What is the Validation Probability?
High-throughput -omics experiments like those based on
microarrays, next generation sequencing, and mass
spectrometry produce lists of significant results. Those
lists are usually defined based on a false discovery
rate (FDR) threshold. Leek JT, Rasgon JL, and Taub MA (2011) "A statistical approach to selecting and confirming validation
targets in -omics experiments" BMC Bioinformatics In Press [preprint] proposed a new statistical approach
for validation. The idea is to validate a random subset
of the significant results, rather than simply those at the
top of the list. Based on the proportion of these results
that validate on an independent technology, it is
possible to calculate the probability that the results
have been validated. The validation probability is a
score that measures the certainty with which
high-throughput results have been validated. Scores
greater than 0.5 are evidence that the validation was
successful, with higher scores indicating greater
confidence in the validation.
How to Use the Calculator
To use the validation calculator:
- Input the false discovery rate used to define the
significant results in the original analysis.
- Input the number of randomly sampled significant
results that were tested for validation with an
independent technology.
- Of the results tested for validation, input the
number that validated.
- Click validate and the validation probabillity is
calculated.
Additional Resources
The methods behind this calculator were published
in Leek JT, Rasgon JL, and Taub MA (2011) "A statistical approach to selecting and confirming validation
targets in -omics experiments" BMC Bioinformatics In Press [preprint]. Please see the
manuscript for further details and examples. R code is available from the R
code page. The R code implements the validation
calculation performed by the online calculator and
includes additional functionality. It will be submitted to Bioconductor shortly.
