Short Read Simulator: Simple Math For Coverage ?

Question

[Pierre Lindenbaum]

I'm writing my own short read read simulator (mostly for fun, I known I'm reinventing the wheel).

The 'beta' source code is available here: http://code.google.com/p/variationtoolkit/source/browse/trunk/src/shortreadsim.cpp

Here is my question:

• each short read has a length = 'short_ read_length'
• a gene has a length = 'segLength'

how many short-reads do I need to create to have an average coverage = 'N' for this gene ? (simple math ?)

Answer 1

[Michael]

ahem, yes very simple I guess, given all read come only from this gene, that's what you want right?

let r be the read length, g be the gene length, r<=g then each read contributes coverage c = r/g. So you need N/(r/g) reads. E.g. if r=g and aimed coverage is 1 then you need 1 read. Let alone those reads that do not fully overlap the gene, they contribute ofc only partially. If you want to include these into the calculation and want the exact coverage, you can only do it after you placed each read randomly and then recompute the coverage.

Or did I miss something completely??

[Pierre Lindenbaum]

Thanks Michael, I must be very tired... :-)