In article <9308311856.AA03849 at genetics.com>
mcolbert at GENETICS.COM writes:
>I am looking for information on neural network packages that can be used for
>protein and nucleic acid sequence analysis. Specifically, we have novel
>sequences and want to see what (if anything) they may be similar to within the
>databases. The colleague I am asking this question for has already run
>traditional types of homology searches (blast,fasta) with not much success.
>The profile analysis offered in the GCG package is more sensitive than a
>traditional homology search, but creating a profile requires having multiple
>sequences that you align and search with, and we have only one sequence at a
>time. I have done some work with neural networks (very limited) so we
>thought we would at least ask around.
A while ago I read an article in which they used neural networks to predict cod
ing and noncoding sequences in a piece of genomic DNA (I don't have the article
at hand but I can look it up if you want). This is the only use of a neural ne
twork in sequence analysis I am aware off. Please if someone else has more info
rmation share it with this group. It is a very interesting subject!
Henk van de Kamer
CHLAMY at SARA.NL