Input preparation

The main input of phuEGO is a tab-delimited text file with two columns of no headers. The first column contains protein uniprot IDs and the second column contains aggregated log2-fold change (LFC) or other scores for each protein. The score dictates the probability by which the random walk process will restart from the associated node, subject to influence by the scores of other nodes within the same layer (read more about the algorithm personalized pagerank <https://igraph.org/python/doc/api/igraph._igraph.GraphBase.html#personalized_pagerank>). User can view the example input data.

Aggregating phosphosites (active development)

A phosphoproteomics experiment typically produced a list of phosphosites with LFC between test and control samples. To perform network propagation using protein-protein interaction network, the phosphosites need to be aggregated into a protein list. In the main analysis of the publication, this is performed by filtering phosphosites on functional scores, and aggregating using maximal LFC. The users can follow the same practice, or adjust it where they deem appropriate.

The authors of phuEGO are currently developing additional functions for this step, which will be released in near-future update of the package. Besides functional score, predicted signs of regulation of the phosphosites on protein functions could also be used during the aggregation step.