I think this is a potential "Nodalpoint paper of the month". Using MD5 hashes of sequence fragments to create unique identifiers (Smith et al., 2005).
We present a rapid and efficient method to map sequence identifiers across databases. The method uses the MD5 checksum algorithm for message integrity to generate sequence fingerprints and uses these fingerprints as hash strings to map sequences across databases. The program, called MagicMatch, is able to cross-link any of the major sequence databases within a few seconds on a modest desktop computer.
2005, 'MagicMatch -- cross-referencing sequence identifiers across databases.', Bioinformatics, vol. 21, no. 16, pp. 3429-30
Comments
Well, it's a useful paper
Well, it's a useful paper but the idea is around for a long time and I know that several applications use similar approaches beyond what Smith et al already mention in their paper. Still helpful and worth the publication but a paper of the month? Then again ... not that easy to find a universally interesting paper in the current issues of the bioinformatics journals.
contributions welcome :)
I think this is a snazzy little idea and was something that I personally hadn't seen before.
Regarding "paper of the month" - it's a pretty relaxed thing. More than one a month is fine and the title is not meant to imply that it must be an outstanding publication - it can just be something that caught your eye for whatever reason, good or bad.
I like methods with some application over "just methods" too - so send 'em in!
You're right, bioinformatics
You're right, bioinformatics being what it is (i.e. non-specific), it is not easy to find a "universally interesting" paper.
While MD5 hashing strings to create unique ids is not new, I haven't seen it used in the way they propose in the paper. Although their performance comparison seems a little bit too good: >1000-fold and 61-fold speed-up. Pointers to other applications would be interesting (if you get the chance).
Papers that I find interesting are biased towards alternative splicing (my current research). I might consider posting some of my personal picks to my own weblog. In the mean time Steve Mount has a nice post on current papers analysing alternative splicing microarrays here.
I guess I could change the segment to "Microarray data normalization method of the month"...
Universally interesting to me ...
I tend to prefer papers that describe surprising biological findings using bioinformatics over methodological improvements. However, let's not discuss whether a paper should appear in what section and what delimits bioinformatics, computational biology and statistics in the first place overly and focus on the content.
[Note to self: Stop commenting on it then...]