Project objective: To provide a user-friendly, web-based analytical pipeline for comparative metagenomic studies. In particular, METAGENassist allows users to take bacterial census data from different environment sites or different biological hosts, and perform comprehensive multivariate statistical analyses on the data. These multivariate analyses can be done using either taxonomic or automatically generated phenotypic labels and visualized using a variety of high quality graphical tools. The bacterial census data can be derived from 16S rRNA data, NextGen shotgun sequencing or even classical microbial culturing techniques.
Data input formats: Users upload a taxonomic profile in one of several supported formats (details). A taxonomic profile file contains the names of microbial species (or other taxonomic classes) and their relative abundance in at least 2 samples; accepted formats include CSV format, or data generated by mothur, QIIME, MG-RAST, MEGAN, and STAMP. Users will also benefit by uploading an optional file containing metadata for each sample (CSV format).
Data processing: METAGENassist performs: 1) taxonomic name normalization; 2) automated taxonomic-to-phenotypic mapping using nearly 20 different phenotypic categories; 3) data integrity/quality checks and 4) data normalization via normalization by constant sum, normalization by a reference feature, sample specific normalization or auto/Pareto/range scaling.
Statistical analysis: METAGENassist offers a wide array of commonly used statistical and machine learning methods including: univariate statistics - fold change analysis, t-tests, volcano plots, one-way ANOVA, correlation analysis;multivariate statistics - principal component analysis (PCA) and partial least squares - discriminant analysis (PLS-DA);clustering - dendrograms, heatmaps, K-means clustering, self organizing feature maps (SOM); supervised classification - random forests and support vector machine (SVM).
Data output: Upon completion, METAGENassist generates a variety of well-annotated tables and colorful, labeled graphs in an anti-aliased PNG format. PDF versions for some plots are also available. The processed data and images are available for download.
Please Cite:
David Arndt, Jianguo Xia, Yifeng Liu, You Zhou, An Chi Guo, Joseph A. Cruz, Igor Sinelnikov, Karen Budwill, Camilla L. Nesbø and David S. Wishart METAGENassist: A comprehensive web server for comparative metagenomics Nucleic Acids Research 2012 Jul;40(Web Server issue):W88-95. Epub 2012 May 29.
0 comments:
POST A COMMENT