# Bioinformatics

METAGENE-1 is a metagenomic foundational model co-developed by researchers from the University of Southern California, Prime Intellect, and the Nucleic Acid Observation Station. With 7 billion parameters trained on 1.5 trillion base pairs of DNA and RNA sequences sourced from human wastewater samples, METAGENE-1's primary function is to assist in public health applications, including disease surveillance, pathogen detection, and early identification of emerging health threats. Its strength lies in capturing comprehensive genomic information distribution within the human microbiome, demonstrating robust generalization capabilities.

Chai Discovery is a platform dedicated to decoding life interactions, potentially encompassing bioinformatics, genomics, or related fields. It aims to unveil the complex interactions between living organisms through advanced technological means. The significance of this product or technology lies in its potential to provide deep insights and data support for life sciences, medical research, and related fields.

ChatCell is an application framework that makes single-cell analysis easier and more intuitive by leveraging natural language processing. It first transforms single-cell RNA sequencing data into a format easily understandable by large language models (LLMs). It then uses templates to integrate this single-cell language with task descriptions and target results, creating comprehensive single-cell instructions. To enhance the LLM's expertise in the single-cell domain, it undergoes vocabulary adaptation using a specialized single-cell vocabulary library to expand the model's knowledge base. Subsequently, it utilizes unified sequence generation to empower the model to proficiently execute various single-cell tasks.

KG-RAG is a task-agnostic framework that combines the explicit knowledge from a knowledge graph with the implicit knowledge of a large language model. Here, we utilize a massive biomedical knowledge graph SPOKE as the provider of biomedical context. A key feature of KG-RAG is its ability to extract 'prompt-relevant context' from the SPOKE knowledge graph, defined as the minimal context required to respond to a user prompt.