Mapping a resequenced genome

[carlos]

Hi,

I would like to use bowtie2 to map a low coverage genome sequence over a reference genome. Libraries are Illumina ( 170 pb ) and I used latests bowtie2 (bowtie2-2.2.3). Using this options:
bowtie2 --phred64 -x ref.scaffolds -p 38 -I 0 -X 370 -1 forward.fq -2 reverse.fq -S output.sam

either with local or end-to-end my reads "aligned concordantly exactly 1 time" are low, any idea of how I can improve my options?

Bests,

Carlos

with end-to-end stats are:

207650332 reads; of these:
207650332 (100.00%) were paired; of these:
49664549 (23.92%) aligned concordantly 0 times
55571363 (26.76%) aligned concordantly exactly 1 time
102414420 (49.32%) aligned concordantly >1 times
----
49664549 pairs aligned concordantly 0 times; of these:
555481 (1.12%) aligned discordantly 1 time
----
49109068 pairs aligned 0 times concordantly or discordantly; of these:
98218136 mates make up the pairs; of these:
74270285 (75.62%) aligned 0 times
8235025 (8.38%) aligned exactly 1 time
15712826 (16.00%) aligned >1 times
82.12% overall alignment rate

and with --local alignment

207650332 reads; of these:
207650332 (100.00%) were paired; of these:
22728112 (10.95%) aligned concordantly 0 times
41051553 (19.77%) aligned concordantly exactly 1 time
143870667 (69.29%) aligned concordantly >1 times
----
22728112 pairs aligned concordantly 0 times; of these:
324840 (1.43%) aligned discordantly 1 time
----
22403272 pairs aligned 0 times concordantly or discordantly; of these:
44806544 mates make up the pairs; of these:
29720852 (66.33%) aligned 0 times
2927354 (6.53%) aligned exactly 1 time
12158338 (27.14%) aligned >1 times
92.84% overall alignment rate

[GenoMax]

Is this data really old (that you have selected phred64 scores)? How long are the reads?

[carlos]

yes, a bit old, Illumina paired end (80 bp after trimming)

[Brian Bushnell]

Reads aligning multiple times is a function of read length, reference repeat content, and possibly library bias; there's not much you can do about that once you have existing data. However, it's also affected by pairing, and if I am reading your stats correctly, your reads are not being paired correctly. Are you sure the two files go together, and that the order of the reads is the same? If you trimmed the two files separately, that will have broken the pairing order. So specifically, what tool did you use for trimming, and what was the command?