Taking into account phenotype frequencies

[julesjacobsen]

For example in Decipher ARID1B

Code    Phenotype   Affected patients with Sequence Variants in ARID1B  Percentage of phenotyped patients with Sequence Variants in ARID1B  Affected patients in DECIPHER   Percentage of 12981 phenotyped patients in DECIPHER p value Odds ratio
HP:0001601  Laryngomalacia  4   12.90%  29  0.22%   6.00 × 10-07   76.59 ( 24.97 - 234.98 )
HP:0012718  Morphological abnormality of the gastrointestinal tract     8   25.81%  363     2.80%   1.44 × 10-06   12.34 ( 5.48 - 27.78 )
HP:0002307  Drooling    4   12.90%  36  0.28%   1.47 × 10-06   59.81 ( 19.79 - 180.73 )

HP:0001601 and HP:0002307 are highly significant phenotypic markers of ARID1B pathology.

[damiansm]

Is this related to phenotype annotation frequency data as provided by Orphanet to weight our scores. Need to wait until the Bayesian implementation to do properly