HTSmix represents the structure of signals in the screen using linear mixed models, normalizes and summarizes the phenotypes to make them comparable across samples, and outputs a list of hits while controlling the False Discovery Rate. The methodology is appropriate for experimental designs with at least two control samples profiled throughout the screen.
Laboratory for Statistical Proteomics and Bioinformatics , Purdue University
- Linux/ MacOsX/ Windows
- R package
:: MORE INFORMATION
D. Yu, J. Danku, I. Baxter, S. Kim, O. K. Vatamaniuk, D. E. Salt, O. Vitek.
“Noise reduction in genome-wide perturbation screens using linear mixed-effect models”,
Bioinformatics, 27, p. 2173-2180, 2011.