[EMBOSS] EDNAFULL

[pmr at ebi.ac.uk]

Ryan Golhar writes:

> Does anyone know know how EDNAFULL was derived?  Ie, How were the scores
> it uses determined?
>
> Google Scholar didn't turn anything up and I'm not finding anything in
> the EMBOSS documentation...

We should add more information to the header in the matrix file. Will do
for the next release.

The DNA matrices are simply the ones supplied by NCBI in
ftp://ftp.ncbi.nih.gov/blast/matrices/ as files NUC4.2 (EDNAMAT) and
NUC4.4 (EDNAFULL).

NUC4.2 simply scores 5 for a match, and -4 for a mismatch.

NUC4.4 scores 5 for a match, but provides appropriate scores for ambiguity
codes so that, for example, R:A scores +1 (rounded up average of -4, -4,
5, 5)

The matrices were created by Todd Lowe in 1992. I am not sure where Todd
was working at that time.

Hope that helps,

Peter Rice