Skip to main content
. 2015 Dec 7;150(1):64–73. doi: 10.1093/toxsci/kfv261

FIG. 1.

FIG. 1.

Overview of the workflow. First, 2 datasets were transformed into document-based form, on which author-topic model was constructed. As a result, 2 matrices, NRs-topic and topic-genes probabilities, were estimated. Through shared topics, the associations between NRs and genes were investigated. (‘TP’ denotes a TGP expression profile; ‘NR’ denotes a nuclear receptor; ‘T’ denotes a topic; ‘G’ denotes a gene).