ability of GPT-4 to compute quantitative metrics specific to diabetes found in an Ambulatory Glucose Profile (AGP). Then, using two independent clinician graders, we evaluated the accuracy, completeness, safety, and suitability of qualitative descriptions produced by GPT-4 across five different CGM analysis tasks. GPT-4 performed 9 out of the 10 quantitative metrics tasks with perfect accuracy across all 10 cases. The clinician-evaluated CGM analysis tasks had good performance across measures of accuracy [lowest task mean score 8/10, highest task mean score 10/10], completeness [lowest task mean score 7.5/10, highest task mean score 10/10], and safety [lowest task mean score 9.5/10, highest task mean score 10/10]. Our work serves as a preliminary study on how generative language models can be integrated into diabetes care through data summarization and, more broadly, the potential to leverage LLMs for streamlined medical time series analysis.

  PublisherOA PDFDOI

• A standards-based approach to digital health research: implementing the people heart study

  Journal of the American Medical Informatics Association · 2025-09-17

  articleOpen accessSenior author

  OBJECTIVE: To assess whether HL7 Fast Healthcare Interoperability Resources (FHIR) can underpin a fully standards-based, end-to-end digital research architecture, demonstrate it in a live study, and quantify its benefits for interoperability and development efficiency. MATERIALS AND METHODS: We designed a generalizable standards-based architecture to accelerate digital health research relying on FHIR as the sole transactional model throughout a participant research lifecycle starting from API-based study discovery to results. It was instantiated for People Heart Study, a real-world digital health cardiovascular-risk assessment study with its protocol transformed into FHIR resources (eligibility, consent, tasks, and results). Evaluation examined workflow coverage, validator conformance across independent servers, and points requiring custom extensions or app logic. RESULTS: The architecture was implemented using cloud managed FHIR stores including an illustrative public research discovery API for first-/third-party apps. A participant-facing iOS app was published on the App Store. Our evaluation reveals that 6 of 10 research app workflows could be executed entirely from FHIR artifacts; 2 were partially standards-driven and 2 remained limited requiring custom development. All FHIR resources passed structural, semantic validation with minimal custom extension usage and terminology integrity issues. DISCUSSION: Our approach addresses persistent challenges in digital health research by enhancing data interoperability, minimizing redundant development, and supporting the full research lifecycle. The architecture aligns with national priorities and complements healthcare standardization efforts. CONCLUSION: By leveraging FHIR, our architecture enables generalizability, interoperability, and reuse across diverse digital health research contexts, transforming study design into data modeling rather than software development, and fostering a more inclusive and agile digital health ecosystem.

  PublisherOA PDFDOI

• The missing value of medical artificial intelligence

  Nature Medicine · 2025-11-25 · 1 citations

  article1st authorCorresponding
  PublisherDOI

Recent grants

• NIH Grant R01LM010125

  NIH · $1.2M · 2012

• NIH Grant U54HL117798

  NIH · $36.7M · 2018

• The Integration of Trans-omics for Precision Medicine (TOPMED) and Other Heart, Lung, Blood and Sleep (HLBS) Data Sets with the Data Commons

  NIH · $8.8M · 2017–2022

• NIH Grant U01CA091429

  NIH · $6.7M · 2007

• NIH Grant U01LM005877

  NIH · $912k · 1997

Frequent coauthors

• Shawn N. Murphy

  357 shared

• Kenneth D. Mandl

  Harvard University

  231 shared

• Tianxi Cai

  Harvard University

  226 shared

• Nathan Palmer

  220 shared

• Katherine P. Liao

  Harvard University

  200 shared

• Susanne Churchill

  Harvard University

  187 shared

• Griffin M. Weber

  179 shared

• Stanley Y. Shaw

  166 shared

Labs

• ZaklabPI