RESEARCH: DUD-E
FOLDING PROJECT #12214 PROFILE
PROJECT TEAM
Manager(s): Louis SmithInstitution: University of Pennsylvania
WORK UNIT INFO
Atoms: 313,529Core: 0x22
Status: Public
Related Projects
TLDR; PROJECT SUMMARY AI BETA
This project uses computer simulations to study how proteins interact with small molecules (like drugs). They're focusing on proteins from the DUD-E dataset, which has a bunch of different proteins and known binding information. One example is Acetylcholinesterase, important for nerve function and targeted by pesticides and some medicines.
Note: This TLDR is a simplication and may not be 100% accurate.OFFICAL PROJECT DESCRIPTION
In this series of projects we are simulating proteins that are part of the DUD-E benchmark data set for protein-ligand interactions, using simulations initialized from Alpha Fold. Simulation methods to study protein-small molecule interactions are of critical importance to the early stages of drug discovery, but most methods have a poor balance of accuracy relative to cost.
Much of the development process for new compounds happens via screening large libraries of compounds for activity against target proteins believed to be relevant for a disease.
Lending focus to this search makes developing new molecules into drugs more economical and faster. In order to do this kind of methods development, good reference data that is widely available is essential.
A classic dataset for benchmarking structural methods attempting to predict protein-ligand interactions known as DUD-E has been widely used because it has diverse proteins, and each protein is bound to a fairly large collection (usually more than fifty) of small molecules for which the ability to bind the receptor have been measured experimentally.
Using Folding@Home, we will create large reference quality simulations of these proteins.
Because we know how such simulations, and the binding methods we or others may test on them, should look and function we have a great yardstick for improving the methods we have and developing new ones. In this project series we have the following systems, many of which are known for their medical relevance in addition to having been extensively studied with both simulation and experiment in the past. 12201 - ACES: Acetylcholinesterase that is critical to nervous system function in animals.
It is the target of pesticides, and also numerous drugs.
If targeted in the correct way, it can reduce neural swelling.
This sequence happens to be from the Pacific Electric ray, Torpedo Californica, which was a landmark discovery in biomedical efforts to isolate neurotransmitter receptors and led to a mechanistic understanding of myasthenia gravis. 12202 - AKT2: serine-threonine kinase taking part in the insulin signal transduction pathway.
Implicated in some cancers, it has been a target of drug development campaigns in the past. 12203 - AMPC: A critical antibiotic resistance gene, it is a beta lactamase capable of opening the critical structural feature of celphalosporin-type antibiotics, rendering them ineffective. 12204 - BACE1: Beta secretase 1, an aspartic acid protease that helps form myelin sheaths in neurons.
It is the major generator of amyloid-beta peptides in neurons, and therefore is implicated in Alzheimer's disease. 12205 - BRAF: B-raf is involved in sending signals involved in cell growth, and as such is considered a proto-oncogen.
It is a serine/threonine kinase that has several known inhibitors, some of which are now anti-cancer medications. 12206 - CASP3: a caspase-type protease that participates in the execution of apoptosis, the process of programmed cell death.
It also acts to cleave one of the amyloid forming proteins and is therefore implicated in Alzheimer's dementia. 12207 - CDK2: one of the cyclin dependent kinases, this protein is a checkpoint kinase that signals transitions between growth and DNA synthesis phases in the cell cycle.
Dysfunction in this checkpoint is associated with cancer; inhibiting CDK2 can arrest cell cycle in cases of abmormal growth, so it has been an anti-cancer target for some time. 12208 - CSF1R: Colony stimulating factor 1 receptor, when bound by cognate ligands, will promote survival, proliferation and differentiation of many myeloid cell types.
It is thus involved in disease and is targeted in therapies for cancer, neurodegenerative diseases, nad inflammatory bone diseases. 12209 - DPP4: Dipeptidyl peptidase-4, a protein that cuts up certain other proteins on the surfaces of most cells.
Important in immune regulation, signal transduction, and apoptosis, molecules inhibiting its enzymatic activity can help treat type 2 diabetes because the peptide hormones (GLP-1, and GIP) are degraded by DPP4.
Thus, inhibiting DPP4 prolongs the effects of these hormones. 12210 - EGFR: Epidermal growth factor receptor; its deficient signaling is associated with Alzheimer's dementia, whereas its over-expression is a common characteristic of tumor cells.
It is thus an oncogene that is targeted by numerous anti-cancer molecules and drugs.
Many of these are targeted at the tyrosine kinase domain, because hampering its function prevents excessive transduction of the signals these receptors would otherwise send to the nucleus of the tumor cell. 12211 - ESR1: Estrogen Receptor Alpha is critical to many tissue differentiation processes across the body, and has been targeted by various drugs to both enhance and suppress its effects depending on associated conditions.
12212 - FA10: Coagulation factor X is an enzyme in the coagulation signaling cascade for forming blood clots.
It is a serine endopeptidase, and has been targeted by inhibitors to reduce coagulation in medical contexts where that is desirable.
.
RELATED TERMS GLOSSARY AI BETA
Acetylcholinesterase
An enzyme that breaks down acetylcholine in the nervous system.
Acetylcholinesterase is a crucial enzyme found in the nervous system. It's responsible for breaking down acetylcholine, a neurotransmitter involved in muscle movement and nerve signaling. Medications that inhibit acetylcholinesterase can be used to treat conditions like Alzheimer's disease.
DUD-E
Directory of Useful Decoys Enhanced
DUD-E is a database of protein-ligand complexes used for benchmarking computational methods in drug discovery. It contains diverse proteins with experimentally determined binding affinities to various small molecules.
Alpha Fold
A deep learning algorithm for predicting protein structures.
AlphaFold is a revolutionary artificial intelligence system developed by DeepMind that can accurately predict the three-dimensional structures of proteins. This breakthrough has immense implications for understanding protein function and advancing drug discovery.
Protein-Ligand Interactions
Interactions between proteins and small molecules (ligands).
Protein-ligand interactions are essential for many biological processes. When a drug binds to a protein target, it can modulate its activity and have therapeutic effects. Understanding these interactions is crucial for drug development.
Drug Discovery
The process of identifying and developing new drugs.
Drug discovery is a complex multi-stage process that involves identifying promising drug candidates, optimizing their properties, and conducting rigorous testing to ensure safety and efficacy. It's a crucial step in bringing new therapies to patients.
Folding@Home
A distributed computing project for protein folding research.
Folding@Home harnesses the power of volunteer computers to simulate protein folding, a fundamental process in biology. This vast computational effort aids in understanding protein structure and function, contributing to advancements in drug discovery and disease treatment.
PROJECT FOLDING PPD AVERAGES BY GPU
Data as of Tuesday, 14 April 2026 06:35:29|
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 4090 AD102 [GeForce RTX 4090] |
Nvidia | AD102 | 18,078,861 | 1,344,474 | 13.45 | 1 hrs 47 mins |
| 2 | GeForce RTX 4080 AD103 [GeForce RTX 4080] |
Nvidia | AD103 | 14,097,600 | 1,257,359 | 11.21 | 2 hrs 8 mins |
| 3 | GeForce RTX 4070 Ti AD104 [GeForce RTX 4070 Ti] |
Nvidia | AD104 | 9,953,358 | 1,117,162 | 8.91 | 2 hrs 42 mins |
| 4 | GeForce RTX 3090 Ti GA102 [GeForce RTX 3090 Ti] |
Nvidia | GA102 | 7,278,113 | 1,003,570 | 7.25 | 3 hrs 19 mins |
| 5 | GeForce RTX 4070 AD104 [GeForce RTX 4070] |
Nvidia | AD104 | 7,135,881 | 1,005,211 | 7.10 | 3 hrs 23 mins |
| 6 | GeForce RTX 3090 GA102 [GeForce RTX 3090] |
Nvidia | GA102 | 6,768,268 | 987,273 | 6.86 | 3 hrs 30 mins |
| 7 | GeForce RTX 3080 Ti GA102 [GeForce RTX 3080 Ti] |
Nvidia | GA102 | 6,462,604 | 973,972 | 6.64 | 3 hrs 37 mins |
| 8 | GeForce RTX 3080 GA102 [GeForce RTX 3080] |
Nvidia | GA102 | 5,743,323 | 933,940 | 6.15 | 3 hrs 54 mins |
| 9 | GeForce RTX 3080 Lite Hash Rate GA102 [GeForce RTX 3080 Lite Hash Rate] |
Nvidia | GA102 | 4,520,754 | 881,633 | 5.13 | 4 hrs 41 mins |
| 10 | GeForce RTX 4060 Ti AD106 [GeForce RTX 4060 Ti] |
Nvidia | AD106 | 3,964,935 | 831,717 | 4.77 | 5 hrs 2 mins |
| 11 | GeForce RTX 3070 Ti GA104 [GeForce RTX 3070 Ti] |
Nvidia | GA104 | 3,925,568 | 822,711 | 4.77 | 5 hrs 2 mins |
| 12 | RTX 4000 SFF Ada Generation AD104GL [RTX 4000 SFF Ada Generation] |
Nvidia | AD104GL | 3,831,698 | 852,911 | 4.49 | 5 hrs 21 mins |
| 13 | GeForce RTX 2080 Ti TU102 [GeForce RTX 2080 Ti] M 13448 |
Nvidia | TU102 | 3,779,171 | 824,234 | 4.59 | 5 hrs 14 mins |
| 14 | GeForce RTX 3070 Lite Hash Rate GA104 [GeForce RTX 3070 Lite Hash Rate] |
Nvidia | GA104 | 3,163,182 | 781,215 | 4.05 | 5 hrs 56 mins |
| 15 | GeForce RTX 3060 Ti GA104 [GeForce RTX 3060 Ti] |
Nvidia | GA104 | 3,117,052 | 766,697 | 4.07 | 5 hrs 54 mins |
| 16 | GeForce RTX 3070 GA104 [GeForce RTX 3070] |
Nvidia | GA104 | 3,023,130 | 768,251 | 3.94 | 6 hrs 6 mins |
| 17 | GeForce RTX 4060 AD107 [GeForce RTX 4060] |
Nvidia | AD107 | 2,609,155 | 778,572 | 3.35 | 7 hrs 10 mins |
| 18 | GeForce RTX 3060 Ti Lite Hash Rate GA104 [GeForce RTX 3060 Ti Lite Hash Rate] |
Nvidia | GA104 | 2,305,676 | 667,580 | 3.45 | 6 hrs 57 mins |
| 19 | GeForce RTX 2080 Super TU104 [GeForce RTX 2080 SUPER] |
Nvidia | TU104 | 2,289,688 | 688,226 | 3.33 | 7 hrs 13 mins |
|
|
|||||||
| 20 | RTX A4000 GA104GL [RTX A4000] |
Nvidia | GA104GL | 2,008,975 | 687,477 | 2.92 | 8 hrs 13 mins |
| 21 | GeForce RTX 3060 Lite Hash Rate GA106 [GeForce RTX 3060 Lite Hash Rate] |
Nvidia | GA106 | 1,764,775 | 620,985 | 2.84 | 8 hrs 27 mins |
| 22 | GeForce RTX 2080 Rev. A TU104 [GeForce RTX 2080 Rev. A] 10068 |
Nvidia | TU104 | 1,610,831 | 583,050 | 2.76 | 8 hrs 41 mins |
| 23 | Geforce RTX 3050 GA106 [Geforce RTX 3050] |
Nvidia | GA106 | 1,141,663 | 546,013 | 2.09 | 11 hrs 29 mins |
| 24 | RTX A2000 GA106 [RTX A2000] |
Nvidia | GA106 | 868,173 | 568,993 | 1.53 | 15 hrs 44 mins |
PROJECT FOLDING PPD AVERAGES BY CPU BETA
Data as of Tuesday, 14 April 2026 06:35:29|
Rank Project |
CPU Model |
Logical Processors (LP) |
PPD-PLP AVG PPD per 1 LP |
ALL LP-PPD (Estimated) |
Make |
|---|