Fix inconsistent subtype function

[peterk87]

Sample SRR1958005 returns multiple subtype calls 2.1; 2.2 with all subtype calls being 2; 2.1; 2.2. Subtypes 2.1 and 2.2 are not consistent with each other.

Testing bio_hansel -s enteritids on the Enterobase genome assembly of SRR1958005 (46121.fasta) produced the following results:

sample	scheme	subtype	all_subtypes	tiles_matching_subtype	are_subtypes_consistent	inconsistent_subtypes	n_tiles_matching_all	n_tiles_matching_all_total	n_tiles_matching_positive	n_tiles_matching_positive_total	n_tiles_matching_subtype	n_tiles_matching_subtype_total	file_path
46121	enteritidis	2.1; 2.2	2; 2.1; 2.2	134979-2.1; 1402592-2.1; 2483597-2.2	TRUE		652	652;652	14	13;36	3	2;25	/home/peter/Downloads/46121.fasta

Other samples which should have been are_subtypes_consistent == False include:

• SRR3049064; all_subtypes == "2; 2.1; 2.2.2.1.1"
• SRR1968431; all_subtypes == "2; 2.1; 2.2.2.1.1"
• ERR1069299; all_subtypes == "2; 2.1; 2.2.2"
• SRR1968272; all_subtypes == "2; 2.1; 2.2.2"

See function here:
https://github.com/phac-nml/bio_hansel/blob/master/bio_hansel/utils.py#L80

Also, would be a good idea to add a test for that function.

[mgopez]

When you mean inconsistent, is that because in 2.1 and 2.2 the ._ value is different?

[peterk87]

Yes, so the inconsistent subtypes would be 2.1 and 2.2. I originally had the function try to find the highest resolution subtype (i.e. longest string of numbers) and use that as the reference to compare against, but in cases where the highest resolution subtypes are equal length, the function fails to report inconsistencies.

[mgopez]

I think I found an issue.
https://github.com/phac-nml/bio_hansel/blob/ddee5c948b40722e2fe11282809fa4b05faf3292/bio_hansel/utils.py#L98
I believe it's supposed to freq >= 1.

If both strings are same length, then >= 1 would add both to the incon_subtypes list.