bwa-mem keeps failing to locate the index files

Question

[pairedttest]

Okay, so I am at my wits end here. I just want to run bwa-mem on some fastq files but bwa-mem does not seem to want to run despite everything being in the proper place. Here is what I have done so far. I have a few fastq files that I am trying to align to the hg38 genome. I ran bwa-index with the following script

#SBATCH --mem-per-cpu=200G

bwa index hg38.fa hg38

And it generated 6 files, all with the hg38 prefix. hg38.fa.amb, hg38.fa.ann, hg38.fa.bwt, hg38.fa.fai, hg38.fa.pac, hg38.fa.sa.

All of these files are in the same working directory as my fastq files. Then I tried to run bwa-mem with the following script

bwa mem hg38 4810_R1.fastq.gz 4810_R2.fastq.gz > 4810_aln.sam 
bwa mem hg38 4811_R1.fastq.gz 4811_R2.fastq.gz > 4811_aln.sam
bwa mem hg38 4812_R1_fastq.gz 4812_R2.fastq.gz > 4812_aln.sam
bwa mem hg38 4813_R1.fastq.gz 4813_R2.fastq.gz > 4813_aln.sam
bwa mem hg38 4814_R1.fastq.gz 4814_R2.fastq.gz > 4814_aln.sam
bwa mem hg38 4815_R1.fastq.gz 4815_R2.fastq.gz > 4815_aln.sam
bwa mem hg38 4816_R1.fastq.gz 4816_R2.fastq.gz > 4816_aln.sam
bwa mem hg38 4817_R1.fastq.gz 4817_R2.fastq.gz > 4817_aln.sam

But it keeps returning the error [E::bwa_idx_load_from_disk] fail to locate the index files

So I have no idea what to do. I've tried multiple times with multiple different prefixes but it returns the same error. all of the files are in the same working directory.

[SIMONE]

in my code the reference is .fasta, not .fa, i don't this could be the issue

Answer 1

[Istvan Albert]

The proper invocation of the indexing is:

bwa index hg38.fa

that creates an index under the name of hg38.fa or

bwa index -p hg38 hg38.fa

which will create an index as hg38. Both would work fine.

In your case you are using the indexing command incorrectly, and the actually index name is hg38.fa

[ATpoint]

The indexing should still run with OPs code, but result in index files hg38.fa.* so bwa mem hg38.fa 4810_R1.fastq.gz 4810_R2.fastq.gz > 4810_aln.sam should do. Essentially, the trailing hg38 in OPs command line is simply ignored due to missing -p. Side-note @OP, don't store sam files, do at least bwa mem (...) | samtools view -o out.bam.