Intelligent identification of multi-level nanopore signatures for accurate detection of cancer biomarkers

Abstract

To achieve accurate detection of cancer biomarkers with nanopore sensors, the precise recognition of multi-level current blockage events (signature) is a pivotal problem. However, it remains rather a challenge to identify the multi-level current blockages of target biomarkers in nanopore experiments, especially for the nanopore analysis of serum samples. In this work, we combined a modified DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm with the Viterbi training algorithm of the hidden Markov model (HMM) to achieve intelligent retrieval of multi-level current signatures from microRNA in serum samples. The results showed that the developed intelligent data analysis method is highly efficient for processing the large-scale nanopore data, which facilitates future application of nanopores to the clinical detection of cancer biomarkers.