Bowtie and Color Space

[cutcopy11]

Hey, I tried the latest release of Bowtie, which now supports color space.

I downloaded the dataset from the link below and got the results further below after slightly less than 4 hours.

(I study sequence patterns in Nucleosomal DNA.)

http://www.ncbi.nlm.nih.gov/sra/?term=SRA001023

ftp://ftp.ncbi.nlm.nih.gov/sra/static/SRX000/SRX000425/
(both files)

/Desktop/bowtie-0.12.0-beta1$ ./bowtie -C -q -a -m 1 c_elegans_ws200_c c_elegans_425.fastq c_elegans_425_positions.txt | awk -v OFS='\t' '{if($2 == "-") {$4 += (length($5)-1)} ; print $0}'

# reads processed: 107422570
# reads with at least one reported alignment: 35370418 (32.93%)
# reads that failed to align: 66794981 (62.18%)
# reads with alignments suppressed due to -m: 5257171 (4.89%)
Reported 35370418 alignments to 1 output stream(s)

I think the quality of reads using Solid is not as good as Solexa (i.e. more mismatches). I used the default 2 mismatch threshold.

I'm going to try to compare with BFAST someday once I figure out how to make proper masks for C. elegans. Bowtie is nice for those who have no idea how these programs work and are not computer science majors. The manual is very easy to understand and one can simply download the indices for commonly used genomes from the website.

-Clayton

[nilshomer]

Hey, I tried the latest release of Bowtie, which now supports color space.

I downloaded the dataset from the link below and got the results further below after slightly less than 4 hours.

(I study sequence patterns in Nucleosomal DNA.)

http://www.ncbi.nlm.nih.gov/sra/?term=SRA001023

ftp://ftp.ncbi.nlm.nih.gov/sra/static/SRX000/SRX000425/
(both files)

/Desktop/bowtie-0.12.0-beta1$ ./bowtie -C -q -a -m 1 c_elegans_ws200_c c_elegans_425.fastq c_elegans_425_positions.txt | awk -v OFS='\t' '{if($2 == "-") {$4 += (length($5)-1)} ; print $0}'

# reads processed: 107422570
# reads with at least one reported alignment: 35370418 (32.93%)
# reads that failed to align: 66794981 (62.18%)
# reads with alignments suppressed due to -m: 5257171 (4.89%)
Reported 35370418 alignments to 1 output stream(s)

I think the quality of reads using Solid is not as good as Solexa (i.e. more mismatches). I used the default 2 mismatch threshold.

I'm going to try to compare with BFAST someday once I figure out how to make proper masks for C. elegans. Bowtie is nice for those who have no idea how these programs work and are not computer science majors. The manual is very easy to understand and one can simply download the indices for commonly used genomes from the website.

-Clayton

I would set the # of mismatches to 10% of the read length since SOLiD error can be up to 10% and a SNP will eat up two mismatches in color space.

[xuying]

Can you tell me how to convert csfasta and qual files to csfastq? Thanks!

[snetmcom]

exactly. you need to increase mismatch tolerance levels. it takes 2 mismatches for a single SNP, 1 mismatch is system error. Most people run 6 mismatches for a 50bp tag if you have the processing power.

[pmiguel]

Can you tell me how to convert csfasta and qual files to csfastq? Thanks!

Hi Xuying,
You are much more likely to get an answer to a question if you start a new thread. Especially if the current thread is unrelated to your question.

--
Phillip

[spenthil]

Can you tell me how to convert csfasta and qual files to csfastq? Thanks!

I can't, but google can: http://www.google.com/search?rlz=1C1...les+to+csfastq

[nilshomer]

I can't, but google can: http://www.google.com/search?rlz=1C1...les+to+csfastq

If it is going to be that way, then try

http://lmgtfy.com/?q=convert+csfasta+and+qual+files+to+csfastq

[Mr Mutundes]

paired reads

has anyone had success using bowtie with colorspace mate pair inputs?

[June]

exactly. you need to increase mismatch tolerance levels. it takes 2 mismatches for a single SNP, 1 mismatch is system error. Most people run 6 mismatches for a 50bp tag if you have the processing power.

But, Bowtie only can set the mismatch up to 3 (-v). How could you fix it?

[Chipper]

But, Bowtie only can set the mismatch up to 3 (-v). How could you fix it?

It allows up to 3 mismatches in the seed, you don't need to set the -v parameter. You might need to increase -e to allow high-quality mismatches for SNPs though.

It is quite amasing to align 10 M reads in 15 min on a single node... But it seems strange that --best is not default.