[ymc]

Is this the right way to use bowtie to align reads from an 1000 genome sample (e.g. NA18526) to the human genome?

First, I downloaded three files from 1000genomes
ftp://ftp.1000genomes.ebi.ac.uk/vol1....filt.fastq.gz
ftp://ftp.1000genomes.ebi.ac.uk/vol1....filt.fastq.gz
ftp://ftp.1000genomes.ebi.ac.uk/vol1....filt.fastq.gz

I downloaded the hg19 index files from
ftp://ftp.cbcb.umd.edu/pub/data/bowt...g19.ebwt.1.zip
ftp://ftp.cbcb.umd.edu/pub/data/bowt...g19.ebwt.2.zip

Unzipped them and put them in the indexes subdirectory.

Then I run the following command on my dual hex-core machine with 48GB RAM:

time ./bowtie -p 12 --chunkmbs 200 -S hg19 -1 NA18526/ERR031854_1.filt.fastq -2 NA18526/ERR031854_2.filt.fastq NA18526.sam

It took about 65min to finish. Am I doing it right?

What is the use of the small ERR031854.filt.fastq at 1000genomes.org??

Thanks a lot in advance!

[gntc]

Duplicate outputs

I have some reads that were reported to map to hg19 uniquely (only one place in the genome) but they have two lines of output. The two lines are completely identical. This happened for many reads. Has anyone else had this problem?

[amaer]

the M option in Bowtie2 beta1.3

Hi,

I have a question about the use of the M parameter in Bowtie2 beta1.3. I understand it's the number of valid, distinct alignments found for each read (M+1 actually). And reports only 1, the best alignment of THOSE examined.

If a read maps to multiple places, in human genome for example, I expect that Bowtie2 to possibly have different best alignment reported for different M values.

Is it possible to have a 36 bp read NOT mapped at all (to human genome) with M =1, but mapped in other places with a higher values for M? If yes, why?

Thanks!

[Xi Wang]

It seems that it has nothing to do with M if a read is not mappable. Because in this mode, "The search terminates when it can't find more distinct valid alignments, or when it finds M+1 distinct alignments, whichever happens first." As the read is unmappable, the searching will always ends up with the former condition, no matter whatever M is.

[amaer]

But the read IS mappable (1 mismatch only for 36 bp)! And I can see that when I increase M. I mean, with M 1, it did not map anywhere on the human genome, but it did so when I increased M (keeping all other parameters constant - N, L, i etc. And when I use -a (all) alignments option I get tens of alignments. I find that very strange.
I expect that regardless of the M value, at least one of the alignments would be reported, not necessarily the best one.

[Xi Wang]

Seems its a bug of Bowtie2. Probably you may report it to the developers.

[LizBent]

Hi, Bowtie noob question here. I have some paired read data where there are unpaired reads (that is, these reads were removed in preprocessing because they were too short, so that there are sequences in read1.fastq that are not present in read2.fastq, and vice versa). Can I use this with Bowtie, or should I go back to the raw data and re-process it (to remove primers and adapters) this time without removing any sequences? The data was originally prepared for use with Trinity, which supports unpaired reads in paired end data.

[rsinha]

Bowtie2: How to reproduce example alignment

I am running Bowtie2 to reproduce the example mentioned in the Manual but all in vain. Here is the example I am trying to reproduce:
End-to-end alignment example

Read: GACTGGGCGATCTCGACTTCG
Reference: GACTGCGATCTCGACATCG

Alignment:
Read: GACTGGGCGATCTCGACTTCG
|| || | || || || |||| || ||
Reference: GACTG--CGATCTCGACATCG

I will be glad if someone let me know what parameters were used to find above alignment. OR someone who is able to do it

[Dario1984]

Any plans to implement --best and --strata for paired-end mode ? I get a lot less reads mapping with -v 3 -m 1 because it isn't present.

[Tobias Kockmann]

-m parameter in "end-to-end" mode

Dear Bowtie developers/users,

I have seen the following line of code in a Nature protocol for aligning 36 bp quality-pre-filtered ChIP-seq reads to the Drosophila genome. I was asking myself, if they are really sensible:

bowtie -q -m 1 -v 3 --best --strata

As I under stood from the Bowtie publication, -v mode will create "end-to-end" hits, but totally ignores quality information. The alignment ranking is therefor purely based on the number of mismatches (stratum). Since the max. number of mismatches is set to 3, valid alignments will be allowed to contain up to 3 mismatched position over the complete 36mer. The reporting option -m 1 means that Bowtie will only report an alignment, if the read is unique.

Lets assumes the following scenarios for a single read:

a.)
alignment a, perfect match
alignment b, 1 mismatched position

Bowtie will report a, right?

b.)
alignment a, perfect match
alignment b, perfect match

Bowtie will report no match for the read, since both valid alignments have the same rank.

But, what do the options --best and --strata trigger in "end-to-end" mode??? Do they do anything?

Greetings,
Tobi

[M4love]

Hey I had a similar question. I have dowloaded bowtie latest version, but am new to use it. infact i am dumb enough not to know how to open it or use it. I extracted the file and am trying to open it in "terminal" and even R non of it is working.

[LizBent]

Hey I had a similar question. I have dowloaded bowtie latest version, but am new to use it. infact i am dumb enough not to know how to open it or use it. I extracted the file and am trying to open it in "terminal" and even R non of it is working.

Did you follow the installation instructions? If you did, and were successful, you can type

bowtie

into the prompt and it will give you the help screen. If you were not, you will get an error. If you need help installing bowtie, find someone locally to help you. There are instructions on the website for how to install bowtie.

[M4love]

Did you follow the installation instructions? If you did, and were successful, you can type

bowtie

into the prompt and it will give you the help screen. If you were not, you will get an error. If you need help installing bowtie, find someone locally to help you. There are instructions on the website for how to install bowtie.

Hey where should i type Bowtie
in the R software or in the terminal?

[M4love]

Thanks a lot though. I really feel stupid

[M4love]

Hey where should i type Bowtie
in the R software or in the terminal?

and I followed these instructions

http://javastat.stat.wvu.edu/tcltk/Bowtie.html