Research section Research projects and references

Research: CANCER Folding Project #17603

Project #17603 overview

Project Summary AI Beta

This project tries to improve how we study protein shapes using computer simulations. It focuses on finding important shapes of a protein called MET kinase, which is linked to lung cancer. By choosing the most interesting shapes to simulate, it hopes to learn more efficiently than traditional methods.
Automated summary; simplified and may not be fully accurate.

Project team

Manager(s)
Sukrit Singh
Institution
Memorial Sloan-Kettering Cancer-Center

Work unit

Atoms
59,897
Core
OPENMM_22
Status
Beta

No related projects listed.

Source material

Official Project Description

This project is an attempt at implementing adaptive sampling in Folding@home.

Adaptive sampling is a way of enhancing sampling of protein conformational space by selectively launching simulations from the most "valuable" work units. Identifying druggable states or exploring conformational state space relevant to disease is an existing challenge.

The embarassingly parallel nature of Folding@home allows us to massively scale up our exploration.

However, the underlying methods still rely on luck to a large extent – we must discover the states in work units as the dataset grows in size and more work units are run.

This can be an incredibly inefficient process, wasting work units on regions of state space that are irrelevant or uninteresting to the question at hand.

Adaptive Sampling is a way to tackle this inefficiency.

Using iterative rounds, where we collect the work units so far and select the "best/most valuable" conformational state worth exploring.

New simulations and work units are launched from these most valuable work units, hopefully more efficiently exploring state space.

This project is identical in calculation to 16497, exploring conformations of MET kinase, involved in non-small-cell lung carcinoma, but acting as a test bed.

Performance data

Hardware Performance for Project 17603

Compare community-sampled Folding@Home output for the GPUs and CPUs processing this project.

Data as of Friday, 14 August 2026 15:32:09

GPU PPD Averages

Rank
Project
Model Name
Folding@Home Identifier
Make
Brand
GPU
Model
PPD
Average
Points WU
Average
WUs Day
Average
WU Time
Average
1 GeForce RTX 3080 Ti
GA102 [GeForce RTX 3080 Ti]
Nvidia GA102 5,590,626 288,223 19.40 1 hrs 14 mins
2 GeForce RTX 3090
GA102 [GeForce RTX 3090]
Nvidia GA102 4,947,845 275,336 17.97 1 hrs 20 mins
3 GeForce RTX 3070 Lite Hash Rate
GA104 [GeForce RTX 3070 Lite Hash Rate]
Nvidia GA104 3,331,531 241,653 13.79 1 hrs 44 mins
4 GeForce RTX 2080 Rev. A
TU104 [GeForce RTX 2080 Rev. A] 10068
Nvidia TU104 3,095,323 227,725 13.59 1 hrs 46 mins
5 GeForce RTX 3060 Ti Lite Hash Rate
GA104 [GeForce RTX 3060 Ti Lite Hash Rate]
Nvidia GA104 2,719,187 225,335 12.07 1 hrs 59 mins
6 GeForce RTX 2060
TU104 [GeForce RTX 2060]
Nvidia TU104 2,013,689 204,814 9.83 2 hrs 26 mins