[sparks]

Hi Susan,
I think you could align this using Novoalign as the N's won't count as full mismatches, only P=0.25 of mismatch and hence penalty of 6. Building the index with a k-mer length of 7 might improve performance. If you'd like to discuss further you can contact me via email at colin at novocraft <.>com

Colin

[axiom7]

Polonator data with gaps

Hi Ben,
I have some output from a Polonator. This is paired-end data with gaps. For instance, the raw data is 26 base pairs. The researcher asserts this to be 2x15mers with a gap of two nucleotides between base 7 and 8, and between 20 and 21. He also asserts that the spacing between the two 15mers is between 500 and 1500 bases. I used a perl script to insert "NN" in the two gaps, and to create two mated fasta files. Ran the following:

bowtie -t -p 8 -v 3 -m 100 -I 500 -X 1500 -f --ff -a -1 mate1.fa -2 mate2.fa

This seemed to run reasonably and I /think/ I am asking for alignments with 1 additional mismatch beyond the 2 gapped nucleotides.

Problem occurs in the second set of data. The researcher asserts the 26 base pairs to have a 6 nucleotide gap, but when I attempt to run the above bowtie command (after processing the raw data with my perl script) with "-v 7" I get an error message: "-v arg must be at most 3". Am I out of luck here? Am I asking bowtie to do something for which it is not designed?
Thank you.
Susan

[Ben Langmead]

I was playing around with the new bowtie version, and I noticed that specifying the --best flag resulted in an output that had a different number of mappings than when the flag is omitted.

I'll take a look. You used --al and --un. Is the number of reads different in those files, or in the alignment output? Or both?

[Ben Langmead]

When I attempt to align mated paired-end sequence reads and output the file
in SAM format, I receive a segmentation fault.

Hi kraigrs. I'll definitely take a look. Is it possible for me to get those input files from you? I can't reproduce that with the reads and indexes that I've tried on my end.

Thanks,
Ben

[kraigrs]

Trouble with bowtie printing SAM output

When I attempt to align mated paired-end sequence reads and output the file
in SAM format, I receive a segmentation fault. If I try the same thing
without the -S/--sam option, it works fine. Here is what I am getting:

EEB-WITT5:Bowtie wittkopp-lab$ ./bowtie -q -k 1 --sam --best
--solexa1.3-quals dmel-all-CDS-r5.21 -1
./mel_sim_data/Hybrids/s_2_1_sequence.txt -2
./mel_sim_data/Hybrids/s_2_2_sequence.txt > s_2_sequence.sam

Segmentation fault

Any help in this matter would be greatly appreciated! Again, I would like
this output to be in SAM format. I tried converting the bowtie output to
SAM but the bowtie2sam.pl script doesn't do that.

[jmj1091]

Hi Ben,
I was playing around with the new bowtie version, and I noticed that specifying the --best flag resulted in an output that had a different number of mappings than when the flag is omitted. However, in the bowtie documentation, it says that --best does not change which alignments are considered valid. I tried this with multiple sets of reads, and each time the number of alignments differed. Is this a bug or am I misunderstanding the effects of --best? The other flags I was using (in both cases) were -S, -p, --solexa1.3-quals, --al, and --un.
Thanks!

[Ben Langmead]

Ben, thanks for your so kind reply. I'm also wondering whether the "-3" option for the length of reads setting is correct. For example, if the length of my reads is 36bp and I want to the length just to be 25bp, so I set "-3 11" . Is that correct?

Yes, "-3 11" will turn an 36bp read into a 25bp read with 11 bases trimmed from the (3') end. To trim from the front (5') end, use "-5".

Thanks,
Ben

[lix]

Ben, thanks for your so kind reply. I'm also wondering whether the "-3" option for the length of reads setting is correct. For example, if the length of my reads is 36bp and I want to the length just to be 25bp, so I set "-3 11" . Is that correct?

Thanks again,
lix

[Ben Langmead]

Hi lix,

I am mapping the ChIP-seq reads(raw sequences) to the reference genome(homo sapiens) and I want to the length of reads to be 25bp, with 2 mismatches to be allowed. So, I choose the option " --best -r -3". Is this option setting correct for mapping?

For raw reads, yes, use "-r". For allowing up to 2 mismatches, use "-v 2". It's also reasonable to use "--best"; it's noticeably slower than the default mode, but you might prefer having the best-ness guarantee.

I also build a reference genome index using "bowtie-build" option and is there any difference using this compared with the pre-built indexes on the website?

No difference. If you build an index using any of the scripts in the "scripts" subdirectory of the bowtie package, you should end up with an index that's effectively identical to the corresponding pre-built index. All pre-built indexes are built with bowtie-build.

Thanks,
Ben

[lix]

Hi Ben,

Thank you for Bowtie. It seems to be widely used and I am running it recently. There are some option settings I'm confused and I hope you can give me some kind suggestions:

I am mapping the ChIP-seq reads(raw sequences) to the reference genome(homo sapiens) and I want to the length of reads to be 25bp, with 2 mismatches to be allowed. So, I choose the option " --best -r -3". Is this option setting correct for mapping? If not, how the option should be setted?
I also build a reference genome index using "bowtie-build" option and is there any difference using this compared with the pre-built indexes on the website?

Thanks,

-lix

[para_seq]

In effect, that's what the -v/-n settings do. Ns count as mismatches, so if the alignment policy is "-v 2", reads with more than 2 Ns will not align. If Bowtie's options don't do exactly what you need, you could also write a very simple script that filters out reads according to your own criteria beforehand.



That depends very much on the quality of the data. Some sets of reads have many, many Ns, and some have Ns systematically at certain positions. My advice is to try various parameters and see what seems to give the best result.

Thanks,
Ben

Thank very much, Ben.

[Ben Langmead]

1. I wonder if there is a switch in Bowtie to filter out reads that contain more than certain number of 'N' s.

In effect, that's what the -v/-n settings do. Ns count as mismatches, so if the alignment policy is "-v 2", reads with more than 2 Ns will not align. If Bowtie's options don't do exactly what you need, you could also write a very simple script that filters out reads according to your own criteria beforehand.

2. How many 'N's in each read (e.g. 35-nt long) do people usually allow for Illumina data? Thank you.

That depends very much on the quality of the data. Some sets of reads have many, many Ns, and some have Ns systematically at certain positions. My advice is to try various parameters and see what seems to give the best result.

Thanks,
Ben

[Ben Langmead]

But it would be a lot cleaner just to have a --circular option.

You're correct that there is no --circular option. Your workaround is good, though.

Thanks,
Ben

[SillyPoint]

Circular genomes

Does Bowtie make any provision for circular genomes?

I could fake it out by appending N-1 bases from the start of the reference onto the end of it, where N is the read length. But I'd need a separate index for each value of N.

Or perhaps not. If I use a large N, some reads will map both to the real position near the start, and to the same sequence in the appended copy. Running with --best and -k 1 (the default), I should get a single report at one of those positions.

But it would be a lot cleaner just to have a --circular option.

--TS

[Ben Langmead]

Can Bowtie do DNA methylation aligment? I haven't found in your manual. How to define the parameter?

There are ad hoc ways of using Bowtie to do DNA methylation alignment, e.g., you can index and query the strands of your target genome separately and change the bases to mimic the bisulfite reaction (Cs -> Ts, for the most part). Refinements to this scheme are also possible (e.g. treat lone Cs and CpGs differently when mimicking the bisulfite reaction). As usual, the two main problems are loss of signal (more mismatches -> fewer reads align) and signal bias (depending on the conversion scheme, there might be an inherent coverage bias toward methylated or toward unmethylated sites).

I am also looking at trying to support bisulfite more directly in the future.

Thanks,
Ben