Use bowtie2 to analysis DNA-reseq data

[Lv Ray]

Hi dear,
I am going to deal with my big DNA-reseq data by bowtie2! I have finished index buliding , now I am preparing for mapping. But there're so many arguments to adjust! can I just use all the default ? but it seems not good.
Can anybody tell me a complete solution or a paper about bowtie2-mapping pipeline? please Add your detail reasons for every argument, that is best!

[gringer]

Use the default options initially, because it's reasonable to assume that the people who wrote the program have a good idea about the behaviour of their program, and have set up the defaul options to work the best in a general situation. You're going to be splitting hairs (and racking up bioinformatics time) by playing around with bowtie options, so save that for when the reviewers of your Nature paper get back to you.

[Lv Ray]

gringer, thanks for your answer ! I will use the default options to try my data. However, I don't agree your second point that the default options is the best. Becuase I think they set so many options in order to adapt other situations.

[gringer]

Not the best in all situations, but they should be the best in most situations. To say that is not the case is implicitly saying that you understand the program better than the people who wrote it, which is a dangerous path to travel down.

If you're interested in differential expression and fold-change differences of 1.1x, then fiddling with the options will probably have a big impact on your results. However if you're only fishing for the most obvious differences (i.e. "statistically significant"), and fold changes of 2x or more, then a few more reads here and there shouldn't affect the outcome of your analysis (and if it does, you probably shouldn't trust your answer).

[Lv Ray]

Thank you very much.