Informative signals in microchip-based electrophoresis: mathematical models and estimates

The paper considers possible lines of developing algorithms for processing and estimating analytical signals of microchip electrophoretic devices: simulation of informative signals (peaks) and non-informative components (base line drift and high-frequency noise). Based on the information obtained, sources of interference can be revealed, the optimal algorithms for the signal estimation have been found, and criteria for neighboring peaks separation have been modified. Efficiency of conventional economical methods for signal pre-processing, namely, trend compensation and a filtration, as well as validity of the new technique for peak area estimation and applicability of the main component method in multi-channel measurements when the peak boundaries can hardly be localized, have been shown for electrophoregramms obtained by separating the DNA fragment mix on a microfluidic chip.