Sohini Ramachandran, PhD
Brown University
Biography:
Sohini Ramachandran is the Hermon C. Bumpus Professor of Biology at Brown University, where she also holds a courtesy appointment in Computer Science. As the Founding Director of Brown’s Data Science Institute, she led the university’s data science initiatives from 2020 to 2024, having previously directed the Center for Computational Molecular Biology (2017–2022).
Dr. Ramachandran holds a PhD from Stanford University and completed a postdoctoral fellowship at Harvard University, where she was also a Junior Fellow in the Harvard Society of Fellows. Her research has been recognized with numerous honors, including a Sloan Research Fellowship, the Pew Biomedical Scholar award, and a Presidential Early Career Award in Science and Engineering (PECASE). She currently serves as the Program Director for Brown’s NIH T32 Predoctoral Training Program in Biological Data Science. A dedicated educator, she has been honored with the Henry Merritt Wriston Fellowship and the Philip J. Bray Award for Excellence in Teaching. Her self-competitive spirit is fueled by having misspelled 'succedaneum' at the 1995 National Spelling Bee.
Abstract:
As human genetic datasets scale, the translation of complex scientific analyses of human genomes into the public-facing information ecosystem remains poorly understood. We present the first large-scale metascientific analysis of how human genetics research is synthesized and consumed publicly, analyzing 3,050,193 historical revisions and 56,098 discussions across 25 years of Wikipedia data spanning 6,737 demographic pages. Our analyses demonstrate that scientific findings in human genetics are being integrated into public-facing knowledge platforms in ways that frequently promote genetic essentialism—the false notion that ethnicity, race, and nationality are fixed, immutable biological categories. Our results reveal that genetic framing has grown substantially over time: 55.5% of the 1,000 most-visited demographic pages feature genetics terminology, and 30.4% include dedicated 'Genetics' sections, generating an estimated annual reach of 90 million views. This trend is amplified among nationalities, where 67.8% contain genetics terms and 52.5% have dedicated sections. Furthermore, this public synthesis heavily shapes downstream artificial intelligence; chatbots cite Wikipedia in 92.5% of responses about nationalities, incorporating genetics terminology in 19.2%. Grokipedia leverages genetics even more, with 93.1% of pages on ethnicity, nationality and race using genetic terms and 73.1% containing genetics sections. These findings demonstrate that human genetics research is increasingly central to how populations are defined publicly and by AI. To prevent misinterpretations, the field of human genetics must pair our technical advances with an urgent, deliberate commitment to improve how our research is communicated, contextualized, and consumed in the public sphere.
Sohini Ramachandran is the Hermon C. Bumpus Professor of Biology at Brown University, where she also holds a courtesy appointment in Computer Science. As the Founding Director of Brown’s Data Science Institute, she led the university’s data science initiatives from 2020 to 2024, having previously directed the Center for Computational Molecular Biology (2017–2022).
Dr. Ramachandran holds a PhD from Stanford University and completed a postdoctoral fellowship at Harvard University, where she was also a Junior Fellow in the Harvard Society of Fellows. Her research has been recognized with numerous honors, including a Sloan Research Fellowship, the Pew Biomedical Scholar award, and a Presidential Early Career Award in Science and Engineering (PECASE). She currently serves as the Program Director for Brown’s NIH T32 Predoctoral Training Program in Biological Data Science. A dedicated educator, she has been honored with the Henry Merritt Wriston Fellowship and the Philip J. Bray Award for Excellence in Teaching. Her self-competitive spirit is fueled by having misspelled 'succedaneum' at the 1995 National Spelling Bee.
Abstract:
As human genetic datasets scale, the translation of complex scientific analyses of human genomes into the public-facing information ecosystem remains poorly understood. We present the first large-scale metascientific analysis of how human genetics research is synthesized and consumed publicly, analyzing 3,050,193 historical revisions and 56,098 discussions across 25 years of Wikipedia data spanning 6,737 demographic pages. Our analyses demonstrate that scientific findings in human genetics are being integrated into public-facing knowledge platforms in ways that frequently promote genetic essentialism—the false notion that ethnicity, race, and nationality are fixed, immutable biological categories. Our results reveal that genetic framing has grown substantially over time: 55.5% of the 1,000 most-visited demographic pages feature genetics terminology, and 30.4% include dedicated 'Genetics' sections, generating an estimated annual reach of 90 million views. This trend is amplified among nationalities, where 67.8% contain genetics terms and 52.5% have dedicated sections. Furthermore, this public synthesis heavily shapes downstream artificial intelligence; chatbots cite Wikipedia in 92.5% of responses about nationalities, incorporating genetics terminology in 19.2%. Grokipedia leverages genetics even more, with 93.1% of pages on ethnicity, nationality and race using genetic terms and 73.1% containing genetics sections. These findings demonstrate that human genetics research is increasingly central to how populations are defined publicly and by AI. To prevent misinterpretations, the field of human genetics must pair our technical advances with an urgent, deliberate commitment to improve how our research is communicated, contextualized, and consumed in the public sphere.
