The TranscriptFormer Single-cell Model | Dr. James Pearce
Автор: BioTuring Team
Загружено: 2025-08-11
Просмотров: 469
We were thrilled to welcome Dr. James Douglas Pearce, Staff AI/ML Research Scientist at Chan Zuckerberg Initiative, for an inside look at #TranscriptFormer: A Generative Cross-Species Cell Atlas Across 1.5 Billion Years of Evolution.
One of the biggest challenges in modern biomedical research is understanding how individual cells behave and function, especially across different species and disease states.
To address this, CZI is advancing AI-powered virtual cell models:
computational tools designed to simulate biology across species, time, and biological conditions.
TranscriptFormer represents a major leap forward in this effort. It’s the first generative, multi-species AI model for single-cell transcriptomics, trained on over 112 million cells from 12 species spanning 1.5 billion years of evolution. This unprecedented dataset enables TranscriptFormer to classify cell types, predict disease states, and simulate gene interactions — even in species and conditions it hasn’t seen before.
This model is a crucial step toward CZI’s vision of virtual cell models that accelerate discovery and transform the way we understand, diagnose, and treat disease.
Доступные форматы для скачивания:
Скачать видео mp4
-
Информация по загрузке:
