Abstract

BACKGROUND:

This study was undertaken because of the accidental observation that a sample of 60+ beads was obtained by the instrument from a completely dry, unused well in a 96 well plate. Others have observed unexplained outliers in replicated wells. The problem was first observed on an older instrument, and replicated on a new instrument.

METHODS AND RESULTS:

Data is presented from two instruments using a multiple blank following well experiment that shows a surprising amount of carryover that has an unexpected nature. When it occurs, it does not necessarily decline from one well to the next. There appears to be two types of carryover, one that is small, predictable and declines consistently, and another which is potentially very large, unpredictable, and does not decline. The former can be compensated for or ignored. The latter cannot be addressed without using multiple replicated samples or an intraplex method.

CONCLUSION:

This problem has significance for analysis of results obtained with suspended microarray instruments. A special notation is made that biostatisticians need to be made aware of these results before experiments are undertaken and data generated for them to analyze. The problem can be handled by enough replicated samples, or an intraplex method. The applicability of these results to oligonucleotide based assays is unknown.