Analysis of genomic markers: Make it easy with the R package MPAgenomics

Quentin Grimonprez (1) , Alain Celisse (2) , Guillemette Marot (1, 3)

1 MODAL - MOdel for Data Analysis and Learning
2 LPP - Laboratoire Paul Painlevé - UMR 8524
3 METRICS - Evaluation des technologies de santé et des pratiques médicales - ULR 2694

MPAgenomics, standing for multi-patients analysis of genomicmarkers, is an R-package which enables to study several copy number andSNP data profiles at the same time. It offers wrappers from commonlyused packages (aroma [Bengtsson, 2004], changepoint [Killick et al., 2013],cghcall [van de Wiel et al., 2007], glmnet [Friedman et al., 2010]) to of-fer a pipeline for beginners in R. Thus, normalization, segmentation andcalling of copy-number or SNP data profiles can be easily performed atthe same time. The special architecture of [Bengtsson, 2004] packages isautomatically created and outputs to right formats for further analysesare suggested. Therefore, more advanced users can use their own methodat one point if they prefer to separate the global analysis in several steps.In addition to these useful wrappers, we propose a strategy to improvethe choice of the penalty parameter used in [Killick et al., 2013] for seg-mentation. As far as multi-patients analysis is concerned, we suggest toselect relevant markers associated with a given response thanks to wrap-pers with the R-packages glmnet [Friedman et al., 2010] and HDPenReg(still under development). Functionalities already implemented in ourpackage HDPenReg, standing for penalized regression in high dimension,will be briefly presented.

Mots clés: segmentation; genomic marker selection; SNP array; calling; normalization; multi patient analysis

Domaines: Statistiques [stat]; Bio-Informatique, Biologie Systémique [q-bio.QM]

Quentin Grimonprez, Alain Celisse, Guillemette Marot. Analysis of genomic markers: Make it easy with the R package MPAgenomics. SMPGD 2014, Jan 2014, Paris, France. , 2014. ⟨hal-01091543⟩