Causal Models Need Causal Data - Xaira’s X-Cell model for Drug Discovery (Bo Wang & Ci Chu, Chief Discovery Officer & Chief AI Scientist)
Quick Answer
Xaira's X-Cell model leverages the X-Atlas dataset to enhance drug discovery by utilizing 30x more information, overcoming limitations of smaller datasets.
Quick Take
The shift from autoregression to diffusion and CRISPR-based experiments enables predictive modeling of gene expression changes in human cells, aiming to revolutionize AI-driven drug development.
Key Points
- X-Cell model trained on X-Atlas dataset enhances predictive capabilities for gene expression.
- 30x more information improves model performance beyond 1.5B parameters.
- CRISPR-based experiments generate extensive raw data for X-Atlas.
- New approach shifts from autoregression to diffusion for better results.
- Promotions of Ci Chu and Bo Wang highlight strategic importance of data-rich drug development.
DeepSignal Analysis
What happened
Xaira's X-Cell model utilizes the X-Atlas dataset to enhance drug discovery by providing significantly more information than previous datasets. The model shifts from autoregression to diffusion and incorporates CRISPR-based experiments, aiming to improve predictive modeling of gene expression changes in human cells.
Key evidence
- The X-Cell model is based on the X-Atlas dataset, which offers approximately 30 times more information than smaller datasets, addressing limitations in predictive modeling.
- Models trained on CELLxGENE have been shown to describe relationships between cell types and states but struggle to predict changes in RNA expression effectively.
- Xaira's approach includes CRISPR-based experiments that run millions of tests in parallel, generating the raw data necessary for the X-Atlas dataset.
Why it matters
The advancement in data richness is crucial for AI-driven drug development, as it allows for better predictive modeling of gene expression changes. This could lead to more effective drug discovery processes and a deeper understanding of cellular mechanisms, which are essential for developing targeted therapies.
What to watch
Source Excerpt
Xaira Therapeutics is all in on data generation for model building! We talk with Bo Wang and Ci Chu about how and why.
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from Latent Space
See more →![[AINews] SpaceXAI launches Grok 4.5, first Opus-class model post Cursor acquisition](https://substackcdn.com/image/fetch/$s_!8D6O!,f_auto,q_auto:good,fl_progressive:steep/https%3A%2F%2Fpbs.substack.com%2Fmedia%2FHMuQw2BXUAAJaQd.png)
[AINews] SpaceXAI launches Grok 4.5, first Opus-class model post Cursor acquisition
SpaceXAI has launched Grok 4.5, a new coding-focused model that is 3x larger than Grok 4.3, priced at $2 per million input tokens. Positioned as an Opus-class model, it aims for efficiency and speed, outperforming competitors like GPT-5.6 and Opus 4.8 in cost-effectiveness.
