bwa sampe: paired ends with different names

Question

[mufernando]

I'm trying to align sequences I downloaded from the SRA with the dmel reference genome. After I align each paired ended fastq with the reference (with bwa aln), I try to run bwa sampe to get the .SAM.

This part of my shell script follows:

bwa aln -n 0.1 -t 8 dmel_ref.fasta "$i"\_"1.fastq" > "$i"\_"1.sai" 2>log/"$i"\_"1.sai.log"
bwa aln -n 0.1 -t 8 dmel_ref.fasta "$i"\_"2.fastq" > "$i"\_"2.sai" 2>log/"$i"\_"2.sai.log"
bwa sampe -a 800 -r '@RG\tID:foo\tSM:bar' dmel_ref.fasta "$i"\_"1.sai" "$i"\_"2.sai" "$i"\_"1.fastq" "$i"\_"2.fastq" > "$i.sam" 2>log/"$i.sam.log"

The tail of the .sam.log (STDERR from bwa sampe) is:

[infer_isize] inferred maximum insert size: 600 (6.52 sigma)
[bwa_sai2sam_pe_core] time elapses: 4.27 sec
[bwa_sai2sam_pe_core] changing coordinates of 818 alignments.
[bwa_sai2sam_pe_core] align unmapped mate...
[bwa_paired_sw] 832 out of 2416 Q17 singletons are mated.
[bwa_paired_sw] 114 out of 4353 Q17 discordant pairs are fixed.
[bwa_sai2sam_pe_core] time elapses: 0.55 sec
[bwa_sai2sam_pe_core] refine gapped alignments... 0.16 sec
[bwa_sai2sam_pe_core] print alignments...
[bwa_sai2sam_pe_core] paired reads have different names: "SRR1525696.81815875", "SRR1525696.81815878"

How can I fix this? I didn't do any previous trimming, I just used the downloaded .fastq from SRA.

Thank you!

Answer 1

[GenoMax]

Perhaps out of sync files were uploaded to SRA by error.

Run repair.sh from BBMap (find the syntax on the linked page) to see if you can fix the issue.

[mufernando]

I've tried that already, but I get the following error:

/usr/local/bin/repair.sh: line 71: /usr/local/bin//calcmem.sh: No such file or directory
/usr/local/bin/repair.sh: line 72: parseXmx: command not found
/usr/local/bin/repair.sh: line 76: freeRam: command not found
java -ea -Xmxm -cp /usr/local/bin/current/ jgi.SplitPairsAndSingles rp in=SRR1525696_1.fastq in2=SRR1525696_2.fastq out=SRR1525696_1_fixed.fastq out2=SRR1525696_2_fixed.fastq outs=singleton.fastq
Invalid maximum heap size: -Xmxm
Error: Could not create the Java Virtual Machine.
Error: A fatal exception has occurred. Program will exit.
Error: A fatal exception has occurred. Program will exit.

[GenoMax]

Are all BBmap suite scripts in one folder? It appears that you may have moved repair.sh out of the BBMap folder into /usr/local/bin.

You could manually run this command in the directory where the fastq files are present.

java -ea -Xmx2g -cp /usr/local/bin/current/ jgi.SplitPairsAndSingles rp in=SRR1525696_1.fastq in2=SRR1525696_2.fastq out=SRR1525696_1_fixed.fastq out2=SRR1525696_2_fixed.fastq outs=singleton.fastq

[mufernando]

oops! that was it. thanks!