Training

IMPACT Training Workshop 3

InterPro for Functional Annotation - understanding the added value of InterPro's component databases

This 2-day IMPACT workshop focuses on understanding the InterPro database, how some of the member databases contribute to it and ultimately how InterPro gains from their input. This workshop will provide insights into the biological significance of the results obtained from using a particular database, and into the level of confidence that should be attached to their annotation. This will allow sensible interpretation of what that database tells us about a given family of sequences, and crucially, how this information differs between the databases, and why.
The workshop includes hands-on tutorials to shed light on some of the complexities, and to illustrate the strengths and weaknesses (and areas of optimum application) of the different approaches.

Target audience

Researchers using or planning to use InterPro.

Learning Objectives

This workshop is a valuable opportunity to learn how its component databases add value to InterPro, and hence how InterPro can add value to your analyses. The presentations and tutorials are augmented with user-driven practical sessions, providing further opportunities to explore InterPro with your own 'problem' sequences.

Background Information

InterPro is an integrated resource containing annotated signatures for a wide range of protein families. The signatures tend to be diagnostic of particular protein functions or structural features, making them useful tools for characterising unknown sequences, particularly from high-throughput genome sequencing projects.

Amongst the original contributor databases were the motif- and domain-based repositories: PROSITE, PRINTS and Pfam. Later, more explicit structural information was added via resources such as Gene3D. Today, InterPro represents an amalgam of 11 databases, each of which provides uniquely specialised insights into the protein families they encode.
Understanding how these different components are brought together to provide a coherent picture of family relationships can hence be a bewildering task.

Each of the partner databases uses a different diagnostic approach: some use single motifs (encoded as 'regular expression' patterns), some exploit multiple-motifs (encoded as 'fingerprints'), and others use complete domains (encoded as HMMs or profiles). Each of these methods performs differently, sometimes in ways that are not obvious to those who don't use them routinely; but these differences are important, as they ultimately determine what we see, and how we interpret our results.