Research Article
BibTex RIS Cite

Genomik Veri Setlerinin LASSO ve Elastik Net Regresyon Yöntemleri ile Analizi

Year 2022, Volume: 13 Issue: 3, 485 - 496, 20.12.2022
https://doi.org/10.22312/sdusbed.1201417

Abstract

Amaç: Bu çalışmanın amacı büyük boyutlu genomik veri setlerinin değişken seçim yöntemleri kullanılarak daha küçük boyutlara indirgenip daha az maliyet ve zaman ile analizlerin gerçekleştirilebileceğini göstermektir.
Gereç ve Yöntem: Bu çalışmada NCBI veri tabanından Bioconductor yardımı ile R programına aktarılan GDS4906 numaralı veri seti kullanılmıştır. Veri seti 10-katlı çapraz doğrulama ile LASSO ve Elastik Net regresyon yöntemleri kullanılarak analiz edilmiştir.
Bulgular: Veri seti LASSO regresyon yöntemi ile analiz edildiğinde veri setinden 5 adet gen seçilmiş olup, sonrasında farklı iterasyonlarda seçilen değişkenler ve değişken sayılarında farklılık gözlendiğinden kararlılık seçimi yöntemi uygulanarak 2 adet gen seçilmiş ve modelin R2 değeri 0,85 olarak bulunmuştur. Aralıklı arama yöntemi kullanılarak uygulanan Elastik Net regresyon yönteminde 19 adet gen seçilmiş ve R2 değeri 0,92 olarak bulunmuştur.
Sonuç: Elde edilen sonuçlara göre LASSO ve Elastik Net regresyon yöntemlerinin genomik veri setlerinde iyi bir performans gösterdiği anlaşılmıştır.

References

  • [1] Khuri, A. I. 2013. Introduction to Linear Regression Analysis. 4th edition by Douglas C. Montgomery, Elizabeth A. Peck, G. Geoffrey Vining. International Statistical Review.
  • [2] Pripp, A. H., Stanišić, M. 2017. Association between biomarkers and clinical characteristics in chronic subdural hematoma patients assessed with lasso regression. PLoS ONE 12(11).
  • [3] Kohannim, O., et al. 2012. Discovery and Replication of Gene Influences on Brain Structure Using LASSO Regression. Front Neurosci, 6(115).
  • [4] Çiftsüren, N. M., Akkol, S. 2018. Prediction of internal egg quality characteristics and variable selection using regularization methods: ridge, LASSO and elastic net. Archives Animal Breeding, 61(3), 279-284.
  • [5] Cho, S., Kim, H., Oh, S., Kim, K., Park, T. 2009. Elastic-net regularization approaches for genome-wide association studies of rheumatoid arthritis. BMC Proc., 3(7), 25.
  • [6] KOAH Veri seti. 2013. NCBI, National Center for Biotechnology Information. https://www.ncbi.nlm.nih.gov/sites/GDSbrowser?acc=GDS4906 (Erişim Tarihi: 10.01.2022).
  • [7] Tibshirani, R. 1996. Regression Shrinkage and Selection via the Lasso. Journal of the Royal Statistical Society: Series B (Statistical Methodology),58(1), 267-288.
  • [8] Zou, H., Hastie, T. 2005. Regularization and Variable Selection via the Elastic Net. Journal of the Royal Statistical Society: Series B (Statistical Methodology),67(2), 301-320.
  • [9] Segal, M., Dahlquist, K., Conklin, B. 2003. Regression approach for microarray data analysis. J Computnl Biol., 10(6), 961–980.
  • [10] Meinhausen, N., Bühlmann, P. 2010. Stability selection. Journal of the Royal Statistical Society: Series B (Statistical Methodology),72(4), 417-473.
  • [11] Sill, M., Hielscher, T., Becker, N., Zucknick, M. 2014. c060: Extended Inference with Lasso and Elastic-Net Regularized Cox and Generalized Linear Models. Journal of Statistical Software, 62(5), 1-22.
  • [12] Becker, N., Werft, W., Benner, A. 2012. Benner A. penalizedSVM: Feature Selection SVM Using Penalty Functions. R package version 1.1. http://CRAN.R-project.org/package= penalizedSVM. (Erişim Tarihi: 11.10.2022).
  • [13] Froehlich, H., Zell, A. 2005. Efficient Parameter Selection for Support Vector Machines in Classification and Regression via Model-Based Global Optimization. In Proceedings of the International Joint Conference of Neural Networks., 31 Temmuz-4 Ağustos, Canada.
  • [14] Bioconductor. https://bioconductor.org (Erişim Tarihi: 07.11.2022).

Analysis of Genomic Data Sets by LASSO and Elastic Net Regression Methods

Year 2022, Volume: 13 Issue: 3, 485 - 496, 20.12.2022
https://doi.org/10.22312/sdusbed.1201417

Abstract

Objective: The purpose of this study is to show that large-sized genomic datasets can be reduced to smaller sizes using variable selection methods, and that analysis can be performed with less cost and time.

Materials and methods: This study uses dataset number GDS4906, which is transferred from the NCBI database to the R program using Bioconductor. The dataset was analyzed using LASSO and Elastic Net regression methods with 10-fold cross-validation.

Results: When the dataset is analyzed using the LASSO regression method, 5 genes were selected from the dataset and 2 genes were selected and the R2 values of the model were found as 0.85 by applying the determination selection method, as the variables and variable numbers selected in different iterations were then different. In the Elastic Net regression method applied using the interval search method, 19 genes were selected and R2 were found as 0.92.

Conclusion: According to the results obtained, LASSO and Elastic Net regression methods have shown a good performance in the genomic datasets.

References

  • [1] Khuri, A. I. 2013. Introduction to Linear Regression Analysis. 4th edition by Douglas C. Montgomery, Elizabeth A. Peck, G. Geoffrey Vining. International Statistical Review.
  • [2] Pripp, A. H., Stanišić, M. 2017. Association between biomarkers and clinical characteristics in chronic subdural hematoma patients assessed with lasso regression. PLoS ONE 12(11).
  • [3] Kohannim, O., et al. 2012. Discovery and Replication of Gene Influences on Brain Structure Using LASSO Regression. Front Neurosci, 6(115).
  • [4] Çiftsüren, N. M., Akkol, S. 2018. Prediction of internal egg quality characteristics and variable selection using regularization methods: ridge, LASSO and elastic net. Archives Animal Breeding, 61(3), 279-284.
  • [5] Cho, S., Kim, H., Oh, S., Kim, K., Park, T. 2009. Elastic-net regularization approaches for genome-wide association studies of rheumatoid arthritis. BMC Proc., 3(7), 25.
  • [6] KOAH Veri seti. 2013. NCBI, National Center for Biotechnology Information. https://www.ncbi.nlm.nih.gov/sites/GDSbrowser?acc=GDS4906 (Erişim Tarihi: 10.01.2022).
  • [7] Tibshirani, R. 1996. Regression Shrinkage and Selection via the Lasso. Journal of the Royal Statistical Society: Series B (Statistical Methodology),58(1), 267-288.
  • [8] Zou, H., Hastie, T. 2005. Regularization and Variable Selection via the Elastic Net. Journal of the Royal Statistical Society: Series B (Statistical Methodology),67(2), 301-320.
  • [9] Segal, M., Dahlquist, K., Conklin, B. 2003. Regression approach for microarray data analysis. J Computnl Biol., 10(6), 961–980.
  • [10] Meinhausen, N., Bühlmann, P. 2010. Stability selection. Journal of the Royal Statistical Society: Series B (Statistical Methodology),72(4), 417-473.
  • [11] Sill, M., Hielscher, T., Becker, N., Zucknick, M. 2014. c060: Extended Inference with Lasso and Elastic-Net Regularized Cox and Generalized Linear Models. Journal of Statistical Software, 62(5), 1-22.
  • [12] Becker, N., Werft, W., Benner, A. 2012. Benner A. penalizedSVM: Feature Selection SVM Using Penalty Functions. R package version 1.1. http://CRAN.R-project.org/package= penalizedSVM. (Erişim Tarihi: 11.10.2022).
  • [13] Froehlich, H., Zell, A. 2005. Efficient Parameter Selection for Support Vector Machines in Classification and Regression via Model-Based Global Optimization. In Proceedings of the International Joint Conference of Neural Networks., 31 Temmuz-4 Ağustos, Canada.
  • [14] Bioconductor. https://bioconductor.org (Erişim Tarihi: 07.11.2022).
There are 14 citations in total.

Details

Primary Language Turkish
Subjects Health Care Administration
Journal Section Araştırma Articlesi
Authors

Merve Vergili 0000-0002-5570-422X

Hikmet Orhan 0000-0002-8389-1069

Publication Date December 20, 2022
Submission Date November 10, 2022
Published in Issue Year 2022 Volume: 13 Issue: 3

Cite

Vancouver Vergili M, Orhan H. Genomik Veri Setlerinin LASSO ve Elastik Net Regresyon Yöntemleri ile Analizi. Süleyman Demirel Üniversitesi Sağlık Bilimleri Dergisi. 2022;13(3):485-96.

SDÜ Sağlık Bilimleri Dergisi, makalenin gönderilmesi ve yayınlanması dahil olmak üzere hiçbir aşamada herhangi bir ücret talep etmemektedir. Dergimiz, bilimsel araştırmaları okuyucuya ücretsiz sunmanın bilginin küresel paylaşımını artıracağı ilkesini benimseyerek, içeriğine anında açık erişim sağlamaktadır.