The challenges of interpreting phosphoproteomics data: A critical view through the bioinformatics lens

Abstract

During the last decade, there has been great progress in high-throughput (HTP) phosphoproteomics and hundreds or even thousands of phosphorylation sites (p-sites) can now be detected in a single experiment. This success is attributable to a combination of very sensitive Mass Spectrometry instruments, better phosphopeptide enrichment techniques and bioinformatics software that are capable of detecting peptides and localizing p-sites. These new technologies have opened up a whole new level of gene regulation to be studied, with great potential for therapeutics and synthetic biology. Nevertheless, many challenges remain to be resolved; these concern the biases and noise of these proteomic technologies, the biological noise that is present, as well as the incompleteness of the current datasets. Despite these problems, the datasets published so far appear to represent a good sample of a complete phosphoproteome of some organisms and are capable of revealing their major properties.