Redux with a "P" option does not return P-groups for two-field or three-field alleles

[gbiagini]

When supplying a two-field allele to ard.redux() with the "P" option, the allele name is generally returned instead of the P-group.

This may primarily affect alleles with additional four-field variants.

For example, ard.redux("B*07:02","P") returns B*07:02 (as does ard.redux("B*07:02:01","P")), while ard.redux("B*07:02:01:01","P") returns B*07:02P.
Similarly, ard.redux("DRB1*01:01","P") returns DRB1*01:01 (as does ard.redux("DRB1*01:01:01","P")), while ard.redux("DRB1*01:01:01:01","P") returns DRB1*01:01P.

For an example of an allele without a four-field variant that behaves normally, ard.redux("B*15:14","P") returns B*15:14P, as expected.

The expected behavior is that ard.redux() would return the same P-group for the two-field, three-field, and four-field equivalents of an allele.

The addition of P-groups to ard.redux() is very helpful and I thank you for this tool and for your time. I'm happy to aid in debugging if desired.

[pbashyal-nmdp]

Thanks Giovanni for the detailed examples. I'll take a look at what's causing the difference.

[pbashyal-nmdp]

Hello @gbiagini,

Just added the fix for P group mapping. Could you test it on your code ?
You can install from the current code with pip by:

pip install git+https://github.com/nmdp-bioinformatics/py-ard

[gbiagini]

Hi @pbashyal-nmdp

I tested the same alleles listed above with the same results. Maybe I'm missing something about getting the updated code?

I tried installing py-ard from the standard pip install py-ard as well as the command you provided, with the same results.

I also tried reinstalling the IMGT database for version 3.52.0, but that didn't seem to change anything either.

I tested both the Python module and CLI. Results of the CLI pasted below.

[pbashyal-nmdp]

It gives a correct answer on mine.

$ pyard -g "B*07:02" -r "P"
B*07:02P

It's not published yet(hopefully soon) so you can't do pip install py-ard yet.

Can you run pyard-status command and post it here? It should list what data you have.

$ pyard-status
-------------------------------------------
IMGT DB Version: Latest (3520)
You're up to date. 3520 is the most recent version.
File: /var/folders/nm/nyjkgrfn3fg207v5z5rz1w8m0000gp/T/pyard/pyard-Latest.sqlite3
Size: 558.89MB
-------------------------------------------
|Table Name          |Rows                |
|-----------------------------------------|
|dup_g               |                  67|
|dup_lgx             |                   1|
|g_group             |               14445|
|p_group             |               21157|
|lgx_group           |               14445|
|exon_group          |               13140|
|p_not_g             |                1690|
|xx_codes            |                1995|
|who_group           |               36036|
|alleles             |               39386|
|exp_alleles         |                  84|
|who_alleles         |               37068|
|mac_codes           |             1113880|
-------------------------------------------

[gbiagini]

I got it working just now!

Instead of forcing a re-install of the database through py-ard, I uninstalled py-ard again, deleted the entire cache in the /tmp/pyard directory, and then reinstalled again with the pip install git+https://github.com/nmdp-bioinformatics/py-ard command you provided.

Now "B*07:02" reduces to "B*07:02P".

Previously, I was forcing a re-install of the database after uninstalling and reinstalling py-ard. A clean install may have been what fixed the inconsistency on my end, can't say for sure.

[gbiagini]

Thanks for all the help @pbashyal-nmdp!

[pbashyal-nmdp]

Thanks for testing it and reporting the issue. It'll be part of py-ard with the next release of 1.0.1.