Dear all,
Our lab have done some aCGH on dog cancer and we were able to identify
a few regions, we are interested in finding out what genes are there in
these regions . However, the dog genome is poorly annotated (only around
1300 genes have been annotated). So my idea is to convert the dog genome
regions into ortholog human regions and find out what genes are there in
those human regions. This way is not 100% accurate, but I am guessing 80-90%
of the time it does the right job.
I am using the liftover tool from UCSC genome browser (CanFam2 to
Hg19). If I set the "Minimum ratio of bases that must remap" to 0.8, about
30% of the dog regions can remap into human, and if I set the number to 0.5,
about 80% of the regions can remap. So my first question is "Is 0.5
considered too loose of a criteria"?
Also, if I allow "multiple regions", more regions are able to remap.
But the output are extremely fragmented (I went from 22regions, average size
of 100kb to 400 regions, average size of 0.5-2kb). The sum of the size of
the regions also went from 11.5mb to 0.34mb. So I lost most of the regions
somehow. Overall, I decided to go with "does NOT allow multiple regions".
Another thought I have is some of dog regions are big (over 1mb), if I
chop one big regions into 10 small regions (i.e. 10 regions around 100kb
each), would it help the "liftover tool" to remap more precisely? And if I
were to do this, what size should I start with?
My overall goal is to find out what genes are there in these dog
regions. Can you please let me know if I am on the right track, or maybe
there is a better way of approaching the problem?
Thank you so much for your time and effort!
Cheers
Washington
_______________________________________________
Genome maillist - [email protected]
https://lists.soe.ucsc.edu/mailman/listinfo/genome
