Category: News

mixOmics 5.0-2 update

The major changes of this new update is the perf() function that supersedes valid() and offers a variable stability measure across the different folds.

The pls() and spls() functions have been modified and are now following the same framework coding.

See the CRAN page here.

The mixOmics website will be updated shortly for the major changes of these functions. Remember that you can subscribe to our newsletter (mixOmics updates, workshops) as indicated here.

Changes in 5.0-2

 

New features:
————-
– The valid function has been superseded by the perf function. Although similar in essence, few bugs have been fixed to estimate the performance of the sPLS and sPLS-DA models with no selection bias. A variable stability frequency has been added to the output. Functions spls.model and pls.model have been removed.

Bug fixes:
———-
-pls and spls function have been modified and ‘harmonised’ w.r.t to scaling. Loading vectors a and b are now scaled to 1. Latent variables t and u are not scaled (following Table 21 of the Tenenhaus book – which is in French, sorry!).

-the argument abline.line has been set to FALSE by default in all plotIndiv functions.

-the warnings messages in the plot functions have been fixed

– tune.multilevel for one factor has been fixed.

ECCB’14 Tutorial on mixOmics

ECCB Tutorial T04. Multivariate methodologies for the exploration of large biological data sets. Application in R using the mixOmics package

Date: Sunday Sept 7, 2014
Venue: FORUM building, Faculté de Médecine, 4, rue Kirschleger, Strasbourg
Time: 9am – 5.30pm (registration starts from 8am)
Contact: mixomics[at]math.univ-toulouse.fr
More details: http://www.eccb14.org/program/tutorials/mixomics

 How to register: http://www.eccb14.org/registration Registration includes lunch & coffee breaks on the day of the workshop and the tutorial material (.pdf and/or print). ECCB Tutorial rates: 110 € (academic) or 60 € (student)

Description

The objective of this tutorial is to introduce the fundamental concepts behind projection-based approaches and illustrate their application on some exemplar studies using the R package mixOmics.

Multivariate projection approaches are useful exploratory tools to get a first understanding of large and complex data sets. These approaches are extremely efficient on large data sets, and can also answer complex questions. Such approaches include Principal Component Analysis (PCA, Joliffe 2002) and other variants, Partial Least Squares regression (PLS, Wold 2001), PLS-Discriminant Analysis, Canonical Correlation Analysis (CCA, Hotelling 1936). These approaches enable the reduction of the dimension of the data by projecting them into a smaller subspace. Recent developments proposed the so-called `sparse’ approaches, which include Lasso penalisations to allow variable selection (Tibshirani 2001).

PCA is the oldest and most popular multivariate technique but often, little is known about how this approach is solved and what are the limitations. More sophisticated approaches like PLS and CCA have recently been extended to deal with the large dimension (sparse PLS, or regularized CCA) and were proven to bring biologically meaningful results in many studies. Contrary to PCA, PLS and CCA enable the integration of two types of data sets.

Since 2009, we have implemented many multivariate approaches and their sparse variants in the R package mixOmics to be used by the statistical and bioinformatics community. Full tutorials are given on our website: http://perso.math.univ-toulouse.fr/mixomics/

In this tutorial, we will focus on the application of these approaches to medium and high throughput biological data (transcriptomics, metabolomics, proteomics data) using PCA, CCA, PLS, PLS-DA and the variants that the mixOmics team and collaborators have developed.

 

Presenters

The presenters are all key developers of mixOmics.

-Dr Kim-Anh Lê Cao (The University of Queensland Dimantina Insitute, Brisbane, Australia). Kim-Anh is a biostatistician researcher in the University of Queensland, Brisbane, Australia. Her institute has a particular focus on severe and chronic diseases such as cancer and diseases involving the immune system, including arthritis, chronic infections, and diabetes. Together with the mixOmics team, Kim-Anh continues to develop methodologies to analyse complex biological studies.

-Dr Sébastien Déjean (Institut de Mathématiques de Toulouse, Université de Toulouse, France). Sebastien is a statistician research engineer in the Universite de Toulouse. Through his support activities to research, he contribues to various projects particularly in the fields of high throughput biology and information retrieval systems.

-Dr Ignacio González (Institut de Mathématiques de Toulouse, Université de Toulouse, Institut National de la Recherche Agronomique, France). Ignacio is working at the plateforme de bioinformatique et biostatistique de Toulouse. Ignacio has been working in several wet laboratories (INSERM, INRA, CNRS, INSA) where he provided statistical support. He has considerable experience in analyzing a vast range of biological data.

Target Audience

Postgraduate students, postdoctoral fellows and researchers with basic statistical knowledge, in need to

-explore large data sets

-use graphical techniques to better visualize data

– apply multivariate projection methodologies to large data sets.

Prerequisite and requirements

We expect the audience to have a good working knowledge in R. Attendees are requested to bring their own laptops, having installed the software RStudio http://www.rstudio.com/ and the R package mixOmics.

Version 4.1-3 is on CRAN now

Changes in 4.1
================

New features:
-------------
- New S3 method valid for objects of class psl, spls, plsda and splsda
- New select.var function to directly extract the selected variables from spls, spca, sipca
- New data set vac18 for multilevel data

Article published explaining correlation circle plots, relevance networks and CIM

Our manuscript ‘Insightful graphicalt outputs to explore relationships between two “omics” data sets has been published and explains how to interpret Correlation Circle plots, how relevance networks and CIM are generated from rCCA and sPLS.

Check this very colourful manuscript[intlink id=”202″ type=”page”]here[/intlink]!

General presentation about mixOmics

A new general presentation about mixOmics is available (and should be updated for major update of the package) in the [intlink id=”204″ type=”page”]Presentation Section[/intlink].

Lê Cao K.-A. Unravelling `omics’ data with the mixOmics R package, Illustration on several studies. General presentation on mixOmics (last updated 05/04/2012) [Presentation]

(s)IPCA

Independent Principal Component Analysis (IPCA)

In some case studies, we have identified some limitations when using PCA:

  • PCA assumes that gene expression follows a multivariate normal distribution and recent studies have demonstrated that microarray gene expression measurements follow instead a super-Gaussian distribution
  • PCA decomposes the data based on the maximization of its variance. In some cases, the biological question may not be related to the highest variance in the data

Instead, we propose to apply Independent Principal Component Analysis (IPCA) which combines the advantages of both PCA and Independent Component Analysis (ICA). It uses ICA as a denoising process of the loading vectors produced by PCA to better highlight the important biological entities and reveal insightful patterns in the data.

IPCA offers a better visualization of the data than ICA and with a smaller number of components than PCA.

How to choose the number of components:

The kurtosis measure is used to order the loading vectors to order the Independent Principal Components.  We have shown that the kurtosis value is a good post hoc indicator of the number of components to choose, as a sudden drop in the values corresponds to irrelevant dimensions.

Sparse Independent Principal Component Analysis (sIPCA)

Similar to the [intlink id=”129″ type=”page”]sparse PCA[/intlink] version implemented in mixOmics, soft-thresholding is applied in the independent loading vectors in IPCA to perform internal variable selection.

How to choose the number of variables to select:

The number of variables to select is still an open issue. In our paper we proposed to use the Davies Bouldinmeasure which is an index of crisp cluster validity. This index compares the within-cluster scatter with the between-cluster separation.

More details about how to use the ipca.R function in the[intlink id=”233″ type=”page”] case study[/intlink].

References

New methods: multilevel analyses

A multilevel approach has been added for cross-over design experiments (up to two cross factors), in collaboration with A/Prof B. Liquet (Universite de Bordeaux, France).  This approach takes into account the complex structure of repeated measurements from different assays, where different treatments are applied on the same subjects to highlight the treatment effects within subject separately from the biological variation between subject.

Two different frameworks are proposed:

  • a discriminant analysis (method = ‘splsda’) enables the selection of features separating the different treatments
  • a integrative analysis  (method = ‘spls’) enables the interaction of two matched data sets and the selection of subset of correlated variables (positively or negatively) across the samples. The approach is unsupervised: no prior knowledge about the samples groups is included.

The multilevel function first decomposes the variance in the data sets X (and Y) and applies either sPLS-DA or sPLS on the within-subject deviation. One or two-factor analyses are available for sPLS-DA.

Associated functions include: multilevel.R, tune.multilevel.R, pheatmap.multilevel.R (see examples in methods, graphics and case studies).

This is our first step towards repeated measurements designs.

The package has been updated to version 4.0-1 to implement these methodologies. It now requires the library ‘pheatmap’.

Web-interface

  • R package and Methods: IPCA and sparse IPCA functions have been implemented (as well as their associated S3 functions). IPCA stands for Principal Component Analysis with Independent Loadings. It is a combination of the advantages of both PCA and Independent Component Analysis (ICA). PCA is a powerful exploratory tool if the biological question is related to the highest variance. ICA was recently proposed in the literature as an alternative to PCA as it optimizes an independence condition that can give more meaningful components. A preprint can be available upon request.
  • R package and Data: The Liver Toxicity study data has been updated to provide geneBank IDs and gene titles
  • R package and Data: Two other data sets have been added: Prostate Tumor study (gene expression) and Metabolomic study of Yeast (metabolomics).
  • Web interface: We are making good progress on our associated web-interface (now deployed on  http://mixomics.qfab.org). Few illustrative examples are also available, and you can download the illustrative examples and run any type of analysis trough the interface. We are currently developing a ‘next level analysis’ to provide pathway enrichment analyses and give the functional annotation of the selected genes using the iHOP database. Do not hesitate to give us some feedback!
  • webinterface
  • ‘sletter: we now have a newsletter, to subscribe, send an email to mixomics[at]math.univ-toulouse.fr with no subject in the body.