Macromolecular Modeling Blog ™

The importance of secondary structure prediction to bioinformatics, modeling and structure prediction cannot be overstated. In a somewhat recent paper by Palopoli et al., their group found that combining the results of mutliple algorithms (with their JSSPrediction methodology) gave major improvements in prediction accuracy (see table at right). This ‘combine and conquer’ strategy is a tried and true strategy for multiple prediction problems in biological systems.

With this poll, we’d like to see who you go to for your secondary structure predictions and hear your thoughts. Are there algorithms not in the poll that you like? Do you combine algorithms? Do you like the servers or prefer in-house command-line installs? What are the challenges? Is an 88% prediction accuracy the ceiling or will we one day get to 95%?

For Rosetta, the standard is a mix of PsiPred, Jufo, SAM, and Prof.

 Loading ...

Amelie Stein, Roland A Pache, Pau Bernadó, Miquel Pons, Patrick Aloy (FEBS Journal, 2009)

This recent mini-review by Stein et al. focuses on the mechanisms that enable dynamic, transient, short lived interactions in cellular networks. Of special interest are the always popular “motif recognition domain”-”short flexible peptide” interactions. However, post translational modifications and regulation by disorder are also discussed. We concise the review further to some basic/interesting/anecdotal/”pondering worthy” points.

(more…)

PLoS ONE launched a new collection titled “Structural Biology and Human Health: Medically Relevant Proteins from the SGC” which makes use of three dimensional molecular animation technology.

The SGC (a public-private partnership created to place 3D structures of proteins of medical relevance into the public domain) have been able to take the original research articles published in the Collection and create ‘enhanced’ versions of them. As a result each of the research articles is now also available as an ‘interactive’ version, incorporating user manipulable, three-dimensional molecular structures.

Readers of these enhanced articles first need to download a plug-in for their browser but are then able to click on hyperlinked text within the article to ‘fly’ to the relevant position within the molecule, and to interact with it at will (by zooming, rotating, animating, and exploring). The seamless integration of interactive 3D structures into the actual text of the article provides considerable new functionality for readers, and it is hoped it will lead to new insights and discoveries.

This technology is simmilar to the Proteopedia ongoing project which enables the annotation of structural data with online animation using the wiki and Jmol frameworks.

It seems that the prevalence of both methods would only go as far as the authors of newly published papers would invest time to create these beatiful, educating animations.