Bwa producing no alignment

Question

[ThePresident]

I am dealing with Illumina paired-end, 150 bp read-length datasets.

Here's what my fastq files looklike:

@SRR3405394.1 1 length=151
NCGACTGAGGTAATTACAGTTCTTCGGTTCCAGCCAGCTGTCTCAGTTTATGGACCAGAACAACCCGCTGTCTGAGATTACGCACAAACGTCGTATCTCCGCACTCGGCCCAGGCGGTCTGACCCGTGAACGTGCAGGGTACAGATCGGAA
+SRR3405394.1 1 length=151
#<GGGGAGGIGGGIGGIIIIIGIIIIIIIIIIGIIIIIIIGGIIGIGGIIIIIGIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIGIGGIGGIIGGIIIIIGGGIIIIIGGIIIIGIIIGIGII<GGGGAGIGGAGG

I use this command to align with bwa:

bwa aln -t8 index 1.fastq > 1.sai
bwa aln -t8 index 2.fastq > 2.sai
bwa sampe index 1.sai 2.sai 1.fastq 2.fastq > aln_bwa.sam

When I look at my SAM file, I see this:

@SQ SN:selected LN:4641665
@PG ID:bwa  PN:bwa  VN:0.7.13-r1126 CL:bwa sampe index 1.sai 2.sai 1.fastq 2.fastq
SRR3405394.1    77  *   0   0   *   *   0   0   NCGACTGAGGTAATTACAGTTCTTCGGTTCCAGCCAGCTGTCTCAGTTTATGGACCAGAACAACCCGCTGTCTGAGATTACGCACAAACGTCGTATCTCCGCACTCGGCCCAGGCGGTCTGACCCGTGAACGTGCAGGGTACAGATCGGAA #<GGGGAGGIGGGIGGIIIIIGIIIIIIIIIIGIIIIIIIGGIIGIGGIIIIIGIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIGIGGIGGIIGGIIIIIGGGIIIIIGGIIIIGIIIGIGII<GGGGAGIGGAGG
SRR3405394.1    141 *   0   0   *   *   0   0   GTACCCTGCACGTTCACGGGTCAGACCGCCTGGGCCGAGTGCGGAGATACGACGTTTGTGCGTAATCTCAGACAGCGGGTTGTTCTGGTCCATAAACTGAGACAGCTGGCTGGAACCGAAGAACTGTAATTACCTCAGTCGTAGATCGGAA GGAAGIIGIIIGIGIIGGGIIGGGGIGGIIAGGGIIGGGGGIGIGGGGIIGGGIGGGGGIIIIIGGIIIIIIIIIIIIGIIIGIIIIGIIIGGIIIGIIIIAGGIIIGGAGIGGGGGGI<GGGGIIIIIGGGIGIIIGIIIIIAGGIIIIG

All my reads are flagged 77 / 141, i.e. they're not aligned. What's the problem? I have a feeling that the quality scores are not recognized, but that's just a wild guess based on some quick reading.

Thank you in advance, TP

[WouterDeCoster]

Do you have a reason to use bwa aln rather than bwa mem?

[ThePresident]

I am interested in extracting aberrant pairs (same orientation fwd-fwd and bkw-bkw combinations). I know that bwa aln is able to mark them properly but I am not sure about bwa mem. Do you happen to know if bwa mem handles these type of pairs the same way as bwa aln does?

[Fabio Marroni]

Just out of curiosity: why are you running aln and index at the same time? For what I remember aln is used to align and should create a sai output, while index is used to create an index of a fasta reference file.

[WouterDeCoster]

I guess index is a placeholder for the path to his index file?

[ThePresident]

Correct, index is placeholder for my index file.

[Fabio Marroni]

Is it possible that you have quality in Illumina 1.3+ or 1.5+ while (I think) bwa is now expecting Illumina 1.8+? Try giving a look at these posts:

What Is The Default Quality Encoding Expected By Bwa?

Convert Illumina 1.3 To Illumina 1.5

https://en.wikipedia.org/wiki/FASTQ_format

[ThePresident]

I did this awhile ago but I believe I verified that I have 1.8+ Illumina score. If I remember correctly FastQC can detect what scores are used.

[Fabio Marroni]

mmm, yes, I noticed you have # in the quality, which cannot be Illumina 1.3+ nor 1.5+.

1) Why do you think it is something in the quality? Did you receive some warning?

2) Have you tried building a small reference identical to one of your reads and aligning against it?

[ThePresident]

1) Googling here and there led me to think that quality scores might be a problem but now I know it shouldn't be. 2) Nope, but that's a good idea. I'll try it.