Re: [Denovoassembler-devel] clean up / simplify CoverageDistribution function

[Eccles, D. <dav...@mp...>]

From: Eccles, David
> ... and [as I've just noticed] I get a segfault on the 454 data with 
> my code when doing this.... I'll try to nail that down in the next day or
so.

Because there weren't already unit tests for Kmer a.compareTo(&b), I didn't
think to put them in my code. I only discovered the bug from a segfault that
happened because Ray was looking for a palindromic sequence with different
starting bases, and couldn't find it (something like A300330303033003A vs the
reverse-complement T300330303033003T). The problem was that it was clearing
the first/last bases when it wasn't meant to (i.e. first bases known), and
not clearing them when it was meant to (i.e. at least one first base
unknown).

https://github.com/gringer/ray/commit/451f766a5ce8b7da4ce6d340c234e7d4c5c7ca4
f

... I guess Kmer unit tests will be my next mini-project for Ray.

I've been thinking again about unknown first bases and assembly, and think
that probably the best thing to do would be to only store unknown first bases
in the Kmer academy. In other words, graduation would require both a known
first-base, and for the k-mer to occur more than once. It would mean that
there's no messing round with unknown bases when doing the assembly. However,
it would also mean that an assembly couldn't be done on sequences that all
have a misread at their second position.

-- David