Design and Prediction of Protein–Protein Interaction
||Design and Prediction of Protein–Protein Interaction|
||SNIC Medium Compute|
||Sinisa Bjelic <email@example.com>|
||2021-09-01 – 2022-09-01|
We are using computational modeling techniques to build proteins that interact with biologically relevant targets (proteins, receptors, transporters, ion channels etc.). Specifically for this project we plan to target transferrin receptor as a way to boost exposure of therapeutics across the blood–brain barrier (BBB). The receptor is a multidomain protein anchored in the BBB membrane involved in the physiological facilitation of iron uptake. Here I will develop protein carriers for transferrin receptor-mediated passage of the BBB by using computational protein design. Protein structures will be built from fragments, amino acid conformations will be sampled by rotamer search, and atom positions will be minimized w.r.t. a force field. We plan on generating millions of structures for efficient conformational sampling.
The long-term goal is to couple therapeutics — as for example drugs against Alzheimer’s — to the designed carriers to increase the brain uptake and cure neurological disorders.
We also use the resources for calculation of chemical modifications on proteins. Since we want to calculate probability of a chemical modification on a protein we need to do it on every amino acid in the sequence. The proteins we are working on in this project are kinases, and the large family leads to many calculations on related homologs.
In addition we are working on studying coiled-coil protein interfaces by Rosetta de novo folding techniques. Coiled coils are the simplest model of protein–protein interaction, consisting of two helical subunits, which are used in this proposal. We are going to generate a large binding dataset of dimeric coiled coils to assess the validity of our methods and to improve upon these. Mainly, our goal is to build heuristics to identify binders from non-binders for this set of proteins. The results will be used for better understanding of protein–protein binding interaction on molecular level.
We are using all the compute time allocated and would like to increase our allocation to 50000h/month as it has been limiting for our research.