ENBIS-17 in Naples

9 – 14 September 2017; Naples (Italy) Abstract submission: 21 November 2016 – 10 May 2017

Covariates Selection in High-Dimensional Data Under Dependence, Application in Oncology

11 September 2017, 16:40 – 17:00


Submitted by
Anne Gégout-Petit
Anne Gégout-Petit (Université de Lorraine), Aurélie Gueudin (Université de Lorraine), Bérangère Bastien (Transgene), Yaojie Shi (Inria Grand Est, BIGS)
In order to select proteomics and/or transcriptomics expressions associated to the effect of a treatment, we have developed a methodology of selection and ranking of covariates linked to a variable of interest in a context of high-dimensional data under dependence but few observations. The methodology imbricates successively rough selection, clustering of variables whose aim is to detect the different independent blocks of covariates, decorrelation of variables using Factor Latent Analysis inside each block, selection using aggregation of adapted methods and finally ranking through bootstrap replications. The method can be adapted to a variable of interest that is qualitative or quantitative. Simulations are developed in order to show the interest of the different steps of the method.

Return to programme