Transcriptome prediction performance across machine learning models and diverse ancestries.

Autor:	Okoro PC; Program in Bioinformatics, Loyola University Chicago, Chicago, IL, USA., Schubert R; Department of Mathematics and Statistics, Loyola University Chicago, Chicago, IL, USA., Guo X; Institute for Translational Genomics and Population Sciences, The Lundquist Institute and Department of Pediatrics at Harbor-UCLA Medical Center, Torrance, CA, USA., Johnson WC; Department of Biostatistics, University of Washington, Seattle, WA, USA., Rotter JI; Institute for Translational Genomics and Population Sciences, The Lundquist Institute and Department of Pediatrics at Harbor-UCLA Medical Center, Torrance, CA, USA., Hoeschele I; Fralin Life Sciences Institute, Virginia Tech, Blacksburg, VA, USA.; Department of Statistics, Virginia Tech, Blacksburg, VA, USA.; Wake Forest School of Medicine, Winston-Salem, NC, USA., Liu Y; Department of Medicine, Duke University School of Medicine, Durham, NC, USA., Im HK; Section of Genetic Medicine, Department of Medicine, University of Chicago, Chicago, IL, USA., Luke A; Department of Public Health Sciences, Parkinson School of Health Sciences and Public Health, Loyola University Chicago, Maywood, IL, USA., Dugas LR; Department of Public Health Sciences, Parkinson School of Health Sciences and Public Health, Loyola University Chicago, Maywood, IL, USA.; Department of Human Biology, Faculty of Health Sciences, University of Cape Town, Cape Town, South Africa., Wheeler HE; Program in Bioinformatics, Loyola University Chicago, Chicago, IL, USA.; Department of Biology, Loyola University Chicago, Chicago, IL, USA.; Department of Computer Science, Loyola University Chicago, Chicago, IL, USA.
Jazyk:	angličtina
Zdroj:	HGG advances [HGG Adv] 2021 Apr 08; Vol. 2 (2). Date of Electronic Publication: 2021 Jan 05.
DOI:	10.1016/j.xhgg.2020.100019
Abstrakt:	Transcriptome prediction methods such as PrediXcan and FUSION have become popular in complex trait mapping. Most transcriptome prediction models have been trained in European populations using methods that make parametric linear assumptions like the elastic net (EN). To potentially further optimize imputation performance of gene expression across global populations, we built transcriptome prediction models using both linear and non-linear machine learning (ML) algorithms and evaluated their performance in comparison to EN. We trained models using genotype and blood monocyte transcriptome data from the Multi-Ethnic Study of Atherosclerosis (MESA) comprising individuals of African, Hispanic, and European ancestries and tested them using genotype and whole-blood transcriptome data from the Modeling the Epidemiology Transition Study (METS) comprising individuals of African ancestries. We show that the prediction performance is highest when the training and the testing population share similar ancestries regardless of the prediction algorithm used. While EN generally outperformed random forest (RF), support vector regression (SVR), and K nearest neighbor (KNN), we found that RF outperformed EN for some genes, particularly between disparate ancestries, suggesting potential robustness and reduced variability of RF imputation performance across global populations. When applied to a high-density lipoprotein (HDL) phenotype, we show including RF prediction models in PrediXcan revealed potential gene associations missed by EN models. Therefore, by integrating other ML modeling into PrediXcan and diversifying our training populations to include more global ancestries, we may uncover new genes associated with complex traits. Competing Interests: Declaration of interests The authors declare no competing interests.
Databáze:	MEDLINE
Externí odkaz:	Zobrazit plný text záznamu