Multi-Task Feature Selection on Multiple Networks via Maximum Flows

Sugiyama, Mahito; Azencott, Chloé-Agathe; Grimm, Dominik G.; Kawahara, Yoshinobu; Borgwardt, Karsten

doi:10.1137/1.9781611973440.23

Cited by 10 publications

(5 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…62, 66), the importance of rare variants (reviewed in refs. 62, 64), the contribution of multiple different alleles at the same locus (allelic heterogeneity) (50,(67)(68)(69), the change in allelic Table S4). Gene models (TAIR10) are shown in B, and AGL50 (AT1G59810) is highlighted in purple.…”

Section: Discussionmentioning

confidence: 99%

Genetic architecture of nonadditive inheritance inArabidopsis thalianahybrids

Seymour

Chae

Grimm

et al. 2016

Proc. Natl. Acad. Sci. U.S.A.

Self Cite

View full text Add to dashboard Cite

The ubiquity of nonparental hybrid phenotypes, such as hybrid vigor and hybrid inferiority, has interested biologists for over a century and is of considerable agricultural importance. Although examples of both phenomena have been subject to intense investigation, no general model for the molecular basis of nonadditive genetic variance has emerged, and prediction of hybrid phenotypes from parental information continues to be a challenge. Here we explore the genetics of hybrid phenotype in 435 Arabidopsis thaliana individuals derived from intercrosses of 30 parents in a half diallel mating scheme. We find that nonadditive genetic effects are a major component of genetic variation in this population and that the genetic basis of hybrid phenotype can be mapped using genome-wide association (GWA) techniques. Significant loci together can explain as much as 20% of phenotypic variation in the surveyed population and include examples that have both classical dominant and overdominant effects. One candidate region inherited dominantly in the half diallel contains the gene for the MADS-box transcription factor AGAMOUS-LIKE 50 (AGL50), which we show directly to alter flowering time in the predicted manner. Our study not only illustrates the promise of GWA approaches to dissect the genetic architecture underpinning hybrid performance but also demonstrates the contribution of classical dominance to genetic variance.T he often observed phenotypic superiority of progeny relative to their parents, or heterosis, is a universal phenomenon and of great importance to plant agriculture. The earliest description of heterosis (also known as hybrid vigor or superiority) dates to Darwin's studies of cross-fertilization in plants. He noticed that intercrossing distantly related individuals gave rise to larger, more vigorous progeny (1). Four decades later, George Shull coined the term "heterosis" (2) for this phenomenon, which he and Edward East had independently described for hybrids of inbred maize in 1908 (3, 4). Heterosis has long been of interest to evolutionary biologists as a potential explanation for the ubiquity of cross-fertilization in plants and animals, but it is also a central component of agricultural breeding programs. The combination of hybrid seed technology and inbred line improvement has driven an unprecedented improvement in maize yield over the past century (5). Despite the economic importance of heterosis and its intensive investigation in a wide spectrum of species, prediction of hybrid performance from parental information remains a major challenge (6).In the terms of quantitative genetics, hybrid vigor (and its opposite, inferiority) describes a deviation of progeny from the phenotypic mean of the parents. This means that heterosis cannot be explained by the addition of the effects of contributing alleles (7). Nonadditive genetic variance can result from a nonlinear phenotypic effect of alleles at one locus, as in the case of dominant/recessive allele pairs in classical genetics, or from epistatic interactions ...

show abstract

Section: Discussionmentioning

confidence: 99%

Genetic architecture of nonadditive inheritance inArabidopsis thalianahybrids

Seymour

Chae

Grimm

et al. 2016

Proc. Natl. Acad. Sci. U.S.A.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Furthermore, easyGWAS supports the automatic import of public phenotypes from AraPheno (https://arapheno.1001genomes.org), a central repository for population-scale phenotype data from Arabidopsis (Seren et al, 2016). Users can also upload their custom summary statistics of GWAS performed in different environments, for example, from offline analyses with PLINK (Purcell et al, 2007) or other third-party tools (Kang et al, 2010;Lippert et al, 2011;Rakitsch et al, 2013;Azencott et al, 2013;Sugiyama et al, 2014;Llinares-López et al, 2015), for visualization, subsequent meta-analysis, or comparison with GWAS results that have already been deposited in easy-GWAS. Detailed descriptions of the different views can be found in the supplemental data (Supplemental Figures 2 to 6 and Supplemental Text 1).…”

Section: Data Repositorymentioning

confidence: 99%

easyGWAS: A Cloud-Based Platform for Comparing the Results of Genome-Wide Association Studies

et al. 2016

Self Cite

View full text Add to dashboard Cite

The ever-growing availability of high-quality genotypes for a multitude of species has enabled researchers to explore the underlying genetic architecture of complex phenotypes at an unprecedented level of detail using genome-wide association studies (GWAS). The systematic comparison of results obtained from GWAS of different traits opens up new possibilities, including the analysis of pleiotropic effects. Other advantages that result from the integration of multiple GWAS are the ability to replicate GWAS signals and to increase statistical power to detect such signals through meta-analyses. In order to facilitate the simple comparison of GWAS results, we present easyGWAS, a powerful, species-independent online resource for computing, storing, sharing, annotating, and comparing GWAS. The easyGWAS tool supports multiple species, the uploading of private genotype data and summary statistics of existing GWAS, as well as advanced methods for comparing GWAS results across different experiments and data sets in an interactive and user-friendly interface. easyGWAS is also a public data repository for GWAS data and summary statistics and already includes published data and results from several major GWAS. We demonstrate the potential of easyGWAS with a case study of the model organism Arabidopsis thaliana, using flowering and growth-related traits.

show abstract

“…Multi-task regularized linear regression Many multi-task variants of the lasso have been proposed [64,33], and can be extended in spirit to various structural regularizers, such as Grace [65]. Assuming T tasks, each containing n t training samples, and denoting by β t the m-dimensional vector of regression weights for task t, the first of these approaches consists in solving the optimization problem defined by Eq.…”

Section: Multi-task Extensionsmentioning

confidence: 99%

“…Multi-task penalized relevance Because of the computational efficiency of graphcut implementations, SConES can be extended to the multi-task setting in such a way as to address these issues. Multi-SConES [65] proposes a multi-task feature selection coupled with multiple network regularizers to improve feature selection in each task by combining and solving multiple tasks simultaneously. The formulation of Multi-SConES is obtained by the addition of a regularizer across tasks.…”

Section: Multi-task Extensionsmentioning

confidence: 99%

Network-Guided Biomarker Discovery

Azencott¹

2016

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

Identifying measurable genetic indicators (or biomarkers) of a specific condition of a biological system is a key element of precision medicine. Indeed it allows to tailor diagnostic, prognostic and treatment choice to individual characteristics of a patient. In machine learning terms, biomarker discovery can be framed as a feature selection problem on whole-genome data sets. However, classical feature selection methods are usually underpowered to process these data sets, which contain orders of magnitude more features than samples. This can be addressed by making the assumption that genetic features that are linked on a biological network are more likely to work jointly towards explaining the phenotype of interest. We review here three families of methods for feature selection that integrate prior knowledge in the form of networks.Comment: 18 pages, 1 figure, published in LNCS 9605 (Machine Learning for Health Informatics

show abstract

Multi-Task Feature Selection on Multiple Networks via Maximum Flows

Cited by 10 publications

References 28 publications

Genetic architecture of nonadditive inheritance inArabidopsis thalianahybrids

Genetic architecture of nonadditive inheritance inArabidopsis thalianahybrids

easyGWAS: A Cloud-Based Platform for Comparing the Results of Genome-Wide Association Studies

Network-Guided Biomarker Discovery

Contact Info

Product

Resources

About