search

uab-cgds-worthey / DITTO

Variant Deleteriousness prediction tool using AI
GNU General Public License v3.0
1 stars 0 forks source link

not giving same predictions when running multiple times #7

Closed ManavalanG closed 1 year ago

ManavalanG commented 3 years ago

In GitLab by @tkmamidi on Oct 11, 2021, 09:52

Observed in CAGI6 challenge...

Below is the variant in different samples. E = Exomiser, D = Ditto, P = combined score

Sample: CAGI6_RGP_TRAIN_18_PROBAND

PROBANDID,SYMBOL,CHROM,POS,REF,ALT,E,D,P,SD,C
# Depth 6
CAGI6_RGP_TRAIN_18_PROBAND,FREM2,chr13,38864311,A,G,0.4639,0.9056842357996026,0.6847921178998013,0,*
CAGI6_RGP_TRAIN_18_PROBAND,FREM2,chr13,38864311,A,G,0.4639,0.9056842357996026,0.6847921178998013,0,*

# depth 8
CAGI6_RGP_TRAIN_18_PROBAND,FREM2,chr13,38864311,A,G,0.4639,0.9338298442747104,0.6988649221373552,0,*
CAGI6_RGP_TRAIN_18_PROBAND,FREM2,chr13,38864311,A,G,0.4639,0.9338298442747104,0.6988649221373552,0,*
ManavalanG commented 3 years ago

In GitLab by @tkmamidi on Oct 13, 2021, 09:55

Found the issue. It has to do with complex columns like VEST4 not being read properly by pandas. Fixed the issue.