Illegal instruction (core dumped) - Indexing a reference genome with STAR

Question

[compuTE]

Hello,

As the title suggests, I'm trying to index a reference genome (.fna and .gff of hg38 from the latest major release of https://www.ncbi.nlm.nih.gov/grc/human) with STAR aligner.

This is my script

#!/bin/bash
#SBATCH -A $group_name
#SBATCH -n 16
#SBATCH -N 1
#SBATCH -t 03:00:00
#SBATCH -J reference_genome_STAR_hg38
#SBATCH -o %j.STAR_reference.out
#SBATCH -e %j.STAR_reference.err

ml GCC/4.9.3-binutils-2.25
ml STAR/2.5.0a

STAR --runThreadN 16 --runMode genomeGenerate --genomeFastaFiles genome.fna --sjdbGTFfile annotation.gff --genomeDir ./reference_genome/ --sjdbOverhang 101

For some reason the job fails and I get

Illegal instruction     (core dumped) STAR --runThreadN 16 --runMode genomeGenerate --genomeFastaFiles genome.fna --sjdbGTFfile annotation.gff --genomeDir ./reference_genome/ --sjdbOverhang 101

On my stderr file.

Any help would be well appreciated, this is the first time I use this software and I'm pretty lost.

Note: I tested it with a smaller genome and annotation file and the problem persists.

[GenoMax]

@Alex Dobin also makes pre-made indexes available here, if that works for you.

[compuTE]

I will get it from here. Thank you! :)

[GenoMax]

How much memory are you assigning to the job? You would need at least 30+GB.

[Devon Ryan]

I'm not sure how they installed STAR, so try providing your own version (you can download a binary from bioconda). Perhaps they installed it in an odd way.