How can I access the consensus sequences of all human ERVs from Dfam?

Question

[mbk0asis]

Hi, everyone!

I'm trying to estimate the expression levels of ERVs in human by mapping the reads on the consensus sequences.

I found a paper that used consensus sequences from Dfam, but I couldn't find the exact file.

I found some files 'hg38_dfam.hits.gz' and 'hg38_dfam.nrph.hits.gz' which had only location information without actual sequences.

Should I extract sequences using the location info, or is there other way?

Can someone give me some hints?

Thank you!

Answer 1

[mlbendall]

You can get the Dfam consensus sequences by downloading the HMM (i.e. DF0000558.hmm) and using hmmemit from the hmmer package:

hmmemit -c DF0000558.hmm > DF0000558.consensus.fasta

[mlbendall]

With the latest version of Dfam there is an API you can use to access the consensus sequences without downloading the HMMs. You would use the /families/{id}/sequence endpoint.

See http://dfam.org/help/api and https://dfam.org/releases/Dfam_3.0/apidocs/

Answer 2

[h.mon]

There are (at least) two pipelines for HERV quantification, TEToolkit ans SalmonTE, you could look into them. What is the paper using consensus DFam?