Fix out-of-bounds error

[standage]

This pull request provides a simple solution to the problem described in #96. If the computed index is greater than the length of the sequence, append A. Otherwise, append the requested nucleotide.

I'm assuming this problem doesn't occur much with whole genome sequences, since it's very rare for a read both to have an indel and line up right at the end of the chromosome. However, when sequencing short amplicons in draft mode this case occurred with high frequency.

There are probably more elegant solutions to this problem. One could:

• draw a random base instead of A
• draw a base from some distribution of nucleotide composition
• append Illumina adapters or other technical sequences to the end of the genomic sequence

But for now, this is a simple solution that allows InSilicoSeq to run without barfing on amplicon sequences.

[codecov]

Codecov Report

Merging #97 into master will decrease coverage by 0.36%.
The diff coverage is 0%.

@@            Coverage Diff             @@
##           master      #97      +/-   ##
==========================================
- Coverage   91.54%   91.18%   -0.37%     
==========================================
  Files          12       12              
  Lines         769      771       +2     
==========================================
- Hits          704      703       -1     
- Misses         65       68       +3

Impacted Files	Coverage Δ
iss/error_models/__init__.py	77.9% <0%> (-1.86%)	⬇️
iss/util.py	97.29% <0%> (-0.91%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update cd73724...e8e7681. Read the comment docs.

[HadrienG]

Good catch, thanks for the PR.

This hasn't happened yet with whole genome sequences.
I like the adapter solution in theory, but in practise we might never encounter two deletions in a row at the end of a sequence, so a A can do.

It'd be a good idea to implement adapter sequences in general. By padding the reads with adapter sequences this bug should disappear and we can get rid of your "A" then 😄

[HadrienG]

I'm merging this and will publish a bugfix release soon. Thanks again!