I would suggest thinking thru what it takes to set up a web site with information that a lab needs. Not everything needs to be known, but it gives an overview of what a company (and even labs) need. (see my web site to see what I've done http://www.mindfullresearch.com)
Hardware: learn some hardware (ie, can you change RAM or install a hard disk?, do you know the difference between an Athlon and Celeron chip? ) System administration (learn Linux or Unix) Networking (how to set up cable modems, DSL, and hook together a system of computers)
Web site: apache web services, php programming language (like HTML-pretty easy)
Database: Mysql or PostGreSQL (two open source (free) databases) or MS Access or Oracle. Learn SQL (Structured Query Language easy) load and extract data or view data from a database: programming language (Perl, Java)
Biochemistry and Molecular Biology and Cell Biology (so he knows the vocabulary)
Know some of the major sites and what information is
available and how to navigate these sites and where the data files are (in
the "download" areas, or "FTP" site).
NCBI, PubMed, Entre Gene,
Mouse Genome Informatics, Rat Gene database, UniGene,
GenBank
Pathways (see KEGG= Kyoto ???),
Gene Cards (at Weizman Inst in Israel),
Swiss-Prot (Expasy),
EMBL, Sanger Institute (PFAM, INTERPROT)
There are also virtual communities on line which can teach a lot, just by reading the pages (don;t worry, even I don;t understand most of it, but you learn alot by osmosis):
http://bio.perl.org/
http://biojava.org/
http://www.open-bio.org/
They talk about nuts and bolt issues, but also about main issues confronting most bioinformaticists today.
I would add two other things:
1) One of the most basic goals of Bioinformaticists (like me involved in databases) is to parse huge amounts of data and make connections amongst them. So one spends an awful lot of time examining ftp sites to see the structure of the files so that they can be downloaded and parsed. Then one has to determine how to match up data from one site to another. This is most commonly done by common IDs, but not all IDs can be matched up one to one.
"Aye, There's the rub."
2) Most of what is going ot be most valuable is not taught inthe classroom. It is learned by doing. If you and your computational colleagues could put together a course project of making a genomic (or proteomic or metabolomic) database on line, that would teach everything I described in my last email.
One other note, I am in the running for a job at the NCI in the caBIG program. I would be able to work form home (in Waco or wherever my home willbe).
see http://cabig.nci.nih.gov/
it would be exciting and would try to link cancer researchers and clinincians thru common tools and vocabulary. Just think, Ph.D.s and MDs talking the same language!!!
