How should I choose? Decoy or not?

Question

[scchess]

Whe I goto: ftp://ftp.broadinstitute.org/bundle/b37/, there are two FASTA files I can use for alignment:

• human_g1k_v37_decoy.fasta.gz
• human_g1k_v37.fasta.gz

The two files are very close in file size. As far as I understand, the decoy version is slighter faster. Other than that, anything else I should consider?

I have a human sample, how should I choose? Speed is not a particular concern for me.

Answer 1

[h.mon]

Oddly, I get an error 550: no such file or directory when trying to reach ftp://ftp.broadinstitute.org/bundle/b37/. Anyway, here (at "Which version should I use?") you will find a short explanation about the decoy genome, and here, a longer explanation.

[kmzhou4]

Is the decopy https://www.ncbi.nlm.nih.gov/assembly/GCA_000786075.2/ released almost 10 years ago incorporated into the current release https://www.ncbi.nlm.nih.gov/assembly/GCF_000001405.40 (year 2022). Is it say to concatenate the above two? Or has anyone done the combination and renaming of from NCBI accessiont to chr__ format?