Hello Yongsheng, The newer assembly is probably the best choice as it is the genome with the most resolution. Keep in mind that there is still considerable work to finalize all of the tracks for the hg19 assembly. Meaning, that there can be a limited set of alternative (derivative or independently corroborative) annotation evidence when compared to the annotation evidence set for hg18.
Most importantly, look out for predicted proteins that were based on hg18, then aligned to hg19. For these there could be some inconsistencies: the hg18 version and the hg19 version may or may not align well. This will depend on if there was change in that region of the genomic assembly that affected the translation and/or if the RefSeq (or another protein track's data) has been updated or not. For a certain set of the gene footprints, you will need to come up with a method to sort through those proteins aligned to genomic sequence with differences. The hg18 version would be essentially incorrect if the hg19 changes are considered to be the "best". It is reasonable to assume that if one were to hand curate the current proteome, some of the hg18 protein versions would be better and some of the hg19 protein versions (as represented in browser tracks) would be better. New variants may be present, the canonical transcript selection may change, etc. Deciding how to define "best" or "better" is a scientific decision you will need to make based on the data and your experimental goals. We hope this helps, Jen ------------------------------------------------ Jennifer Jackson UCSC Genome Bioinformatics Group ----- "Yongsheng Bai" <[email protected]> wrote: > From: "Yongsheng Bai" <[email protected]> > To: "Jennifer Jackson" <[email protected]> > Cc: [email protected] > Sent: Monday, January 4, 2010 1:46:23 PM GMT -08:00 US/Canada Pacific > Subject: RE: [Genome] Question > > Hi Jennifer or Service Support desk, > > I realize there is a big difference in terms of genomic location > annotation > between hg18 and GRCh37(hg19) release version. Which one do you > recommend to > use if I want to check the protein coding sequence information? > > Thanks a lot! > > Yongsheng _______________________________________________ Genome maillist - [email protected] https://lists.soe.ucsc.edu/mailman/listinfo/genome
