P

Pythia 160m C2s

Developed by vandijklab
This is a model fine-tuned on single-cell RNA sequencing data using the Cell2Sentence method, based on the Pythia-160m language model. It is capable of conditional cell generation, unconditional cell generation, and cell type prediction.
Downloads 64
Release Time : 2/14/2024

Model Overview

The model converts single-cell RNA sequencing data into sequences of gene names ordered by expression levels (referred to as 'cell sentences'), enabling large language models to process single-cell transcriptomics data.

Model Features

Cell2Sentence Method
Innovatively converts single-cell RNA sequencing data into sequences of gene names, enabling language models to process transcriptomic data.
Multi-task Capability
Supports three main tasks: conditional cell generation, unconditional cell generation, and cell type prediction.
High Performance
Outperforms similar models in k-nearest neighbor classification and Gromov-Wasserstein distance evaluation.

Model Capabilities

Single-cell transcriptomics data analysis
Conditional cell generation
Unconditional cell generation
Cell type prediction

Use Cases

Biomedical Research
Immune Cell Analysis
Generates expression profiles of specific immune cell types based on immune tissue datasets.
Can be used to study the specificity and function of immune cells.
Cell Type Identification
Predicts the type of unknown cells based on gene expression patterns.
Demonstrates superior classification performance compared to other methods on test data.
Drug Development
Virtual Cell Generation
Generates virtual cell expression data under specific conditions.
Can be used for drug screening and effect prediction.
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase