Difference in output VCF between running on a local PC and supercomputer cluster

Question

[zhou_1228]

I ran a SNP calling program on both local PC and supercomputer cluster. However, the output VCF file generated by supercomputer lack the positive mutation, which we can identify from VCF file produced by local PC. I wonder whether we need adjust some parameters or add some options in our commands when using it on supercomputer, then the mutation will shown. I would truly appreciate your answers. Below is the shell scripts of program I used,

1.first indexing the genome

bwa index -p GlycinMaxbwaidx -a bwtsw GlycinMax.fasta

2.algn

bwa aln -t 4  GlycinMaxbwaidx P1-WD04-3L-F.fastq > P1-WD04-F.bwa
bwa aln -t 4  GlycinMaxbwaidx P1-WD04-3L-R.fastq > P1-WD04-R.bwa

3.build sam file

bwa sampe GlycinMaxbwaidx P1-WD04-F.bwa P1-WD04-R.bwa P1-WD04-3L-F.fastq P1-WD04-3L-R.fastq > P1-WD04.sam

4.convert sam to bam

samtools view -S -b P1-WD04.sam > P1-WD04.bam

5.sort and index bam

samtools sort  P1-WD04.bam -o  P1-WD04.sorted.bam
samtools index P1-WD04.sorted.bam

6.varriant calling

freebayes -f GlycinMax.fasta -p 24 --use-best-n-alleles 4 --pooled-discrete P1-WD04.sorted.bam >P1-WD04.vcf

[Pierre Lindenbaum]

are you using the version same versions of the tools ?

[zhou_1228]

Hi Pierre, thanks for your reply. Yes, I checked the versions of all three tools, BWA, samtools, and freebayes in my SNP calling program, and they are the same between local PC and supercomputer cluster. However, I noticed that the size of final VCF file are different, 19.9MB on supercomputer and 23.2MB on local PC.

[Pierre Lindenbaum]

you can pipe bwa into samtools sort ...

[h.mon]

Are you reads shorter than 70bp? If not, bwa mem is preferred over bwa aln.

Answer 1

[swbarnes2]

For starters, is the vcf the first intermediate file where the supercomputer results differ from the local PC?

And as Pierre says:

Software version differences are the likely answer, and bwa samse | samtools sort works fine, because samtools sort now works on sam files