Piecewise-linear machine-learning model for motif-based prediction of transcription factor binding sites
This repository contains the MATLAB code for a piecewise-linear, data-fusion model for motif-based prediction of transcription factor binding sites. The code was used in the analysis published in the 2010 article Genome-wide histone acetylation data improve prediction of mammalian transcription factor binding sites (Ramsey et al., Bioinform., 2010).
The MATLAB code in this project was written by Stephen Ramsey (@saramsey) except for six contributed modules:
readfastaseqs.m: Harri Lähdesmäki (personal communication)
writebar.m: Matti Nykter (personal communication)
The contributions of Dr. Lähdesmäki and Dr. Nykter to this project are gratefully acknowledged.
This software requires the MATLAB software Stable Noisy Optimization by Branch and FIT (SNOBFIT) by Waltraud Huyer and Arnold Neumaier and which is freely available for download from Universitat Wien (download SNOBFIT)[http://www.mat.univie.ac.at/~neum/software/snobfit/].
This software also requires the MATLAB Statistics Toolbox.
uuidgenfrom MATLAB Central (download uuidgen)[https://www.mathworks.com/matlabcentral/fileexchange/21709-uuid-generation?focused=5103481&tab=function].