Bowtie and reads that failed to align: (100.00%)

michy · 01-26-2011, 01:09 AM

Hello,
Can someone help me please. I have some new paired-end sequence from illumina and the reads are not aligning. What am I doing wrong? An example of the reads are below with the commandline I am using. I would be grateful for any help, cheers.

Read from sequence file #1
@HISEQ2000_0110:4:68:14837:199814#GGGCGG/1
GCTCCAGTATAGGGGAATGCCAGGACTAGGAAGAGGGATTGTGTGGTTTGGAGAGCAGGACGGAGGGAGGGTATACGTCACTATGAGATGGGAAACTTGTA
+HISEQ2000_0110:4:68:14837:199814#GGGCGG/1
ccccc_ca\c[[]X]JYJVRRHSXHMMNYVPVZEIWFFYXXZ_Z]cbccb_J[G_VPZUF^X`H`^YD`BBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBB
@HISEQ2000_0110:4:68:14861:199831#GCACTG/1

Read from sequence file #2
@HISEQ2000_0110:4:68:14837:199814#GGGCGG/2
CACTTTAAAATCTATTTGATCTTGAGGAAATCAGTTGTGTTTCCTAGTTATATAGTCTATCATTTAATAATAGCACTAATGAAGTGTTTAGAAGTAATAAT
+HISEQ2000_0110:4:68:14837:199814#GGGCGG/2
ggggggggggfbfIffffefggggeggggeU^cf_eedfee^ZZ\accdd]eeeeffedfI[bbbcYccbeaeddgggaedTcc_dbfdIc\c\P_ZM_BB

Commandline
./bowtie -t -p 14 -q -I 0 -X 300 -S --chunkmbs 2048 mus -1 ../working/s_temp_1.txt -2 ../working/s_temp_2.txt ../working/s_temp_aligned.sam

fkrueger · 01-26-2011, 01:43 AM

Hi michy,

The command you used looks ok-ish, but

(a) -q is the default, however Bowtie will then assume Phred33 qualities. The HiSeq seems to chuck out Phred64 encoded data though, try specifiying -q --phred64-quals
(b) -I 0 is not not required as it is the default
(c) If you still don't get any alignments you might want to increase -X to a higher value, maybe your insert sizes are just larger than 300bp?

Hope (a) or (c) will fix it.

michy · 01-26-2011, 02:12 AM

Hello fkrueger,
thank you for your reply. Your suggestions esp. a) and c) they did improve the alignment by 15% which means 85% still failed to align. However, I was testing the command on a small subset of the data. I will try some more reads and get back to you.

Thank you again and any other suggestions welcome.

Michy

fkrueger · 01-26-2011, 02:21 AM

May I ask if you used FastQC on your sequence files to find out whether you are sequencing into the adapter on the other side? This might throw mapping efficiencies off quite a bit and can be fixed by trimming adapter sequences off or just shortening the sequences a bit.

michy · 01-26-2011, 02:59 AM

Hello fkrueger,
I've never tried this software but I will now, it looks very useful.

Thank you,
Michy

andrehorta · 02-07-2011, 05:11 AM

I used the command: bowtie -p 4 /rs4244385A

And reads that failed to align: (100.00%)

Can help me?

andrehorta · 02-08-2011, 05:02 AM

Quote:

Originally Posted by andrehorta

I used the command: bowtie -p 4 /rs4244385A

And reads that failed to align: (100.00%)

Can help me?

anyone?????

swbarnes2 · 02-08-2011, 07:42 PM

Your first read has lousy quality scores, and its best blast to mouse has several discrepancies. That's why it won't align. The second read should align fine.

rs4244385 is a human SNP.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Bowtie Paired end 100% failed to align	rahul.m.dhodapkar	Bioinformatics	15	09-12-2013 03:05 PM
Hello and a Question: 50 or 100 bp reads?	kerhard	Introductions	0	02-11-2011 02:13 PM
Bowtie to align reads to single chromosome or region?	jjw14	Bioinformatics	1	10-07-2010 07:20 PM
TopHat+BWT = reads that failed to align ~ 99%	repinementer	Bioinformatics	6	09-28-2010 08:31 PM
Tophat...Mapping reads against Reference with Bowtie [FAILED]	Brajbio	Bioinformatics	0	06-02-2010 01:33 AM

01-26-2011, 01:09 AM	#1
michy Junior Member Location: oxford Join Date: Aug 2008 Posts: 5	Bowtie and reads that failed to align: (100.00%) Hello, Can someone help me please. I have some new paired-end sequence from illumina and the reads are not aligning. What am I doing wrong? An example of the reads are below with the commandline I am using. I would be grateful for any help, cheers. Read from sequence file #1 @HISEQ2000_0110:4:68:14837:199814#GGGCGG/1 GCTCCAGTATAGGGGAATGCCAGGACTAGGAAGAGGGATTGTGTGGTTTGGAGAGCAGGACGGAGGGAGGGTATACGTCACTATGAGATGGGAAACTTGTA +HISEQ2000_0110:4:68:14837:199814#GGGCGG/1 ccccc_ca\c[[]X]JYJVRRHSXHMMNYVPVZEIWFFYXXZ_Z]cbccb_J[G_VPZUF^X`H`^YD`BBBBBBBBBBBBBBBBBBBBBBBBBBBBBBBB @HISEQ2000_0110:4:68:14861:199831#GCACTG/1 Read from sequence file #2 @HISEQ2000_0110:4:68:14837:199814#GGGCGG/2 CACTTTAAAATCTATTTGATCTTGAGGAAATCAGTTGTGTTTCCTAGTTATATAGTCTATCATTTAATAATAGCACTAATGAAGTGTTTAGAAGTAATAAT +HISEQ2000_0110:4:68:14837:199814#GGGCGG/2 ggggggggggfbfIffffefggggeggggeU^cf_eedfee^ZZ\accdd]eeeeffedfI[bbbcYccbeaeddgggaedTcc_dbfdIc\c\P_ZM_BB Commandline ./bowtie -t -p 14 -q -I 0 -X 300 -S --chunkmbs 2048 mus -1 ../working/s_temp_1.txt -2 ../working/s_temp_2.txt ../working/s_temp_aligned.sam

01-26-2011, 01:43 AM	#2
fkrueger Senior Member Location: Cambridge, UK Join Date: Sep 2009 Posts: 625	Hi michy, The command you used looks ok-ish, but (a) -q is the default, however Bowtie will then assume Phred33 qualities. The HiSeq seems to chuck out Phred64 encoded data though, try specifiying -q --phred64-quals (b) -I 0 is not not required as it is the default (c) If you still don't get any alignments you might want to increase -X to a higher value, maybe your insert sizes are just larger than 300bp? Hope (a) or (c) will fix it.

01-26-2011, 02:12 AM	#3
michy Junior Member Location: oxford Join Date: Aug 2008 Posts: 5	Hello fkrueger, thank you for your reply. Your suggestions esp. a) and c) they did improve the alignment by 15% which means 85% still failed to align. However, I was testing the command on a small subset of the data. I will try some more reads and get back to you. Thank you again and any other suggestions welcome. Michy

01-26-2011, 02:21 AM	#4
fkrueger Senior Member Location: Cambridge, UK Join Date: Sep 2009 Posts: 625	May I ask if you used FastQC on your sequence files to find out whether you are sequencing into the adapter on the other side? This might throw mapping efficiencies off quite a bit and can be fixed by trimming adapter sequences off or just shortening the sequences a bit.

01-26-2011, 02:59 AM	#5
michy Junior Member Location: oxford Join Date: Aug 2008 Posts: 5	Hello fkrueger, I've never tried this software but I will now, it looks very useful. Thank you, Michy

02-07-2011, 05:11 AM	#6
andrehorta Member Location: Brazil - Belo Horizonte - UFMG Join Date: Jan 2011 Posts: 14	I used the command: bowtie -p 4 /rs4244385A And reads that failed to align: (100.00%) Can help me?

02-08-2011, 07:42 PM	#8
swbarnes2 Senior Member Location: San Diego Join Date: May 2008 Posts: 912	Your first read has lousy quality scores, and its best blast to mouse has several discrepancies. That's why it won't align. The second read should align fine. rs4244385 is a human SNP.