[BioC] strange behavior with hsahomology

[Nianhua Li]

Hi, Francois,

> Sometimes it seems the scores are in homoURL instead of homoPS. It seems
> like a bug to me, or is it a feature I don't understand?

Yes, it is a bug. Thanks for the report.

>
> > homo <- unlist(as.list(hsahomologyDATA))
>  #one arbitrary case where this is the case
> > x<-grep("396050", homo)
> > as.matrix(homo[(x-2):(x+2)])
>                    [,1]
> 6337.9031.homoOrg  "gga"
> 6337.9031.homoType "m"
> 6337.9031.homoHGID "396050"
> 6337.9031.homoPS   NA
> 6337.9031.homoURL  "66.84"
>
> There seems to be a lot of cases where this happens,

I can bet that you see the problem when and only when homoType=="m". Anyway, I have updated all the
homology data packages (30 of them) in bioc 2.0 repository. The source packages and windows binaries
in bioc 1.9 should be available before the end of the day. You will get something in "homoURL" only
when homoType == "c" and in that case homoPS will be NA. In all other cases, "homoURL" will always
be NA. In hsahomology, homoType can only be "m", "B" or "b", i.e. homoURL is always NA.

> and it would seem
> to me that most of the code using those scores would be rather confused
> by this.

Hope those codes are not confused any more. Now, seriously, do you know any packages that use those
scores? I am curious about this because the source data that we used to generate hsahomology will be
deprecated in Jan, 2007. It is not easy to get those scores from the new data format. So, I am
curious how heavily this data is being used. Thanks.

nianhua