Deep-learning methods based computational modeling

NIH RePORTER · NIH · R01 · $50,000 · view on reporter.nih.gov ↗

Abstract

Project summary Recent advancements in deep learning-based computational protein structure prediction by AlphaFold, RoseTTAFold, ESMFold, and OpenFold methods offer promising opportunities to advance our ongoing NINDS-funded research project entitled “Activation and Inhibition Mechanisms of Calcium-Activated Nonselective Cation Channels” (1R01NS128180). We plan to use current and future deep learning-based computational protein structure prediction methods to predict gating conformational changes in TRP channels with high accuracy to complement our functional studies. We aim to capitalize on these opportunities with this administrative supplement. To fully utilize the extensive capabilities of deep learning-based computational protein structure prediction method capabilities, powerful computational resources are needed with advanced GPU, CPU, RAM, and disk capacity. We are requesting $50,000 NINDS supplement funding to purchase an AI system capable of running all current and future deep learning-based computational protein structure prediction methods. We have identified the ideal AI system configuration from Bizon Technostore. A quote is submitted together with this application.

Key facts

NIH application ID
10816248
Project number
3R01NS128180-02S1
Recipient
UNIVERSITY OF CALIFORNIA AT DAVIS
Principal Investigator
Juan Du
Activity code
R01
Funding institute
NIH
Fiscal year
2023
Award amount
$50,000
Award type
3
Project period
2022-06-01 → 2024-05-31