Greetings Netfolk!,
We have been trying to do a limited evaluation of the utility of Smith and
Waterman searches on massively parallel hardware (MasPar MP-1 and MP-2) using
both BLAZE (from Intelligenetics) and Mpsrch (from Shane Sturrock and John
Collins at the Biocomputing Unit at Edinburgh University).
I was wondering if anyone has done/intending to do something similar? More to
the point has any independent assessment been carried out that anyone is
intending to publish? I have seen some data from Intelligenetics which is very
impressive but then again hardly impartial.
Has anyone any experiences they feel are relevant - for example how necessary is
the rigour of the SW algorithm when most people have been happy to get by with
BLAST and FASTA to date? Is it only a matter of time (with the masses of data
being generated by automatic DNA sequencers) when data parallel solutions will
not only be desirable but essential? Is it not the case that the intellectual
assessments of the results (from whatever method) is the rate limiting step in
sequence analysis?
Any and all responses welcome either to the group or by email - I will
summarise and post responses if there is sufficient interest.
Cheers,
Tony Parsons
(mbpcr at seqnet.dl.ac.uk) or preferably for mail (parsons_a at snd01.pcr.co.uk)
