Building indexes with bowtie-build

bre · 12-31-2009, 10:37 AM

Hello, I am currently trying to align two closely related genomes from multiple sequence fasta files. I have a target genome which is made of 1168 sequences in a single fasta file and I wish to use this to build indexes. Initially I did this with the command "bowtie-build target_file.fa t_index". This ran in a reasonable amount of time and produced 6 index files: t_index.1.ebwt, t_index.2.ebwt, t_index.3.ebwt, t_index.4.ebwt, t_index.rev.1.ebwt, t_index.rev.2.ebwt.
I then ran the command "bowtie t_index -f -t -p4 -l 13 ../PATH/TO/QUERY output.map" to run an alignment, where QUERY is also a multiple sequence fasta file where each sequence length is <= 1024 bases. My output was not what I was expecting; I got only 5.03% alignment when I was expecting over 90%. This lead my to wonder whether or not indexes had been created correctly. Is bowtie-build able to take a multisequence fasta file? Maybe each index file was being overwritten by each following sequence in the target file? I tweaked some parameters and was still not able to get over 6% alignment.

bre · 12-31-2009, 12:56 PM

I tested the indexes by running them against the sequences they were created from and only got 67% alignment so this leads me to further doubt the indexes obtained. Has anyone else had similar problems?

Xi Wang · 01-01-2010, 10:56 PM

I have successfully built the index from multiple sequences in a fasta file and the mapping results seem all right. I guess the problem you encountered is because of the length of the sequence you queried. Such kind of NGS mapping tools is mainly designed for short read alignment.

ZZQ · 03-31-2010, 11:16 PM

I've heard of bowtie, but have not a try yet.

sperry · 04-01-2010, 09:38 AM

For the read lengths you are working with, you are probably better off using a different alignment tool. BWA has an alignment option that specifically accommodates longer read lengths (bwasw). http://bio-bwa.sourceforge.net/

Quote:

Originally Posted by bre

I tested the indexes by running them against the sequences they were created from and only got 67% alignment so this leads me to further doubt the indexes obtained. Has anyone else had similar problems?

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
TopHat Error: Could not find Bowtie index files /bowtie-0.12.5/indexes/.	rebrendi	Bioinformatics	11	06-22-2016 10:55 AM
bowtie index problem (bowtie-build and then bowtie-inspect)	tgenahmet	Bioinformatics	4	09-10-2013 12:51 PM
BOWTIE - Build Indexes - tutorial	andrehorta	Bioinformatics	0	02-07-2011 04:52 AM
bowtie indexes for Soap2?	maojn7488	Bioinformatics	2	07-29-2010 07:32 AM
unzipping human bowtie indexes	Layla	Bioinformatics	8	11-12-2009 03:16 PM

12-31-2009, 10:37 AM	#1
bre Member Location: United States Join Date: Dec 2009 Posts: 11	Building indexes with bowtie-build Hello, I am currently trying to align two closely related genomes from multiple sequence fasta files. I have a target genome which is made of 1168 sequences in a single fasta file and I wish to use this to build indexes. Initially I did this with the command "bowtie-build target_file.fa t_index". This ran in a reasonable amount of time and produced 6 index files: t_index.1.ebwt, t_index.2.ebwt, t_index.3.ebwt, t_index.4.ebwt, t_index.rev.1.ebwt, t_index.rev.2.ebwt. I then ran the command "bowtie t_index -f -t -p4 -l 13 ../PATH/TO/QUERY output.map" to run an alignment, where QUERY is also a multiple sequence fasta file where each sequence length is <= 1024 bases. My output was not what I was expecting; I got only 5.03% alignment when I was expecting over 90%. This lead my to wonder whether or not indexes had been created correctly. Is bowtie-build able to take a multisequence fasta file? Maybe each index file was being overwritten by each following sequence in the target file? I tweaked some parameters and was still not able to get over 6% alignment.

12-31-2009, 12:56 PM	#2
bre Member Location: United States Join Date: Dec 2009 Posts: 11	Self alignment I tested the indexes by running them against the sequences they were created from and only got 67% alignment so this leads me to further doubt the indexes obtained. Has anyone else had similar problems?

01-01-2010, 10:56 PM	#3
Xi Wang Senior Member Location: MDC, Berlin, Germany Join Date: Oct 2009 Posts: 317	I have successfully built the index from multiple sequences in a fasta file and the mapping results seem all right. I guess the problem you encountered is because of the length of the sequence you queried. Such kind of NGS mapping tools is mainly designed for short read alignment. __________________ Xi Wang

03-31-2010, 11:16 PM	#4
ZZQ Junior Member Location: beijing Join Date: Mar 2010 Posts: 3	I've heard of bowtie, but have not a try yet.