This is a test version of Biostars. For the public version, visit https://www.biostars.org.
Human genome build

How different is Human Ensembl Grch37 build 65 from build 75?

I am asking this because I realised that the methylation data I am using was mapped using build 65 and I have used the gtf and fasta files of build 75 for RNASeq, from iGenomes. The annotations were available with the pre-built bowtie2 index in iGenomes so I used that build anyway.

Thanks.

genome annotations ensembl

Probably not a lot (if any). There may be patch differences but they do not change the coordinates.

I think the "builds" you are referring to are the releases of Ensembl. Ensembl release 54 got hg18, so between 54 and 75 the genome build was hg19. Release 76 brought GRCh38, which is the current assembly.

Hi genomax2, thanks for your reply. I am talking about the releases, sorry for the confusion. The nomenclature is different in case of different sources of annotations so I did not refer to hg18 or hg19 because I am using annotations from Ensemble. So yes, Grch37 will be technically hg19 build(accept minor differences).

How different are these two releases in terms of annotations?(65 and 75)

1 answer

If you're using only the primary assembly, then the genome is the same. If you're using the haplotypes and patches, they will change between release 65 and 75.

The genes will be different as they are improved with every release.

Thanks Emily. I have used release 65 with methylation data annotations and release 75 with RNASeq data. To keep everything in sync can re annotate the methylation data,which was used with GRCh65 release 65, with release 75 ? I am using only primary assembly not the patches and haplotyes. Hope this should be the right way.(Because I since the sequence level information isnt changed, I would just like to re-annotate the data rather than remapping and then annotating it with 75 release) Thanks.

If you've just aligned the methylation to the genome, no problem. If you've done any downstream analysis where you compare the methylation to genes, you should match it with the RNASeq.

Log in to answer this question.