KAML: improving genomic prediction accuracy of complex traits using machine learning determined parameters

Yin, Lilin; Zhang, Haohao; Zhou, Xiang; Yuan, Xiaohui; Song, Zhao; Li, Xinyun; Liu, X.

doi:10.1186/s13059-020-02052-w

Cited by 65 publications

(53 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In each year's data, both are ISR performs best, and followed BayesA, BayesB, BayesLASSO, and rrBLUP, BSLMM, BayesC, and DPR. 19,20 . In contrast, rrBLUP is the faster method, while DPR, BayesR, and BSLMM are as same as computationally efficient.…”

Section: Resultsmentioning

confidence: 99%

“…Therefore, in recent years, researchers have regarded phenotype prediction as a critical step in joint functional genomics and genome-wide research 10,18 . However, with the growth of high-throughput genomics data, accurate phenotype prediction requires the development of statistical methods that can simulate all or majors SNPs simultaneously 9,19,20 . Moreover, previous genome-wide association analysis studies have shown that many complex trait phenotypes and diseases have a polygenic genetic background, mainly controlled by many genetic variation sites with smaller effects.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Polygenic Prediction of Complex Traits with Iterative Screen Regression Models

Luo

Gu²

2020

Preprint

View full text Add to dashboard Cite

Although genome-wide association studies have successfully identified thousands of markers associated with various complex traits and diseases, our ability to predict such phenotypes remains limited. A perhaps ignored explanation lies in the limitations of the genetic models and statistical techniques commonly used in association studies. However, using genotype data for individuals to perform accurate genetic prediction of complex traits can promote genomic selection in animal and plant breeding and can lead to the development of personalized medicine in humans. Because most complex traits have a polygenic architecture, accurate genetic prediction often requires modeling genetic variants together via polygenic methods. Here, we also utilize our proposed polygenic methods, which refer to as the iterative screen regression model (ISR) for genome prediction. We compared ISR with several commonly used prediction methods with simulations. We further applied ISR to predicting 15 traits, including the five species of cattle, rice, wheat, maize, and mice. The results of the study indicate that the ISR method performs well than several commonly used polygenic methods and stability.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Polygenic Prediction of Complex Traits with Iterative Screen Regression Models

Luo

Gu²

2020

Preprint

View full text Add to dashboard Cite

show abstract

“…The prediction accuracy of an optimal marker set depends on how well it reflects the characteristics of the markers involved in a phenotype; thus, it is important to construct a marker-set with appropriate markers 14 . In this respect, many studies have adopted approaches that either directly exclude uninformative markers or assign weights to markers according to their contributions in a large set of markers [31][32][33] . These approaches have contributed to improving the accuracy of the www.nature.com/scientificreports/ GP, but simultaneously, it is difficult to select the appropriate markers to be excluded or the weight values to be assigned.…”

Section: Discussionmentioning

confidence: 99%

“…These approaches have contributed to improving the accuracy of the www.nature.com/scientificreports/ GP, but simultaneously, it is difficult to select the appropriate markers to be excluded or the weight values to be assigned. In particular, when these approaches are based on GWAS, obtaining robust weights is problematic due to marker effects or p-values being calculated differently according to the GWAS methods 31 . Moreover, these approaches often require an amount of computation if they conduct modeling based on the large marker set.…”

Section: Discussionmentioning

confidence: 99%

GMStool: GWAS-based marker selection tool for genomic prediction from genomic data

Jeong

Kim

2020

Sci Rep

View full text Add to dashboard Cite

The increased accessibility to genomic data in recent years has laid the foundation for studies to predict various phenotypes of organisms based on the genome. Genomic prediction collectively refers to these studies, and it estimates an individual’s phenotypes mainly using single nucleotide polymorphism markers. Typically, the accuracy of these genomic prediction studies is highly dependent on the markers used; however, in practice, choosing optimal markers with high accuracy for the phenotype to be used is a challenging task. Therefore, we present a new tool called GMStool for selecting optimal marker sets and predicting quantitative phenotypes. The GMStool is based on a genome-wide association study (GWAS) and heuristically searches for optimal markers using statistical and machine-learning methods. The GMStool performs the genomic prediction using statistical and machine/deep-learning models and presents the best prediction model with the optimal marker-set. For the evaluation, the GMStool was tested on real datasets with four phenotypes. The prediction results showed higher performance than using the entire markers or the GWAS-top markers, which have been used frequently in prediction studies. Although the GMStool has several limitations, it is expected to contribute to various studies for predicting quantitative phenotypes. The GMStool written in R is available at www.github.com/JaeYoonKim72/GMStool.

show abstract

“…Because the dataset that we used was not the original data, all the phenotype data had been standardized (mean=0, standard deviation=1). More details about the original dataset can refer to effect loci and many small effect loci (MY) and (3) many loci with small effects (SCS), respectively [21,22].…”

Section: German Holstein Dairy Cattle Datasetmentioning

confidence: 99%

A stacking ensemble learning framework for genomic prediction

Liang

Chang

et al. 2020

Preprint

View full text Add to dashboard Cite

Background: Machine learning (ML) is perhaps the most useful for the interpretation of large genomic datasets. However, the performance of a single machine learning method in genomic selection (GS) was unsatisfactory in existing research. To improve the genomic predictions, we constructed a stacking ensemble learning framework (SELF) integrated three machine learning methods to predict genomic estimated breeding values (GEBVs). Results: We evaluated the prediction ability of SELF by three real datasets and compared the prediction accuracy of SELF, base learners, GBLUP and BayesB. For each trait, SELF performed better than base learners, which included support vector regression (SVR), kernel ridge regression (KRR) and elastic net (ENET). The prediction accuracy of SELF had an average 7.70% improvement compared with GBLUP in three datasets. Except for the milk fat percentage (MFP) traits of the German Holstein dairy cattle dataset, SELF more robust than BayesB in the remaining traits.Conclusions: In this study, we utilized a stacking ensemble learning framework (SELF) to genomic prediction and it performed much better than GBLUP and BayesB in three real datasets with different genetic architecture. Therefore, we believed SEFL had the potential to be promoted to estimate GEBVs in other animals and plants.

show abstract

KAML: improving genomic prediction accuracy of complex traits using machine learning determined parameters

Cited by 65 publications

References 53 publications

Polygenic Prediction of Complex Traits with Iterative Screen Regression Models

Polygenic Prediction of Complex Traits with Iterative Screen Regression Models

GMStool: GWAS-based marker selection tool for genomic prediction from genomic data

A stacking ensemble learning framework for genomic prediction

Contact Info

Product

Resources

About