Welcome to WeFold!WeFold is about bringing people together through social media to create a community that works on advancing the computational prediction of the structures of proteins given their sequence of amino acids.
Proteins are the building blocks of life. They are an extended sequence of amino acids which quickly folds into a compact shape, called the native structure. These structure determines how the protein functions. Genomic sequencing projects have found millions of protein sequences but the experimental approaches used to determine their native structure are too expensive and time consuming to keep pace with those findings. A reliable computational method that can predict protein structure from its sequence will clear this bottleneck and have a significant impact on biology, bioinformatics, and medicine.
The Critical Assessment of protein Structure Prediction (CASP) series of experiments was initiated in 1994 to help advance and assess this field. Every other summer, CASP challenges its participants to submit predicted structures for about a hundred proteins, whose structures have recently been experimentally determined but not yet published. Independent assessors evaluate these submissions once the experimental structures become available. Significant progress has been made in these predictions but major roadblocks still remain.
WeFold was started in 2012 to bring together different groups that participate in CASP and to raise awareness outside the community and reach individuals and organizations that may become incubators of new ideas. The goals of WeFold are twofold: 1) to create hybrid pipelines from the combination of methods contributed by its participants and 2) to collaboratively identify and work on problems that hinder the advancement of the field.
The WeFold experiment has already produced exciting results in the last two CASPs, WeFold for CASP10 and WeFold for CASP11. Currently, we are preparing for CASP12. Please join us either as a CASP participant contributing methods/ideas that can be used for protein structure prediction, Machine Learning participant contributing methods/ideas that can be used for protein scoring, or general participant.
So far we've done 3 rounds of WeFold. Each time the preparation is hectic and the execution is overwhelming and, when CASP is finally over, it blows my mind to think of what we did together. There were lots of new ideas tried, a lot of models generated and shared, a lot of computer power used, and a lot of enthusiasm among the experienced and the young. Most importantly, we got to know each other and the work we do better and we really look forward to meeting in person those with whom we worked so hard during the summer.
I just want to give a big THANK YOU to all the WeFolders that have joined me in this collaborative effort to advance a field that we feel really passionate about and that has consumed and will continue to consume so many hours of our lives.
As a proof of all our effort, I'm posting (in no particular order) all the WeFold abstracts here.
- Subscribe to our site to stay updated via our email list and open forum!
- Want to participate? Please register here to join this year's competition as a new user.
- A subscriber with us already? Simply update your account to gain access to our past CASPs and/or participate in the current CASP!
- A CASP participant already? Login!
Open CASP Forum This forum is for discussion of ideas about protein structure prediction methods. This is an open forum and it does not involve CASP participation. The goal is to bring new ideas to the virtual table and to test them when they become mature for CASP.
CASP Experiment Forum: This forum is for those interested in participating in or directly contributing to the CASP12 experiment. It enables the interaction among groups that contribute different components of the protein structure prediction pipeline. The goal is to leverage expertise at a scale that would be hard to achieve otherwise.
Machine Learning Forum: This forum is for machine learning experts to discuss the protein scoring problem. The goal is to engage the ML community to help the protein structure prediction community to solve the difficult problem of selecting the best computationally generated protein models out of the hundreds of thousands or even millions that the protein structure prediction methods can generate per protein.