Aaron Mackey, and have long advocated for the use of statistical consensus methods in bioinformatics
25 January 2016
The authors compare their work to "simple consensus" methods, but unfortunately not to an already existing statistical consensus method, such as BAYSIC (http://www.biomedcentral.com/1471-2105/15/104). Notably, BAYSIC's unsupervised learning method has the advantage that it does not require a training set, and so may be more widely usable than SomaticSeq.
SomaticSeq vs. BAYSIC?
25 January 2016
The authors compare their work to "simple consensus" methods, but unfortunately not to an already existing statistical consensus method, such as BAYSIC (http://www.biomedcentral.com/1471-2105/15/104). Notably, BAYSIC's unsupervised learning method has the advantage that it does not require a training set, and so may be more widely usable than SomaticSeq.Competing interests
I am an author of the BAYSIC paper