FOLDING PROTEIN BY LOCALLY-CONSTRAINED MINIMIZATION OF A REDUCED STATISTICAL PO

点击翻译按钮获取中文摘要

This subproject is one of many research subprojects utilizing the resources provided by a Center grant funded by NIH/NCRR. The subproject and investigator (PI) may have received primary funding from another NIH source, and thus could be represented in other CRISP entries. The institution listed is for the Center, which is not necessarily the institution for the investigator. The current Teragrid resource allocation for the Sosnick group runs from 08-26-2005 to 08-31-2006. Since we are on track to make use of all of our remaining hours before this time, we would like to request an additional Development Resource Allocation of 30,000 hours. We have used our current allocation to test our program for folding proteins by locally-constrained minimization of a reduced statistical potential. Our reduced representation of protein structure includes only backbone heavy atoms and side chain beta carbons. Our sampling method includes nearest neighbor effects in a discrete simulated-annealing minimization of a statistical potential in dihedral angle space. Torsional angles are constrained to the native basins of the target structure. Due to this structural constraint, our folding algorithm is ideal for testing the foldability of designed protein sequences for which the target structure is known. In fact, we have used our folding routine on the Teragrid to compare the results of our computational folding to Ranganathan's experimental folding of designed WW domain sequences (Socolich et al., Nature. 2005, 437(7058):486-7). Our initial results are extremely encouraging, and we require additional Teragrid resources to complete the project. Highlighting the importance of this need, the experimental results that took many human hours and resources were accomplished computationally on the Teragrid in a short amount of time with modest human effort. With a continuing allocation we would like to apply this method to the folding of larger protein sequences that we are designing using our aforementioned reduced statistical potential. We are currently generating a large number of designed sequences, but we need computational resources so that we can test their suitability to fold into the target structure. We will apply our folding routine to screen these sequences for their ability to fold and then experimentally verify any results by making the proteins in the lab. The ability to rapidly design and fold large numbers of designed sequences in silico would be an outstanding technical accomplishment in the field of protein design. It is thus critical to this research that we have continued access to the computational resources provided by Teragrid.