User menu

Accès à distance ? S'identifier sur le proxy UCLouvain

Combining strong sparsity and competitive predictive power with the L-sOPLS approach for biomarker discovery in metabolomics

Primary tabs

Feraud, Baptiste [UCL] Munaut, Carine Martin, Manon [UCL] Verleysen, Michel [UCL] Govaerts, Bernadette [UCL]

Introduction In the context of metabolomics analyses, partial least squares (PLS) represents the standard tool to perform regression and classification. OPLS, the Orthogonal extension of PLS which has proved to be very useful when interpretation is the main issue, is a more recent way to decompose the PLS solution into predictive components correlated to the target Y and components pertaining to the data X but uncorrelated to Y. This predominance of (O)PLS can raise the question of the awareness of alternative multivariate regression and/or classification tools able to find biomarkers. Actually, the search for biomarkers remains a key issue in metabolomics as it is crucial to very accurately target discriminating features. Objective Most of the time, (O)PLS methods perform well but a drawback often occurs: too many variables can be selected as potential biomarkers even using adapted statistical significance tests. However, for final users (in medical studies for instance), it can be advantageous to deal with only a small number of easily interpretable biomarkers. Methods This drawback is approached in this paper via the use of sparse methods. The sparse-PLS (sPLS), an extension of PLS which promotes an inner variable/feature selection, is an interesting existing solution. But a new intuitive algorithm is proposed in this paper to combine sparsity and the advantages of an orthogonalization step: the “Light-sparse-OPLS” (L-sOPLS). L-sOPLS promotes sparsity on a previously optimized deflated matrix which implies the removal of the Y-orthogonal components. Results A discussion around the compromise between sparsity and predictive modelling performances is provided and it is shown that L-sOPLS produces convincing results, illustrated principally on the basis of 1H-NMR spectral data but also on genomic RT-qPCR data. Conclusion The L-sOPLS algorithm allows to reach better predictive performances than (O)PLS and sPLS while taking into account only a very small number of relevant descriptors.

Document type Article de périodique (Journal article) – Article de recherche
Access type Accès restreint
Publication date 2017
Language Anglais
Journal information "Metabolomics" - Vol. 13, no. 130, p. 15 (2017)
Peer reviewed yes
Publisher Springer New York LLC (New York)
issn 1573-3882
e-issn 1573-3890
Publication status Publié
Affiliations UCL - SSH/LIDAM/ISBA - Institut de Statistique, Biostatistique et Sciences Actuarielles
UCL - SST/ICTM/ELEN - Pôle en ingénierie électrique
Keywords Biomarker discovery ; (O)PLS models ; Feature selection ; Sparse models ; L-sOPLS · 1H-NMR data ; RT-qPCR data
Links
  1. Abdi Hervé, Partial least squares regression and projection on latent structure regression (PLS Regression), 10.1002/wics.51
  2. Afanador N.L., Tran T.N., Buydens L.M.C., Use of the bootstrap and permutation methods for a more robust variable importance in the projection metric for partial least squares regression, 10.1016/j.aca.2013.01.004
  3. Barker Matthew, Rayens William, Partial least squares for discrimination, 10.1002/cem.785
  4. Bartel David P., MicroRNAs: Target Recognition and Regulatory Functions, 10.1016/j.cell.2009.01.002
  5. Bylesjö Max, Rantalainen Mattias, Cloarec Olivier, Nicholson Jeremy K., Holmes Elaine, Trygg Johan, OPLS discriminant analysis: combining the strengths of PLS-DA and SIMCA classification, 10.1002/cem.1006
  6. Chapman Andrew, Saad Yousef, Deflated and Augmented Krylov Subspace Techniques, 10.1002/(sici)1099-1506(199701/02)4:1<43::aid-nla99>3.0.co;2-z
  7. Chun, H., & Keles, S. (2007). Sparse partial least squares regression with an application to genome scale transcription factor analysis. Madison: Department of Statistics, University of Wisconsin.
  8. Chung, D., Chun, H., & Keles, S. (2012). Spls: Sparse partial least squares (SPLS) regression and classification. R package, version, 2, 1–1.
  9. de Jong Sijmen, SIMPLS: An alternative approach to partial least squares regression, 10.1016/0169-7439(93)85002-x
  10. Efron, B., Hastie, T., Johnstone, I., & Tibshirani, R. (2004). Least angle regression. Annals of Statistics, 32(2), 407499.
  11. Féraud Baptiste, Govaerts Bernadette, Verleysen Michel, de Tullio Pascal, Statistical treatment of 2D NMR COSY spectra in metabolomics: data preparation, clustering-based evaluation of the Metabolomic Informative Content and comparison with 1H-NMR, 10.1007/s11306-015-0830-7
  12. Friedman J., Hastie T., & Tibshirani R. (2010). A note on the group lasso and a sparse group lasso, arXiv preprint arXiv:1001.0736 .
  13. Gabrielsson Jon, Jonsson Hans, Airiau Christian, Schmidt Bernd, Escott Richard, Trygg Johan, OPLS methodology for analysis of pre-processing effects on spectroscopic data, 10.1016/j.chemolab.2006.03.013
  14. Geladi Paul, Kowalski Bruce R., Partial least-squares regression: a tutorial, 10.1016/0003-2670(86)80028-9
  15. Giudice Linda C, Kao Lee C, Endometriosis, 10.1016/s0140-6736(04)17403-5
  16. Hastie, T., Tibshirani, R., & Wainwright, M. (2015). Statistical learning with sparsity: The lasso and generalizations. Boca Raton: CRC Press.
  17. Höskuldsson Agnar, PLS regression methods, 10.1002/cem.1180020306
  18. Indahl Ulf G., Liland Kristian Hovde, Naes Tormod, Canonical partial least squares-a unified PLS approach to classification and regression problems, 10.1002/cem.1243
  19. Jung Youngae, Lee Jueun, Kwon Joseph, Lee Kwang-Sik, Ryu Do Hyun, Hwang Geum-Sook, Discrimination of the Geographical Origin of Beef by1H NMR-Based Metabolomics, 10.1021/jf102194t
  20. Lai Eric C., Micro RNAs are complementary to 3′ UTR sequence motifs that mediate negative post-transcriptional regulation, 10.1038/ng865
  21. Lê Cao Kim-Anh, Rossouw Debra, Robert-Granié Christèle, Besse Philippe, A Sparse PLS for Variable Selection when Integrating Omics Data, 10.2202/1544-6115.1390
  22. Lu Bo, Castillo Ivan, Chiang Leo, Edgar Thomas F., Industrial PLS model variable selection using moving window variable importance in projection, 10.1016/j.chemolab.2014.03.020
  23. Mevik Bj�rn-Helge, Cederkvist Henrik Ren�, Mean squared error of prediction (MSEP) estimates for principal component regression (PCR) and partial least squares regression (PLSR), 10.1002/cem.887
  24. Muñoz-Romero Sergio, Arenas-García Jerónimo, Gómez-Verdejo Vanessa, Sparse and kernel OPLS feature extraction based on eigenvalue problem solving, 10.1016/j.patcog.2014.12.002
  25. Nisenblat Vicki, Bossuyt Patrick MM, Shaikh Rabia, Farquhar Cindy, Jordan Vanessa, Scheffers Carola S, Mol Ben Willem J, Johnson Neil, Hull M Louise, Blood biomarkers for the non-invasive diagnosis of endometriosis, 10.1002/14651858.cd012179
  26. Rousseau, R. (2011). Statistical contribution to the analysis of metabonomic data in $${}^1$$ 1 H-NMR spectroscopy (Doctoral dissertation, Université Catholique de Louvain, Belgium), permalink: http://hdl.handle.net/2078.1/75532 .
  27. Stenlund Hans, Gorzsás András, Persson Per, Sundberg Björn, Trygg Johan, Orthogonal Projections to Latent Structures Discriminant Analysis Modeling on in Situ FT-IR Spectral Imaging of Liver Tissue for Identifying Sources of Variability, 10.1021/ac8005318
  28. Tapp Henri S., Kemsley E. Kate, Notes on the practical utility of OPLS, 10.1016/j.trac.2009.08.006
  29. Trygg Johan, Wold Svante, Orthogonal projections to latent structures (O-PLS), 10.1002/cem.695
  30. van Gerven, M. A. J., & Heskes, T. (2010). Sparse orthonormalized partial least squares. In Benelux conference on artificial intelligence.
  31. Wehrens Ron, Chemometrics with R, ISBN:9783642178405, 10.1007/978-3-642-17841-2
  32. Weljie Aalim M., Bondareva Alla, Zang Ping, Jirik Frank R., 1H NMR metabolomics identification of markers of hypoxia-induced metabolic shifts in a breast cancer model system, 10.1007/s10858-011-9486-4
  33. Wiklund Susanne, Johansson Erik, Sjöström Lina, Mellerowicz Ewa J., Edlund Ulf, Shockcor John P., Gottfries Johan, Moritz Thomas, Trygg Johan, Visualization of GC/TOF-MS-Based Metabolomics Data for Identification of Biochemically Interesting Compounds Using OPLS Class Models, 10.1021/ac0713510
  34. Wold, H. (1975). Path models with latent variables: The NIPALS approach (pp. 307–357). New York: Academic Press.
  35. Wold Svante, Trygg Johan, Berglund Anders, Antti Henrik, Some recent developments in PLS modeling, 10.1016/s0169-7439(01)00156-3
  36. Wold Svante, Sjöström Michael, Eriksson Lennart, PLS-regression: a basic tool of chemometrics, 10.1016/s0169-7439(01)00155-1
  37. Zou Hui, Hastie Trevor, Regularization and variable selection via the elastic net, 10.1111/j.1467-9868.2005.00503.x
Bibliographic reference Feraud, Baptiste ; Munaut, Carine ; Martin, Manon ; Verleysen, Michel ; Govaerts, Bernadette. Combining strong sparsity and competitive predictive power with the L-sOPLS approach for biomarker discovery in metabolomics. In: Metabolomics, Vol. 13, no. 130, p. 15 (2017)
Permanent URL http://hdl.handle.net/2078.1/189217