Bioinformatics

Keith Bradnam keith at evol.gene.nottingham.ac.uk
Fri Mar 29 23:15:35 EST 1996


Leonard F. Kuehner wrote:
> I have searched high and low for some information on bioinformatics and have
> so far come up with nothing....snipped...Does anyone in this group know what 
> bioinformatics is exactly.  

Well seeing as I did my MSc in Bioinformatics and am now doing my PhD in it, 
I guess I should know.
Taken in its most literal sense, bioinformatics deals with biological information.
The information in question is DNA, RNA, and protein sequence information.

So for DNA and RNA you can treat sequences of nucleotides (A,T,C,G and U) as 
a string and analyse it in various ways.  For proteins, you have 20 possible
amino acids to look at.  At the end of the day, all the bioinformatician (if that
is the word?) deals with is this raw sequence information, usually stored as
a string on their computer.

As these sequences are now stored extensively in protein and nucleotide
databases (e.g. nearly 700,000 nucleotide sequences now in GenBank) you can
analyse them directly by computer and there are loads of different bits of
software to do this.

Bioinformatics is relatively new, as sequence databases only started accumulating
from the start of the 80's.  It's approaching a new era, as such, as for the
first time, complete genomes are now available for analysis, so now you can
start to tackle questions like "what sort of information is needed to specify an 
individual, or even to specify life?"

> What would someone in the field do, 
On the protein side of things, the big thing, indeed the holy grail, is to
try and use a protein sequence (and nothing else) to try and predict the
3-dimensional structure of the protein.  And you can get quite far with that.
You can also make predictions of protein funcion, membrane spanning regions, 
loop structure, possible antigens and loads of other stuff.
You can also make structural predictions for RNA. For DNA the big thing is
to look at evolution by seeing how sequences have diverged between organisms.
Also you can look at what influences codon usage in different species, or within
different genes within a species (which is what I'm doing now).
You can also look to see what non-coding DNA (95% of the human genome, approx.) does
and how that relates to other genomes.  In many ways, you can start to test
what have only been theories or hypotheses, regarding genetics and evolution.

A very quick example....the third base of each codon in a gene, does not seem to 
determine which amino acid that codon encodes for.  I.e. GGU encodes for glycine
as does GGC, GGA and GGG...in short GGx encodes for glycine.
Early work on the neutral theory suggested that the use of the third base should
therefore be largely random.  However it is not, and sequence analysis has taught us
loads about why the particular choice of third base may be important.

I have some links to various bioinformatics sites at:

http://evol.nott.ac.uk/~pdxkrb/

> and what sort of background
> does one need to work in the field?  

Anything from genetics, molecular biology, computing, statistics,
or in my case...ecology!  There are quite a few different perspectives on 
it, so you can bring different skills to the subject.
Hope this helps

Keith B.
 
  -_-_-_-_-_-_-
    -_-_-_-_-	 We came, we saw, we didn't care for anymore
      -_-_-
	-




More information about the Comp-bio mailing list