Really Low Bowtie2 Alignment Numbers From Demultiplex in STACKS

[scottr08]

In STACKS, I ran process_radtags and get close to 20 million reads. They are paired-end 150bp reads.

/home/srile14/stacks-1.48/process_radtags -p /work/srile14/FastqFiles_MS2_044_Kelly_OysterRADseq/ \
--paired \
-i gzfastq \
-b /work/srile14/demulti --inline_inline \
-o /work/srile14/test-out \
-c -q -r -t 140 -w 0.15 -s 10 \
--renz_1 xbaI \
--renz_2 ecoRI \
--adapter_mm 2 \
--adapter_1 AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC \
--adapter_2 AGATCGGAAGAGCGTCGTGTAGGGAAAGAGTGT \


20272558 total sequences
142938 reads contained adapter sequence (0.7%)
356902 ambiguous barcode drops (1.8%)
0 low quality read drops (0.0%)
19436 ambiguous RAD-Tag drops (0.1%)
19753282 retained reads (97.4%)



When I go to align those reads to the genome I created with bowtie2-build, I get only 14,641 reads.

bowtie2 -q -x /work/srile14/virginica_genome/virginica_genome \
-1 Sample_96.1.fq,Sample_95.1.fq,Sample_94.1.fq,....Sample_1.1.fq
-2 Sample_96.2.fq,Sample_95.2.fq,Sample_94.2.fq,....Sample_1.2.fq
-S /work/srile14/stdout



14641 reads; of these:
14641 (100.00%) were paired; of these:
4171 (28.49%) aligned concordantly 0 times
5341 (36.48%) aligned concordantly exactly 1 time
5129 (35.03%) aligned concordantly >1 times
----
4171 pairs aligned concordantly 0 times; of these:
88 (2.11%) aligned discordantly 1 time
----
4083 pairs aligned 0 times concordantly or discordantly; of these:
8166 mates make up the pairs; of these:
5678 (69.53%) aligned 0 times
1199 (14.68%) aligned exactly 1 time
1289 (15.78%) aligned >1 times
80.61% overall alignment rate


The overall alignment rate is high, but the total number of reads mapped seems extremely low (14,651 out of 20 million). Is there something I am missing? Or is this common for longer reads?

Thanks in advance,

Scott

[mastal]

A very low alignment rate could be a sign that the reads are not from the species you expected.

[SNPsaurus]

Can you make a de novo reference in Stacks and blast the loci sequences to see what they are? As mastal says, it could be that your samples are heavily contaminated with some other species (like bacteria). Or just grab 10 reads from a few samples and blast them. It is always good to go to the starting material and check it out directly.

If your samples are not the exact species as the reference, then it may be that you need to allow more mismatches during the alignment.