A stochastic approach to count RNA molecules using DNA sequencing methods

Abstract

Estimating the number of congeneric mRNA molecules in a preparation is an essential task in DNA technology and biochemistry. However, DNA sequencing methods can only distinguish molecules with different sequences or of different lengths. Recently, it was shown that it is possible to combine RNA molecules with short DNA molecules (tags) such that, with high probability, any two RNA molecules are combined with tags having different sequences. For this technique, we propose a statistical estimator and a confidence interval to determine the number of mRNA molecules in a preparation. In a second approach, the mRNA molecules are lengthened by attaching a random number of linker oligonucleotides. The original number of mRNA molecules is then determined by the number of different lengths obtained from the experiment. We also give estimator and confidence interval for this method. Both methods can be implemented using recent as well as established methods from DNA technology. The methods can also be applied to a larger number of molecules without the need to exhaust the complete preparation. The computation of estimators and confidence intervals can be accomplished by dynamic programming. © Springer-Verlag Berlin Heidelberg 2003.