Introduction

The TimeXNet website provides access to the TimeXNet algorithm which predicts gene networks and pathways activated during a cellular response using time-course biological data in the context of a large molecular interaction network. TimeXNet accepts biological data in the form of fold change values of gene expression, protein expression and protein phosphorylation over time.




TimeXNet implements an algorithm based on a minimum cost flow optimization approach to predict activated pathways in a large molecular network between three sets of genes – initial, intermediate and late response genes based on their time of change.


As shown in the figure below, the TimeXNet algorithm connects genes A, B expressed at early time points to genes I, J expressed at late time points through genes E and F, which are expressed at intermediate time points. In the process, TimeXNet also identifies genes C, D and G, H which themselves do not show any change in expression but are important in connecting the genes that do. In the case of gene expression datasets, these are usually proteins that show changes in phosphorylation.




TimeXNet requires two main sources of information:

  1. Fold change values of gene expression, protein content or phosphorylation levels over time. The genes/proteins can be partitioned into three groups by the time of their highest observed fold change and scored based on the amount of this change.

  2. A large molecular network containing protein-protein, protein-DNA interactions and post-translational modification with edges weighed according to their reliability. The web-based version can directly access highly curated protein-protein interactions networks for multiple species from the HitPredict database.



TimeXNet has no species restriction and has been extensively evaluated with other similar tools. It is also available as a downloadable Java application.


References:

Linking Transcriptional Changes over Time in Stimulated Dendritic Cells to Identify Gene Networks Activated during the Innate Immune Response. Patil A, Kumagai Y, Liang KC, Suzuki Y, Nakai K. PLoS Comput Biol. 2013 Nov;9(11):e1003323. PDF

TimeXNet: Identifying active gene sub-networks using time-course gene expression profiles. Patil A and Nakai K, BMC Systems Biology, 8(Suppl4), S2, 2014. PDF