ようこそ!

Welcome to Martin Frith's homepage at the CBRC

I am a computational biologist working at the CBRC, which is part of AIST. The CBRC is on the island of Odaiba, a futuristic entertainment district near central Tokyo.

Long term aim: decipher genome sequences

	Bible	Human genome
Age	3 thousand years	3 billion years
Length	4 million letters	3 billion letters
Technology	Ark	Brain

Genomes are palimpsests of unthinkable antiquity, which hold the secrets to technology more advanced than any achievement of human civilization. We live in exciting times: genomes have been sequenced only recently, and we have barely begun to decipher them.

Genomes will not be deciphered by one breakthrough, but by incremental steps. In my small way I have contributed to this, for example by discovering a code for transcription initiation, and finding short proteins.

Short term aim: proper sequence comparison

Since biological sequences first became avaliable, the main way of deciphering them has been by comparing them to each other, to look for patterns in the similarities and differences. Until recently, I thought that pairwise sequence comparison was a thoroughly solved problem. It is not. I am working on these questions:

How can we compare sequences with highly biased letter abundances, such as malaria genomes? Current methods do not work well.
How can we tell significant similarities from chance similarities? In theory, this is solved by the BLAST statistics. In practice, the BLAST statistics are confounded by simple repeats, and standard repeat-masking methods do not work.
How can we compare multi-gigabase sequence data (genomes, short reads)? Previous methods can do this only with severe restrictions (either heavy repeat-masking, or restriction to almost exact matches). We believe we have the first real solution: LAST.

Join me

I welcome postdocs and visitors to come and work with me, but you would probably need your own funding. Likely sources include JSPS and HFSP, and there will be others depending on your nationality and other circumstances. Here is a funding guide for Europeans (pdf). Strong quantitative skills are desirable, e.g. from a background in physics or mathematics. Knowledge of biology is not essential, but willingness to learn about and deal with messy biological details is. Here are some project ideas, although original projects are especially welcome. Knowledge of Japanese is not necessary. 日本人も歓迎です。 You can apply to work at the CBRC here.

Alumni

Naoki Osato. To: University of Massachusetts

Publications

You can find most of my published articles by searching PubMed for Frith MC. (A few are just Frith M.)

History

I previously (2004-2006) worked half-time at the IMB, University of Queensland, and half-time at GERG, RIKEN. Before that, I received a Ph.D. in bioinformatics from Boston University (1999-2003), supported by a Howard Hughes Predoctoral Fellowship. Even further back, I survived the Part III Mathematics course at Cambridge University, and got an undergraduate degree in Physics and Philosophy at Oxford University.

Motif analysis

One approach to deciphering sequences is to study motifs (i.e. recurring sequence patterns). Quite recently I developed a method, GLAM2, which can discover motifs with arbitrary insertions and deletions. It is now part of the MEME suite. Less recently, when I was in the Zlab, I made a few motif analysis tools that might still be useful: Clover, Cluster-Buster, GLAM, Repfind, site2genome.

Miscellaneous

Links to websites I find useful.
Some handy bookmarklets.

Contact

Email: martin followed by @ followed by cbrc.jp. This may change periodically to avoid spam. For personal email, please use my gmail.com address.
Address: AIST Tokyo Waterfront Bio-IT Research Building, 2-4-7 Aomi, Koto-ku, Tokyo, 135-0064, Japan. Access.
Tel: I prefer email. Fax: +81-3-3599-8081

Last modified 2010-02-19