Peer-Reviewed Publication
Patterns (N Y)2026;7(9):101635.September 11, 2026Journal Article
EmmaEmb: A quantitative framework for analyzing embedding spaces in molecular biology.
Pia Francesca Rissom1,2, Vít Škrhák2,3, Paulo Yanez Sarmiento1, Jordan F Safer2,4, Connor W Coley5,6, Bernhard Y Renard1,7, Henrike O Heyne1,7,8,9,10, Sumaiya Iqbal2,10
1Hasso Plattner Institute, Digital Engineering Faculty, University of Potsdam, Potsdam, Germany.
2Bioinformatics and Machine Learning, Broad Institute of MIT and Harvard, Cambridge, MA, USA.
3Department of Software Engineering, Charles University, Prague, Czech Republic.
4Faculty of Biology, Technion-Israel Institute of Technology, Haifa, Israel.
5Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA, USA.
6Department of Electrical Engineering, Massachusetts Institute of Technology, Cambridge, MA, USA.
7Hasso Plattner Institute for Digital Health at Mount Sinai, Icahn School of Medicine, New York, NY, USA.
8Finnish Institute for Molecular Medicine (FIMM), University of Helsinki, Helsinki, Finland.
9Program in Medical and Population Genetics, Broad Institute of MIT and Harvard, Cambridge, MA, USA.
10Cancer Data Science, Dana-Farber/Harvard Cancer Center, Cambridge, MA, USA.
Abstract
Embeddings, numerical vectors learned by deep-learning models, are increasingly used to represent complex molecular biology data and support predictive tasks and generative design. There is a growing need for systematic approaches to interpret and explain the information encoded in high-dimensional embedding spaces. Here, we introduce EmmaEmb, a quantitative, model-agnostic framework for geometric…
Create a free account to keep reading
Free members get 10 full research views every month across publications, clinical trials, FDA clearances, adverse events, and NIH grants. No credit card required.
Want unlimited research access? See Pro plans