Short Co-occurring Polypeptide Regions Can Predict Global Protein Interaction Maps

Abstract

A goal of the post-genomics era has been to elucidate a detailed global map of protein-protein interactions (PPIs) within a cell. Here, we show that the presence of co-occurring short polypeptide sequences between interacting protein partners appears to be conserved across different organisms. We present an algorithm to automatically generate PPI prediction method parameters for various organisms and illustrate that global PPIs can be predicted from previously reported PPIs within the same or a different organism using protein primary sequences. The PPI prediction code is further accelerated through the use of parallel multi-core programming, which improves its usability for large scale or proteome-wide PPI prediction. We predict and analyze hundreds of novel human PPIs, experimentally confirm protein functions and importantly predict the first genome-wide PPI maps for S. pombe (∼9,000 PPIs) and C. elegans (∼37,500 PPIs).

ProfPPIdb: pairs of physical protein-protein interactions predicted for entire proteomes

“…Our new methods and new predictions 57 at least double the number of organisms for which sequence-based PPI predictions are 58 available, and they do this in a more consistent way than other method [24]. On top,…”

“…As before in [20], [23], we obtained Many advanced sequence-based PPI prediction methods have been developed. Park and 112 Marcotte [22] showed that PIPE2 [24], AutoCorrelation [31], and SigProd [32] 113 performed well compared to other methods. We showed a profile-kernel SVM to improve 114 over these methods for human and yeast [23].…”

A deep learning framework for improving protein interaction prediction using sequence properties

“…Thus, there is a longstanding interest in using sequence-based methods to model and predict protein interactions (Hashemifar, et al, 2018). Several sequence-based methods have been developed to predict PPIs, such as SigProd (Martin, et al, 2005), AutoCorrelation (Guo, et al, 2008), Yang's work (Yang, et al, 2010), Zhou's work (Zhou, et al, 2011), PIPE2 (Pitre, et al, 2012), You's work (You, et al, 2014), PPI-PK (Hamp and Rost, 2015), Wong's work (Wong, et al, 2015), Sun's work (Sun, et al, 2017), and DPPI (Hashemifar, et al, 2018). The main limitation of these sequence-based methods is only the amino acid compositions of a sequence were considered.…”

Evolution of protein-protein interaction networks in yeast

“…PIPE has been used to identify novel protein interactions, to discover new protein complexes, to predict novel protein functions [32–36], and to produce proteome-wide predicted interaction networks for S . cerevisiae [34], Schizosaccharomyces pombe [33], Caenorhabditis elegans [35] and Homo sapiens [36], among others.…”

“…We do note that interactions from one species can be used to predict interactions in another, even using distant relatives such as human and yeast. However, within-species interactions have been found to be more accurate [33]. …”