CHAPTER 8: Bioinformatics and Statistics: Statistical Analysis and Validation
-
Published:10 Jun 2013
-
Series: Drug Discovery
H. C. J. Hoefsloot, in Comprehensive Biomarker Discovery and Validation for Clinical Application, ed. P. Horvatovich and R. Bischoff, The Royal Society of Chemistry, 2013, pp. 226-242.
Download citation file:
In this chapter guidelines are given for the selection of a few biomarker candidates from a large number of compounds with a relative low number of samples. The main concepts concerning the statistical validation of the search for biomarkers are discussed. These complicated methods and concepts are discussed without using difficult mathematics. So the people who are involved in biomarker research but have no statistical or informatics background are able to understand this chapter. Also for the specialist all the concepts to avoid the problem of false‐positives, these are findings that turn out to be wrong, are discussed and practical advice is given. The situation where no biomarkers are found is discussed and suggestions are made as to how to investigate whether this is due to the statistical method or to the data.