GPS

GPS is a software tool designed to predict protein-DNA interaction events with high spatial resolution from ChIP-Seq data. It was developed to address the challenge of detecting closely spaced homotypic events, key components of invertebrate and mammalian promoters, and enhancers. The random variation in the ChIP fragmentation process makes it difficult to identify the exact location of these events. Still, GPS can overcome this challenge by modeling observed reads using a complexity penalized mixture model and predicting event locations with a segmented EM algorithm.

GPS also has an optional mode that allows it to align common events across distinct experiments. This feature enhances the accuracy of the predictions made by GPS, making it a more reliable tool for researchers. Moreover, GPS has been shown to detect more joint events in synthetic and actual ChIP-Seq data compared to other methods.

One of the key advantages of GPS is its superior spatial resolution. This means it can accurately predict the location of protein-DNA interaction events even when they appear as a single cluster of reads.

GPS is now a part of GEM.