The consequences of sample pooling in proteomics: an empirical study.

Pooling of samples in proteomics experiments might help overcome resource constraints when many individuals are analysed. The measured biological variation should be reduced giving increased power to detect treatment differences. Pooling has been advocated in microarray work but there are few tests of its potential in proteomics. In this study, we examine three issues on which the success of the pooling approach might hinge and provide evidence that: (i) the protein expression in a pool matches the mean expression of the individuals making up the pool for the majority of proteins, although for some proteins the pool expression is different; (ii) the biological variance between pools is reduced compared with that between individuals, as predicted in theory, but this reduction is not as large as expected. A practical consequence of this is that power could be reduced; (iii) proteins detectable in individual samples are usually but not always visible when samples are pooled. We conclude that pooling of samples in proteomics work is a valid and potentially valuable procedure but consideration should be given to these issues in experimental design.