Multi-Channel Series Piezoelectric Quartz Crystal Sensor with Partial Least Square Method for Identification of Clinically Relevant Bacteria

摘要

Routine clinical microbiological detection and identification of bacteria are largely based on biochemical tests. Other simple methods developed recently all perform identification after the isolation of pure cultures. The unavoidable time delay associated with such identification procedures will be fatal to the severely ill patients. In this paper, we describe a novel identification method based on combination of multi-channel series piezoelectric quartz crystal sensor with partial least square method (PLS-MSPQC). With this approach the frequency responsive curves, which have obtained during the detection of the most commonly encountered microorganisms by multi-channel series piezoelectric quartz crystal sensor (MSPQC), can be recognized by partial least square method (PLS). The advantage of proposed method in this research is to incorporate the identification into the cultivation and detection, which is cost-cutting and time-saving. With the proposed PLS-MSPQC system, 47 bacterial strains isolated from hospitalized patients were successfully detected and identified consuming around 20 h. According to statistic analysis, there was no significance discrepancy between PLS-MSPQC method and API 20C AUX method for identification of the four types of bacteria. It shows that PLS-MSPQC method has great potential as a powerful new tool in clinical diagnostic microbiology.