Here are some of the biggest breakthroughs mentioned in the provided references:
Exascale Simulations for SARS-CoV-2:
Conducted exascale simulations of SARS-CoV-2 spike protein, revealing dramatic spike opening and cryptic pockets, which have implications for drug design and understanding viral infectivity (Zimmerman et al., 2021) .
Ab Initio Protein Folding Simulations:
Achieved molecular simulations of ab initio protein folding for the NTL9 protein, providing insights into the protein folding process (Voelz et al., 2010) .
Markov State Models for Protein Dynamics:
Developed Markov State Models (MSMs) to study protein folding kinetics and dynamics, providing a framework to understand protein conformational changes over long timescales (Bowman et al., 2009; Lane et al., 2011) .
RNA Polymerase II Dynamics:
Investigated the dynamics of RNA polymerase II translocation at atomic resolution, elucidating mechanisms of transcription elongation (Silva et al., 2014) .
Ligand Modulation of GPCR Activation:
Used cloud-based simulations to reveal how ligands modulate G protein-coupled receptor (GPCR) activation pathways, advancing the understanding of GPCR function and drug targeting (Kohlhoff et al., 2014) .
Nanotube Confinement Effects on Proteins:
Demonstrated that nanotube confinement can denature protein helices, providing insights into the effects of nanoscale environments on protein structure (Sorin & Pande, 2006) .
Simulation and Experiment in Protein Folding:
Combined simulation and experimental approaches to reveal slow unfolded-state structuring in acyl-CoA binding protein folding, highlighting the interplay between simulations and experiments (Voelz et al., 2012) .
Advances in Markov State Models:
Improved coarse-graining and adaptive sampling techniques in MSMs, enhancing the modeling of biomolecular dynamics (Bowman, 2012; Zimmerman et al., 2018) .
Insights into Allosteric Sites:
Identified potential cryptic allosteric sites within folded proteins using equilibrium fluctuation analysis, suggesting new targets for drug discovery (Bowman & Geissler, 2012) .
GPCR Activation Pathways:
Revealed ligand modulation of GPCR activation pathways through extensive simulations, providing insights into receptor function (Kohlhoff et al., 2014) .
I don't see much in there. Doing the simulations is not the same as confirming the simulations.
The question wasn't did they do they simulation but rather was any major usable outcome validated. Seems very little if anything.
Yes quite a few as other commenters have indicated. Another good one is !boinc@sopuli.xyz. BOINC is an open source platform for volunteer computing that also has hundreds of scientific papers and citations under its belt. There are BOINC projects for medical research, space research, math, you name it, there's probably a BOINC project for it. Anybody can start a BOINC project and you choose which projects you contribute CPU/GPU time to. You can pick more than one at a time. You may recognize some of the people hosting BOINC projects: Large Hardon Collider, Max Planck Institute, University of Washington Institute for Protein Design, etc
It's basically a software you install to donate your computer's idle power (ie typicallywhen the CPU does nothing) to help scientists with the huge amount of calculations that simulating the folding of protein requires and that not even the best supercomputers can achieve alone in a reasonable time. It's distributed science. I'll admit I still don't grasp what folding means in the context of proteins. The programme folding@home started around 2000 and is still running ie you can still donate CPU time today
A protein is like a really long chain of simple monomers (amino acids), that you can think of as a long string of differently coloured beads. The ordering of the beads somewhat determines how the protein functions, but the major factor that determines it is how this long string is bundled up, i.e. "folded" (think of a ball of yarn).
A DNA sequence tells us the sequence of the amino acids in a protein, but tells us nothing about how it is folded. It is of great interest to compute how a protein will fold, given its sequence, because then we can determine how and why it works like it does, and use gene-editing techniques to design proteins to do the stuff we want. This requires huge amounts of computational power, so you get the fold@home project :)
If 10,000 people work one hour in one day, then that day still had 10,000 hours worked even if a day is only 24 hours.
Same with CPUs, many of them run on more than just one core.
I mean I had the programme running for 10 years, donating whenever it was on and idle or on a low load. It was not necessary to be perfectly idle, the programme could take just e.g. 50% of the CPU. (Famously on a core2duo running programmes that were not multi-core).
All in all my computer was probably on 12 hours per day