Y. Larriba, C. Rueda
Peaks of gene expressions are crucial in circadian biology as it determines the time of day when the gene biological function is performed. The 50% of gene expressions in mammals display rhythmic patterns, with expressions varying across tissues. Due to the risk that it entails, in most human gene studies sampling times are unrecorded. This work proposes CIRCUST, a methodology that robustly estimates the temporal order among samples and accurately analyzes molecular rhythms. It relies on the formulation of the problem within the Circular Space and on the use of non-linear models for the analysis of oscillatory signals. CIRCUST has been validated on two controlled experiments. Its great performance is demonstrated for the GTEx database, with high inter-individual variability. For 34 human tissues, the temporal order estimation problem is solved, molecular rhythms are analyzed and clock molecular networks are described to provide, as a culmination, an atlas of human circadian expressions
Keywords: Temporal Order Estimation, Oscillatory Signal Analysis, Non Linear Models, Circular Principal Components, Gene Expression Analysis, Circadian Rhythms
Scheduled
Biostatistics I
June 9, 2022 10:10 AM
A13
