Bayesian Network Model Predicting Personal Phenotypes using Genome-Sequencing Data
Reference:
Chen YC, Douville C, Wang C, Niknafs N, Yeo G, Beleva-Guthrie V, et al. A probabilistic model to predict clinical phenotypic traits from genome sequencing. PLoS Comput Biol. 2014;10(9):e1003825.
Link: https://journals.plos.org/ploscompbiol/article?id=10.1371/journal.pcbi.1003825
This model won the PGP (Personal Genome Project) challenge in CAGI (Critical Assessment of Genome Interpretation) 2013 competition (https://genomeinterpretation.org/content/cagi-2013), in which the model was used to predict 239 human traits for 77 individuals using their WGS (Whole-Genome Sequencing) data.
In the paper, we demonstrate the ability of the model by predicting 146 phenotypes for 130 individuals using their WGS data. Phenotype information and WGS data were downloaded from PGP website: https://www.personalgenomes.org/us
Figure 1 in the paper shows 38 best predicted phenotypes with AUC (Area Under ROC Curve) > 0.7: https://github.com/Yun-Ching-Chen/GIBI/blob/master/images/Figure1.pdf