bwa paired read FR issue

[starbug]

Hi,

I have been trying to run the following bwa mem command:

bwa mem -M -t 16 w_ind.fa SRR412532_1.fastq.gz SRR412532_2.fastq.gz | samtools view -Sb - | samtools sort - SRR412532.sorted && samtools index SRR412532.sorted.bam

And for each iteration get the following:

[M:rocess] read 1777778 sequences (160000020 bp)...
[M::mem_pestat] # candidate unique pairs for (FF, FR, RF, RR): (0, 2, 0, 0)
[M::mem_pestat] skip orientation FF as there are not enough pairs
[M::mem_pestat] skip orientation FR as there are not enough pairs
[M::mem_pestat] skip orientation RF as there are not enough pairs
[M::mem_pestat] skip orientation RR as there are not enough pairs
[M::mem_process_seqs] Processed 1777778 reads in 75.678 CPU sec, 22.324 real sec

This shoud be a paired read dataset and when I check the FASTQ files I see:

gzcat SRR412532_1.fastq.gz | head
@SRR412532.1 FCC02L0ACXX:3:1101:1130:1956/1
NGAGCTTCAGGCCCAGGCCAAGGCTTACTTTGAGAAGACGCAGGAGCAGCTAACACCCCTGGTCAAGAAGGCCGGAACTGACCTGATCAA
+
#1=DDDDEHHHGGIGHI=FHGGGIG>BAFGHC>=BCGG>:?DG?CB;FECE>CC@G1=5,,;?;@ACA;>;2=;'820>@>>>C<>@@A>
@SRR412532.2 FCC02L0ACXX:3:1101:1035:1967/1
NACAAAAAGCAAAATGAATCTAGCTGTCCCTGTCCTGGCCGGTATTCCATCTTCTAGAACCTGTTTCCGTGTTTTTCCTGGAGTGTCTGC
+
#1:BDFFFF?FHHIGEHIB@AHIGICFDEGGIHGEEHIIGEG?FFEDHIGGGIIIG;CGGGEGGHGA3;.;.?B@=@AC>>55(5>5;A#
@SRR412532.3 FCC02L0ACXX:3:1101:1464:1954/1
NTGCTTTTCTGCCCTGGAAGTTGATGAGGCATATGTTCCCAAAGAGTTTAACGCTGAAACATTCACCTTCCATGCAGACTTATGCACACT

gzcat SRR412532_2.fastq.gz | head
@SRR412532.1 FCC02L0ACXX:3:1101:1130:1956/2
GTTGGGAGAGGGCTCCAGACCTGGCCAGTGGGGGTTCTAGGGGCCAGCAGGGGAGGGAAGACAATGGTCTGGACGCCTCACTGGGTGGCA
+
@B@?DDDDAHFHHIFHIJIGIGCHHIGI3?FGII7;FE@GHIBHAEB?@E;>>8?@2'588ACCCDC(44?9@AB><B9(8@A#######
@SRR412532.2 FCC02L0ACXX:3:1101:1035:1967/2
GGGTTTAGGGACCTCCCTGGGTGGAACAATCTACACGGTTGAAGCACTAGCAGGGAGGTTCTAGTGGGATCACACGTTTATTAACATGCT
+
?@@=BDFFGHDFFGGHIGIHCFEHHGGHGHGGGGGHIDBFHIJCFHCAGIFGHGIHEB?;BBC>;AC6A>?=CDCB?A<BDDCDDCA@CC
@SRR412532.3 FCC02L0ACXX:3:1101:1464:1954/2
GGCTTGTGTTTCACCAGCTCAGCAAGTGCAGATTGTTTCTTGACTTGTTTCTCAGCCTCAGGAAGTGTGCATAAGACTGCATGGAAGGTG

I am not sure why there are not enough FR pairs as the reads seem to be in the same order with matching headings... Am probably missing something obvious!!! Any ideas??

All help much appreciated!!

Thanks in advance

[GenoMax]

Are you using current versions of bwa/samtools?

[starbug]

bwa version 0.7.12 / samtools 1.2

Have tried a couple of other SRA datasets with this bwa install and they map really nicely. Bit stumped as to what's different about this one...

[GenoMax]

Have you done a quick count on the two files to make sure that there is an equal number of reads in them? (grep for "^@SRR412532" and do a line count on both files).

[starbug]

Ooh no I hadn't thought of that! Just given that a try now but both showing same count (26,073,197)

[GenoMax]

But it appears that bwa is only reading 1,777,778 sequences based on the log above so something is wrong with the files. Perhaps you can try to download them again?

You can find the fastq files at EBI SRA: http://www.ebi.ac.uk/ena/data/view/SRR412532 if you don't want to deal with SRA.

[starbug]

Have just downloaded the FASTQs from the EBI again (sorry, I should have mentioned I bypassed SRA) and still the same issue. Is it worth downloading them as SRAs with fastq-dump even though it will take longer?

[GenoMax]

Since files at EBI have this problem it may be worth a shot getting them using fastq-dump from SRA.

There are instances where the original data as submitted may have a problem. If the files from SRA don't work either then you should contact SRA support and/or the original submitter to let them know.