generating genome indexes with STAR

Question

[snp87]

Hi, I am trying to generate mouse geneome indexes with STAR to align my RNAseq data. I downloaded the mm10 genome in tar.gz format (Mus_musculus_UCSC_mm10.tar.gz). Do I have to gunzip this first before trying to index the genome? Thank you!

Answer 1

[Devon Ryan]

Yes, you'll need to untar it first.

[snp87]

Thanks so much! After untar-ing when I tried indexing, I don't think it worked.

The command used:

STAR --runThreadN 4 --runMode genomeGenerate --genomeDir Genome_data/star \
--genomeFastaFiles Genome_data/Mus_musculus_UCSC_mm10.tar.gz

Result: Jun 23 14:20:03 ..... Started STAR run Jun 23 14:20:03 ... Starting to generate Genome files Killed: 9

I don't think it was indexed. Do you know what the problem might be?

[Alastair Kerr]

You are still using the tar archive for the fasta files. Use the fasta files extracted from the tar archive

[GenoMax]

STAR is known to require plenty of RAM (30+G) for operations. How much memory do you have access to?

Answer 2

[DareDevil]

You can try the following:

STAR --runThreadN 4 \
--runMode genomeGenerate \
--genomeDir /path/to/genomeDir \
--genomeFastaFiles /path/to/genome/ref.fasta \
--sjdbGTFfile /path/to/annotations.gtf \
--sjdbOverhang 149

--sjdbOverhang is ReadLength-1

Answer 3

[aleferna]

I had the same problem, for me it was becuase I was using a database of all human viruses as opposed to a genome. This had too many contigs, I read in a google post to increase the genomeChrBinNbits 14, that did it.