If you’re a frequent visitor to our website or this blog, you’ll know we talk a lot about accuracy here. When it comes to your research, we understand that obtaining accurate and reliable results is paramount in deciding which software to use. So how accurate is our protein structure prediction program, NovaFold? And how can one test the accuracy of a protein model when the experimental structure is unknown? We can begin to answer these questions by looking at results from the Critical Assessment of Protein Structure Prediction (CASP) experiments.

Every two years, research groups around the world participate in the CASP competition to test their prediction tools against unpublished protein structures in a series of blind studies. The goal of these experiments, as stated on the CASP website, “is to obtain an in-depth and objective assessment of our current abilities and inabilities in the area of protein structure modeling.”

The I-TASSER algorithm, developed by Professor Yang Zhang at the University of Michigan and referred to as the Zhang-Server in the CASP competition, has taken first place in each of the past three competitions. This unique algorithm utilizes a combination of “threading” and “ab initio folding” to predict protein structure, resulting in highly accurate protein models, even in the absence of similar template structures. This is the same algorithm used by NovaFold, which can be accessed through Protean 3D, on our new cloud app, or on a Linux server.

Read our latest white paper to learn more about the CASP competition and how NovaFold utilizes I-TASSER to produce accurate protein models. Then, check out NovaFold on the Cloud to try a structure prediction for yourself. Academic and government users can predict their first structure free!