Analysis of Coulter counting data from nanopores using clustering

Nanopores with diameters of 105nm have been used to detect 60nm diameter beads by Coulter counting. The Coulter counting is performed by applying a constant potential across the nanopore while recording the drop in ionic current upon passage of a nanobead. In this paper, we analyze the use of the amplitude of the current drop and its duration to infer useful information about the nanobead(s) passing through the pore. Analysis of the shape and duration of these current drops enables us to estimate the size distribution of the nanobeads. Current drops corresponding to complete passage of a nanobead through the pore is called an event. The proposed approach uses spectral clustering to cluster the drops in the feature domain as the shape of the clusters is unknown. Spectral clustering accurately finds the natural clusters and the outlying cluster. The outlying cluster is further clustered to find the categories of the events within it.