Tuesday, November 25, 2008

Human annotation of sequences

The "Metagenome Annotation Using a Distributed Grid of Undergraduate Students" is a nice article in PLoS Biology, it remind me than the annotation of sequences is a common problem and hard to implement automatically by programs and a human can "decide" better than a machine (what about AI for sequence annotation?).

I like the article tittle using "distributed grid" terms, maybe it is also considered as "heterogeneous nodes". LOL

So, the strategy is to mix some students, computers with internet access, sequences and a control version system (validated by a supervisor), the work flow is:

Step 1: screen a sequence

Step 2. Validate the annotation

And the best part you can resolve 2 problems in one hit: 1. annotate your sequences, 2. teach the students how to use bioinformatic tools and annotation.

Some years ago, I participate in a similar version of "annotathon", more simplistic because was based in the sequence homology in DBs, but it was very fun.


  1. Grid of un-payed slaves? Hmmmm Ciencias Genomicas? or regular/decaff students?

  2. It's your choice, remember to weight the skills to correctly distribute the jobs.

    In this case, I see some rare cases:
    a) the node is upgraded to supervisor
    b) the node never run correctly (made it dispensable)
    c) nodes co-operations can bring babies. :)