bizarre LINKAGE results - bug ?

I. Fenton ifenton at crc.ac.uk
Fri Nov 11 08:59:39 EST 1994


hello fellow-linkage people,

ok, here's the deal.  i've got a PEDIN.DAT file with a few families in
them, and my DATAIN.DAT file, i run a simple 2-point MLINK run on them
on my pc, and all is lovely except that MLINK gives the last family
a -infinity lod score at 5%,10%,20%,30%,40% recombination values,
*BUT* gives a proper likelihood (lod score = -0.7 or something)
at 0% and 1% distance.  what gives ?!?  i've seen -infinity across
the board before (usually i've cocked-up something in DATAIN.DAT)
but never seen the likelihoods fail at 5% or more, but work at 0%
and 1% !?!

i am using the dos version of linkage, pulled of jurg ott's ftp-site
a few months ago.  note - i ran it on a sparc-station here, and 
sun-linkage gave me proper likelihoods, and UNKNOWN is always happy
so i don't think there is anything wrong with the compatibility
of the genotype data.  i think it's a bug (shock !  horror !) or
simply am i missing something obvious ?

the files are a bit big, so i won't post them to the newsgroup
unless asked, but for reference MLINK is compiled for 14 alleles
(i don't like recoding !) although there are only 9 alleles in
this large (and largely untyped) family, but all the other families
in the PEDIN.DAT file go ok.  also why does the sparcstation version
of linkage give a different result to the pc-version ??!

*ANY* comments/suggestions are welcomed..., esp. from jurg or any
of the other "official" LINKAGE shops...

thanks,

Iain Fenton, Cardiff, UK.




More information about the Gen-link mailing list