Introduction
This app is wrapper of several method for multivariable analysis, aim to do clustering and dimention reduction.
Expression table has to be long table, with column as samples, and row as variables
Grouping information is requied in meta table with at least two colums, first as sample names, second as groups.
Expression table has to be long table, with column as samples, and row as variables
Grouping information is requied in meta table with at least two colums, first as sample names, second as groups.
Upload expression data (Long table)
Expression matrix (long table)
Upload meta file (required)
Meta data
NOTE
Note that this pre-process is optionalOriginal Data
Processed Data
Note that the core function behind IS prcomp, NOT princomp
'princomp' can only be used with more sampless than variables/features
Here we use PLS-DA as a dimention reduction and feature selection method
PLS-DA can be thought of as a 'supervised' version of Principal Component Analysis (PCA) in the sense that it achieves dimensionality reduction but with full awareness of the class labels
Besides its use as for dimensionality-reduction, it can be adapted to be used for feature selection [5] as well as for classification
Since PLS-DA is prone to the problem of overfitting, cross-validation is an important step in using PLS-DA as a feature selector and a classifier
PLS-DA can be thought of as a 'supervised' version of Principal Component Analysis (PCA) in the sense that it achieves dimensionality reduction but with full awareness of the class labels
Besides its use as for dimensionality-reduction, it can be adapted to be used for feature selection [5] as well as for classification
Since PLS-DA is prone to the problem of overfitting, cross-validation is an important step in using PLS-DA as a feature selector and a classifier
Refereneces:
http://mixomics.org/methods/pls-da/
http://mixomics.org/methods/pls-da/
For Exploring mode, the method doing more assessment to decide the numbers of pPC first.
The batch effect (before and after correction using two method) can be quantified and visualized by the box plot.
Download the data from the download tab.
PCA plot is usually used to visual check if there is any obivous batch effects.
Two methods, ComBat and CorrrectBatch (based on probability) are used to do the correction.
You can clearly tell the power of the correction by comparing the PCA plot before and after.
Download the data from the download tab.
You can also check the proportion of PCA variations accross all PCs from PCA anlaysis.
There is for sure some variation beteen the two method correction, but usually highly consistent to earch other.
The batch effect on each pairs of the pPCs using PPCA analysis for the optimized number of pPCs
The batch effect on each pairs of the PCs using PCA analysis for the top 10 PCs
busy, won't be long ...

Analyzing, won't be long ...

Contact
Acknowledgement
This shiny apps uses:
DT, data.table: to table manipulation and display
shinyBS: for progressing status.
plotly: to make ggplot2 plot more interactive
ggplot2: for a lot of plotting
Of course, shiny, rstudio, htmlwidgets, and many more
ChangeLog
V1.0: 20180301, major update, with modules
V0.2: 20170915, re-arrange UI, adding more options
V0.1: 20170823, functional version online
V0.2: 20170915, re-arrange UI, adding more options
V0.1: 20170823, functional version online