High-throughput technologies have greatly advanced the biological research as well as brought demands for new statistical methods. We have developed three statistical methods specifically for microbiome and GWAS data respectively, i.e.,

a) A two-stage microbial association mapping framework which is statistically powerful in detecting disease/trait-associated microbial taxa at the lowest available taxonomic rank with the microbiome data;

b) The jointing model of the longitudinal microbiome data and the survival outcome. A zero-inflated scaled beta regression model with random effects is newly proposed to model the highly skewed relative abundance data;

c) The efficient estimation of odds ratios for candidate SNPs from GWAS following by replication studies.

海报