This technology is an algorithm that uses structural similarities between proteins to determine if two proteins will interact.
Determining the likelihood that proteins or other small molecules may interact is critical for understanding biological processes and developing targeted therapeutics. However, current approaches for identifying these interactions are costly and low throughput. Therefore, there is a need for improved methods capable of integrating various data sources to generate more accurate predictions.
This technology uses structural similarities between proteins to calculate structure-based likelihood ratios (LRs) to determine if two proteins will interact. Using Bayesian statistics, PrePPI also incorporates independent LRs from non-structural information. This current algorithm includes updated sources and features capable of predicting the interaction between structured domains and unstructured peptides. The PrePPI database contains over 1.35 million reliably predicted functional interactions with 130-500 thousand predicted physical interactions. Unlike other computational databases, this technology can provide predictions beyond the available scientific literature or known protein sequences, pointedly guiding experimental design, small molecule screening, and drug development.
This technology has been validated as a working software.
IR CU21167
Licensing Contact: Cynthia Lang