Bowtie2 mapping and paired-end reads

[westeros]

Hi all,

I have started using bowtie2 to map 100bp Illumina reads to a reference genome. Our ultimate goal is to have gene expression estimates.

I have two questions:

1) Bowtie2 use a score to define if a read is aligned or not. What is the best setting to call a read uniquely aligned ? Most papers using small reads allow a few mismatches and the threshold is pretty easy to grasp. But how do we handle it using scores ?

2) What is the best way to use paired-end mappings and read counts ? Is it correct to count the two mates as two reads mapped to a gene ?

Thanks a lot.

[arcolombo698]

Score Threshold

Hello. I am interested in the similar topic regarding score threshold, but I am not sure how to set it.

I would guess the following from the manual

For an alignment to be considered "valid" (i.e. "good enough") by Bowtie 2, it must have an alignment score no less than the minimum score threshold. The threshold is configurable and is expressed as a function of the read length. In end-to-end alignment mode, the default minimum score threhsold is -0.6 + -0.6 * L, where L is the read length. In local alignment mdoe, the default minimum score threshold is 20 + 8.0 * ln(L), where L is the read length. This can be configured with the --score-min option. For details on how to set options like --score-min that correpond to functions, see the section on setting function options.


so if you know the length is 100 bp, you can calculate the min score using the above info.


however is the 100bp the FRAGMENT LENGTH of the template used for the sequencing machine? or the actual read itself? how do you calculate the length of read? I am not sure if this implies the fragment length used to initially prepare the sequence reads. This is important because after the sequencer outputs the reads, normally we have a distribution of read lengths.

[mastal]

It means the read length.

See the section of the manual under the heading Setting Function Options,

http://bowtie-bio.sourceforge.net/bowtie2/manual.shtml#setting-function-options

it tells you how to specify the parameters.

[dpryan]

The length is the length of the read and has nothing to do with the fragment length. If you have a 50bp read and L,-0.6,-0.6, then the minimum score is -31.6.