Why is PhenoSV score always `0.0`?

WGLab / PhenoSV

PhenoSV: Interpretable phenotype-aware model for the prioritization of genes affected by structural variants.

MIT License

16 stars 3 forks source link

Why is PhenoSV score always `0.0`? #10

Closed jsquaredosquared closed 4 months ago

jsquaredosquared commented 4 months ago

Hi,

I have tried to score the examples in the README using PhenoSV-light, but the Phen2Gene and PhenoSV scores are always 0. What could be the reason for this?

I would appreciate your guidance. Thanks

python phenosv/model/phenosv.py --model PhenoSV-light \
--c chr6 --s 156994830 --e 157006982 --svtype 'deletion' \
--noncoding 'tad' \
--HPO 'HP:0000707,HP:0007598'

  Elements  Pathogenicity           Type  Phen2Gene  PhenoSV
0       SV       0.607864  Non-coding SV          0      0.0
1   ARID1B       0.550785       Intronic          0      0.0
2     NOX3       0.155716     Regulatory          0      0.0
3    TFB1M       0.249238     Regulatory          0      0.0

Karenxzr commented 4 months ago

Hi,

I have tried to score the examples in the README using PhenoSV-light, but the Phen2Gene and PhenoSV scores are always 0. What could be the reason for this?

I would appreciate your guidance. Thanks

python phenosv/model/phenosv.py --model PhenoSV-light \
--c chr6 --s 156994830 --e 157006982 --svtype 'deletion' \
--noncoding 'tad' \
--HPO 'HP:0000707,HP:0007598'

  Elements  Pathogenicity           Type  Phen2Gene  PhenoSV
0       SV       0.607864  Non-coding SV          0      0.0
1   ARID1B       0.550785       Intronic          0      0.0
2     NOX3       0.155716     Regulatory          0      0.0
3    TFB1M       0.249238     Regulatory          0      0.0

Hi, it is because genes that the given SV affected doesn’t show any associations with given HPO terms based on phen2gene.

jsquaredosquared commented 4 months ago

But according to the example in the README, there should be some association?

  Elements  Pathogenicity           Type  Phen2Gene   PhenoSV
0       SV       0.664912  Non-coding SV   0.999126  0.664331
1   ARID1B       0.823556       Intronic   0.999126  0.822836
2     NOX3       0.045648     Regulatory   0.837460  0.038229
3    TFB1M       0.533431     Regulatory   0.544762  0.290593

When I score variants using the web tool, there is an association. But when I use the CLI, the values are always 0.

Karenxzr commented 4 months ago

It's working well on my end. Can you double confirm the data file for phen2gene has been installed correctly? (knowledgebase, skewness, weights)

jsquaredosquared commented 4 months ago

I forgot to unzip the file 🤦

It works now.

Thanks for your help 🙂