Collins Lab Blog

The ‘ome and me

“The sequencing of multiple genomes from diverse species has tremendous potential to impact our understanding of biology, both by providing a census of all proteins and by enabling subsequent analysis of their functions.”

A concise statement (understatement?) from Science. 2004 Dec 24;306(5705):2246-9. As a molecular cell biologists we hope to tap into this huge potential to illuminate our area of focus and enable new discoveries that we can test experimentally.

It is a constant source of frustration for us to be aware of the most robust tools to approach the ‘omics potential. As an example, the lab recently laboriously purified two proteins that were ‘strongly predicted’ to physically associate with each other. In our hands the two proteins had no affinity for each other, an annoyingly negative result, made only slightly more meaningful by the fact that we could experimentally recapitulate other previously verified protein interaction partners for the two proteins.

It is not apparent to me if such strategies are being successfully employed by other groups. For instance, take the study from which the introductory quote is taken, Bowers et al. (2004) Use of Logic Relationships to Decipher Protein Network Organization 306:2246-9. This study found 750,000 previously uncharacterized relationships between protein families using very clever logic analysis of phylogenetic profiles to identify functional correlations between proteins.. Fast forward to today, 2012, and it is not clear if the data reported was subsequently used to generate novel hypotheses that were tested experimentally, or as a jumping off point for the experimental verification of new protein functions.

In addition to bioinformatic probing, experimental ‘omics databases proliferate, holding out the tantalizing hope that a small research group can successfully mine the database jungle to pick off a crucial fact that could provide a key to unlock a mystery. I was reminded of this fact today when a student stopped by my office to discuss a Y2H screen that he was getting ready to perform. For the particular protein this student wanted to query, there were no obvious positive controls and so I suggested one approach would be to simply check out some ‘omics screens for possible targets that could be rapidly checked against the prepared bait.

For S. cerevisiae proteins, the SGD does a superb job of collating published information. For other model systems the landscape is more disparate. Some of the approaches underway are incredibly impressive. Take for instance the Allen Brain Atlas, or the Human Protein Atlas, which has analyzed ~12,000 human proteins with specific antibodies. These efforts are akin to the man-on-the-moon effort. The payoff in science, education and medicine should be just as fundamental.

Comments

Leave a Reply

You must be logged in to post a comment.