Disease Risk Prediction

Prediction Combine Multiple Ancestries:

We developed an individualized risk prediction algorithm to combine multiply ancestry genotype data. It is the first model in literatures to incorporate an individual's genetic ancestry proportion into risk prediction.

US Patent, European Patent:

• Prediction models incorporating stratification of data.

Chinese Patent:

• 并入数据分组的预测模型.

Paper: Xia, X., Zhang, Y., Sun, R., Wei, Y., Li, Q., Chong, K. C., Wu, W. K. K., Zee, B. C. Y., Tang H., Wang, M. H. (2022). A Prism Vote method for individualized risk prediction of traits in genotype data of multiple population. PLoS Genetics, 18(10), e1010443. IF: 6.020, Q1, Rank 27/175 (15.43%) in Genetics & Heredity. 

Prediction Integrate Omics Data:

We used Artificial Neural Network to combine DNA methylation and single-nucleotide polymorphisms(SNPs). A significant improvement in prediction accuracy was achieved compared with using only SNPs. 

This paper rank 15% in The Genetic Analysis Workshops (GAW) 20 at San Diego, USA, and has been promoted to publish in BMC Genetics.

 
Paper: Xia, X., Weng, H., Men, R., Sun, R., Zee, B. C. Y., Chong, K. C., Wang, M. H. (2018) Incorporating methylation genome information improves prediction accuracy for drug treatment responses. BMC Genetics, 19(1), 67-71. IF: 2.759.