Phylogenetic Transfer of Knowledge for Biological Networks

Advances in biotechnology have enabled researchers to study molecular bi- ology from the point of view of systems, from focused efforts at functional annotation to the study of pathways, regulatory networks, protein-protein interaction networks, etc. However, direct observation of these systems has proved difficult, time-consuming, and often unreliable. Thus computational methods have been developed to infer such sys- tems from high-throughput data, such as sequences, gene expression levels, ChIP-Seq signals, etc. For the most part, such methods have not yet proved accurate and reli- able enough to be used in automated analysis pipelines. Most methods used to infer biological networks rely on data for a single organism; a few attempt to leverage ex- isting knowledge about some related organisms. Today, however, we have data about a large variety of organisms as well as good consensus about the evolutionary relation- ships among these organisms, so that the latter can be used to integrate the former in a well founded manner, thereby gaining significant power in the analysis. We have coined the term phylogenetic transfer of knowledge (PTK) for this approach to inference and analysis. A PTK analysis considers a family of organisms with known evolutionary re- lationships and "transfers" biological knowledge among the organisms in accordance with these relationships. The output of a PTK analysis thus includes both predicted (or refined) target data for the extant organisms and inferred details about their evolution- ary history. While a few ad hoc inference methods used a PTK approach almost a dozen years ago, we first provided a global perspective on such methods just 6 years ago. The last few years have seen a significant increase in research in this area, as well as new ap- plications. The time is thus right for a review of recent work that falls under this heading, a characterization of the solutions proposed, and a description of remaining challenges.