GFF parsing in count_reads.R

[thomcuddihy]

Had an issue where making ref_gene_df failed due to double the number of locus_tags compared to other features. On investigating, locus_tags contained both "locus_tag=..." and "old_locus_tag=..." entries.

Manually changed count_reads.R from:

locus_tags <- unlist(lapply(gene_attr, function(x) {
    x[grepl("locus_tag", x)]
}))

to:

locus_tags <- unlist(lapply(gene_attr, function(x) {
    x[grepl("^locus_tag", x)]
}))

Might be worth changing all feature greps ("locus_tag", "gene_biotype" and "gene") to "^feature=" e.g.

Didn't think it was worth a PR, since it could also be considered an issue with user input file.

Note sure if related to issue #8

[adamd3]

Thanks for reporting and for providing the fix. This is now merged.