Massaging sequence data

[Matt Rockman]

I have a large matrix of aligned sequences and I'm analyzing it in a
variety of ways. One approach I would like to try is parsimony excluding
silent transitions. I have seen this done by 1) excluding all third
position transversions and 2) recoding leucine codon first positions as R
and Y. 
My question is pretty simple and mechanical: is there any shortcut to
recoding the leu first positions? My matrix is huge and I am not eager to
sort through looking for leu codons and changing the first position. Does
anybody know of a trick for doing this?

(please respond by reply email)
Thanks,
Matt Rockman
mrockman at minerva.cis.yale.edu