Comparative Evaluation of Two Serial Gene Expression Experiments

Stuart G. Baker, 2014

Introduction

This program fits biologically relevant response curves in comparative analysis of the two gene expression experiments involving same genes but under different scenarios and at least 12 responses. The program outputs gene pairs with biologically relevant response curve shapes including flat, linear, sigmoid, hockey stick, impulse and step curves.

Reference

Baker, S.G. Comparative Analysis of Biologically Relevant Response Curves in Gene Expression Experiments: Heteromorphy, Heterochrony, and Heterometry. Microarrays 2014, 3, 39-51.
https://www.mdpi.com/2076-3905/3/1/39

Requirement

Mathematica Version 8 or later.

To run example in manuscript

copy all files into some folder called "FOLDER"
start a new Mathematica session
type SetDirectory["FOLDER"]
type << mfit.m
type MFit[datasetm1,datasetm2,NewFitQ->True]

To try on your own data,

type MFit[dataset1,dataset2,options]

Options

NewFitQ TRUE Fits new data or retrieves data from previous fit
NewNameCORE "mfit" Name prepended to data set name for new fit
StoredNameCORE "mfit" Name prepended to data set name for retrieving fit (usually the same as NewNameCORE)
ModelSetQ full+poly full+poly" fits flat, line, sigmoid, double sigmoid, generalized double sigmoid, and polynomial curves of degree 3, 5, 7 (requires 7 fitted points and recommended with at least 14 time points
    full" fits flat, line, sigmoid, double sigmoid, generalized double sigmoid (requires 7 fitted points and recommended with at least 14 time points).
    "reduced" fits flat, line, sigmoid, double sigmoid curves (requires 6 fitted points and recommended with at least 12 time points)
SplitType alternating "alternating" which fits every other point or "custom"
SplitFitSet {1,4,7,10,13} Applies only for SplitType->"custom", ranked points for fitting the model
ShowPlot FALSE Shows plots for gene pairs with biologically relevant curves
ShowReport FALSE Report on gene pairs with biologically relevant curves
ShowSummary TRUE Shows summary results
ShowRPEAIC FALSE Show relationship of RPE to AIC

dataset={data, geneid, genename, times, xname, yname, plotname, dataname},

data matrix of responses, where rows indicate genes and columns indicate times
geneid a list of gene identification numbers corresponding to data rows
genename a list of gene names corresponding to data rows
times list of observation times
xname name of horizontal axis on plot
yname name of vertical axis on plot
plotname name of data set for plot
dataname name of data set for storing files

File Contents

Download All (ZIP, 8 MB)

File name Description
mfit.m main program and calls all programs
mfitpair.m compute results for model pairs
mfitsupport.m support functions
mfitform.m compute heterochrony and heterometry
mfitkey.m model selection
mfitplot.m compute plots
mfitplotoverlay.m overlay plot
mfitplotside.m side-by-side plot
mfitreport.m report lists of genes by model pair
mfitcore.m core fitting program
mfitdiagnostic.m plots RPE vesus AIC
mfitdata.m generate data in correct format from following files
mfitrawdata1.m raw data for frog example 1 (X.laevis)
mfitrawdata2.m raw data for frog example 2 (X.tropicalis)
mfitgenename.m gene names for frog examples

Notes on raw data for example

The files mfitrawdata1.m and mfitrawdata2.m are structured as follows: There 11299 rows correspond to genes. The 126 = 14 x 9 columns are ordered from left to right as 14 times for replicate 1 from sample 1, 14 times for replicate 2 from sample 1, 14 times for replicate 3 from sample 1, 14 times for replicate 1 from sample 2, 14 times for replicate 2 from sample 2, 14 times for replicate 3 from sample 2, 14 times for replicate 1 from sample 3, 14 times for replicate 2 from sample 3, 14 times for replicate 3 from sample 3. The 14 times are 1.5, 5, 7, 9, 13.25, 14.75, 16.25, 18.25, 19.75, 21.75, 24.75, 27.5, 35, 44.5. The file mfitgenename.m is structured as follows. The first 11299 rows correspond to the genes in the raw data (there are extra rows). The 3 columns are row number, gene ID, and gene name (or nothing if no gene name).

Disclaimer

This code is provided "as is", without warranty of any kind, express or implied, including but not limited to the warranties of merchantability, fitness for a particular purpose and noninfringement. In no event shall the NCI or the individual developers be liable for any claim, damages or other liability of any kind. Use of this code by recipient is at recipient's own risk. NCI makes no representations that the use of the code will not infringe any patent or proprietary rights of third parties.

Last updated: March 18, 2014