University of Cambridge > Talks.cam > MRC Biostatistics Unit Seminars > Signal Identification for Rare and Weak Features: Higher Criticism or False Discovery Rates

Signal Identification for Rare and Weak Features: Higher Criticism or False Discovery Rates

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact Dr Jack Bowden.

Signal identification in large-dimensional settings is a challenging problem in biostatistics, even more so if is the signal is very weak. For this purpose the method of “higher criticism” (HC) has recently been powerfully advocated and shown to be highly effective for determining appropriate decision thresholds. On the other hand, many omics studies now routinely employ a variant of the “false discovery rate” (FDR) for screening for interesting features. In my talk, I will compare signal identification using the HC and FDR approaches, with the aim to provide a better understanding of HC as well as offering a simple explanation for HC’s favorable performance. After a non-technical introduction to both the HC and FDR approaches, I will derive the ideal thresholds corresponding to the HC and FDR criteria and explore their mutual relationships. Subsequently, it can be established that in a rare-weak setting there is a near identity between the HC and a simple FDR threshold. As a result, the decision threshold obtained by the frequentist HC approach is given a natural Bayesian interpretation. This is joint work with Bernd Klaus, EMBL Heidelberg.

This talk is part of the MRC Biostatistics Unit Seminars series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

© 2006-2020 Talks.cam, University of Cambridge. Contact Us | Help and Documentation | Privacy and Publicity