The Chan Zuckerberg Initiative (CZI) has released TranscriptFormer, a generative artificial intelligence model trained on over 112 million single cells from 12 species spanning 1.5 billion years of evolution. This AI model provides a concise and interactive representation of extensive cellular data, enabling the exploration of gene expression patterns across evolutionary timelines. TranscriptFormer can simulate biological processes it has not previously encountered, offering new avenues for understanding cellular biology and potential therapeutic applications. The model represents a significant advancement in cross-species cellular analysis and is part of ongoing efforts to leverage AI in genomics and multi-omics research.
The Cross-Species Cell Atlas: @ChanZuckerberg Initiative Releases TranscriptFormer AI Model Trained on data covering 1.5 billion years of evolution, CZI’s new generative AI model probes cellular biology across species for therapeutic applications. https://t.co/DhNnHelc8E
SGCRNA: Spectral Clustering-Guided Co-Expression Network Analysis Without Scale-Free Constraints for Multi-Omic Data https://t.co/Icd7cprCmL #biorxiv_bioinfo
TransAgent: Dynamizing Transcriptional Regulation Analysis via Multi-omics-Aware AI Agent https://t.co/gjVDrbBjGW #biorxiv_bioinfo