Tumor classification by partial least squares using microarray gene expression data

By Danh V. Nguyen, David M. Rocke. Bioinform., 2002

878 citations recorded by the source corpus.

MOTIVATION One important application of gene expression microarray data is classification of samples into categories, such as the type of tumor. The use of microarrays allows simultaneous monitoring of thousands of genes expressions per sample. This ability to measure gene expression en masse has resulted in data with the number of variables p(genes) far exceeding the number of samples N. Standard statistical methodologies in classification and prediction do not work well or even at all when N < p. Modification of existing statistical methodologies or development of new methodologies is needed for the analysis of microarray data. RESULTS We propose a novel analysis procedure for classifying (predicting) human tumor samples based on microarray gene expressions. This procedure involves dimension reduction using Partial Least Squares (PLS) and classification using Logistic Discrimination (LD) and Quadratic Discriminant Analysis (QDA). We compare PLS to the well known dimension reduction method of Principal Components Analysis (PCA). Under many circumstances PLS proves superior; we illustrate a condition when PCA particularly fails to predict well relative to PLS. The proposed methods were applied to five different microarray data sets involving various human tumor samples: (1) normal versus ovarian tumor; (2) Acute Myeloid Leukemia (AML) versus Acute Lymphoblastic Leukemia (ALL); (3) Diffuse Large B-cell Lymphoma (DLBCLL) versus B-cell Chronic Lymphocytic Leukemia (BCLL); (4) normal versus colon tumor; and (5) Non-Small-Cell-Lung-Carcinoma (NSCLC) versus renal samples. Stability of classification results and methods were further assessed by re-randomization studies.

Subfields: Genetics, Molecular Biology

View Paper: https://www.semanticscholar.org/paper/c42192af3f3f81770b85b7f14170fb2abcce17d7

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Genetics

In Molecular Biology

Citations in the corpus

Prediction by Supervised Principal Components (2006, background): “Nguyen and Rocke (2002) and Hi and Gui (2004) discussed partial least squares (PLS) approaches to survival prediction from microarray data.”
Genetic programming for simultaneous feature selection and classifier design (2006, methodology): “Nguyen et al. [ 51 ], [45] achieved 97.1% (best test) accuracy using the logistic discriminant classifier.”
So you think you can PLS-DA? (2017, methodology): “Besides its use as for dimensionality-reduction, it can be adapted to be used for feature selection [5] as well as for classification [12], [13], [16], [3].”
Semi-Supervised Methods to Predict Patient Survival from Gene Expression Data (2004, methodology): “In particular, we examined three previously published procedures: a method based on SVMs (Li and Luan 2003), a method based on PLS (Nguyen and Rocke 2002a), and an ad hoc procedure that calculated a ‘‘ risk index ’’ for each patient by taking an appropriate linear combination of a subset of the gene”
Overview and Recent Advances in Partial Least Squares (2005, background): “The success of PLS in chemometrics resulted in a lot of applications in other scientific areas including bioinformatics, food research, medicine, pharmacology, social sciences, physiology–to name but a few [28, 25, 53, 29, 18, 22].”
Filter versus wrapper gene selection approaches in DNA microarray domains (2004, methodology): “Apart from the FSS approach, other dimensionality reduction procedures appear in supervised microarray literature such as principal component analysis [49] or partial least square [50].”
Interactions between gut microbiota, host genetics and diet relevant to development of metabolic syndromes in mice (2010, methodology): “…and the abundance distributions of variables are extremely uneven, PLSDA is a useful multivariate analysis tool to deal with this type of data (Nguyen and Rocke, 2002; PerezEnciso and Tenenhaus, 2003; Wang et al., 2004; The ISME Journal Figure 4 Phylogeny and differential abundance and…”
Transfer Learning for Molecular Cancer Classification Using Deep Neural Networks (2019, methodology): “The approaches compared were: 1) performing feature reduction with Recursive Feature Elimination with SVM (SVM-RFE) algorithm, Principal Component Analysis (PCA) algorithm, and Sequential Forward Selection (SFS) algorithm independently, followed by 2) classification with Softmax classifier, linear k”
Data-driven modelling of signal-transduction networks (2006, background): “Early proof-of-principle studies showed that different samples could be classified by PLS based on transcriptiona”
Partial least squares: a versatile tool for the analysis of high-dimensional genomic data (2006, neutral): “It also implements the classification method PLSþLDA presented in ‘Classification problem’ section and discussed by Nguyen and Rocke [38, 39] and Boulesteix [40] as well as the ridge PLS method [51] mentioned in ‘PLS and generalized linear models’ section.”
Bayesian model averaging: development of an improved multi-class, gene selection and classification tool for microarray data (2005, methodology): “Examples of univariate methods include the signal-to-noise ratio (Golub et al ., 1999), the t -test (Nguyen and Rocke, 2002b), the ratio of between-groups to within-groups sum of squares (BSS/WSS) (Dudoit et al ., 2002), the signiﬁcance analysis of microarray (SAM) statistic (Tusher et al ., 2001),…”
A mixture model-based approach to the clustering of microarray expression data (2002, background): “…validation, and understanding of various classes and subclasses of cancer; see, for example, Eisen et al. (1998), Ben-Dor et al. (1999) (2000), Alon et al. (1999), Golub et al. (1999), Hastie et al. (2000), Moler et al. (2000), Nguyen and Rocke (2001), and Xing and Karp (2001), among others.”
mixOmics: An R package for ‘omics feature selection and multiple data integration (2017, methodology): “The assessment is data-driven and similar to the process detailed in Rohart et al. (2016), where one-sided t-tests assess whether there is a gain in performance when adding components to the model.”