Research On A Highly Precise And Automatic Droplet Digital PCR System

Droplet microfluidics can generate monodisperse droplets in a high-throughput manner.Droplet digital PCR(dd PCR)is an important application of droplet microfluidics,which is able to achieve absolute quantification of nucleic acids.In recent years,dd PCR is becoming an enabling technology with wide applications in clinical diagnosis.For these applications,accuracy and objectivity are two key demands.As far as we know,highly precise and automatic dd PCR systems have not yet been reported.To address these challenges,this dissertation established a novel dd PCR system for highly precise and automatic analysis of nucleic acids and conducted a quantitative analysis of these factors with high precision and automation,ranging from droplet generation process to droplet detection process.The research mainly includes the following parts:(1)Highly precise and automatic analysis of droplet generation process.In this dissertation,a highly precise droplet generation apparatus was developed to observe droplet generation processes in real time.Based on this apparatus,a novel cosine similarity algorithm(CSA)method was proposed for automatic online monitoring of dynamic droplet generation processes.The CSA method utilized the periodicity of droplet generation process to calculate droplet generation frequency and its coefficient of variation(CV)with a video clip captured with the high-speed camera in a conventional droplet generation system.The droplet generation frequency was used to calculate the number of generated droplets,and its CV was used to characterize the uniformity of the droplet generation process.To validate the performance of the CSA method,four kinds of dynamic droplet generation processes were studied.These experimental results demonstrated that the CSA method is a highly accurate and automatic method for the online determination of the droplet counts and uniformities of droplet generation processes.(2)Highly precise and automatic analysis of droplet detection process.In this dissertation,a highly precise droplet detection apparatus was developed with confocal microscopy to automatically determine the number of droplets detected with high precision.Based on this apparatus,a novel density-based watershed algorithm(DWA)method was proposed for accurate,robust and automatic classification of dd PCR data.The DWA method adopted gridded data density as a parameter and applied a watershed algorithm to automatically segment the gridded scatter plot into isolated regions.With region selection and merging,the dd PCR data could be classified in an automatic and unsupervised way.The DWA method achieved absolute quantification of nucleic acids with high precision in four orders of magnitude.These experimental results demonstrated that the DWA method achieves accurate,robust,and automatic classification of dd PCR data to determine in a highly accurate and automatic way the number of positive droplets detected.(3)Synthesis of a biocompatible surfactant.A novel method was proposed and carried out to synthesize a biocompatible surfactant by catalyzing the reaction with dimethylformamide(DMF)and extracting the product with a mixed solvent of dichloromethane(DCM)and tetrahydrofuran(THF).The synthesized surfactant maintained high droplet stability and monodispersity as well as high amplification efficiency during dd PCR process.This biocompatible surfactant enables high precision of nucleic acid quantitative analysis with dd PCR.With these aspects combined,a novel dd PCR system was developed in this dissertation.The droplet generation process and droplet detection process were quantitatively studied,and highly precise and automatic quantification of plasmids and nucleic acids from clinical samples was achieved.With these improvement in precision and automation of dd PCR,this research is promising to propel the applications of dd PCR to clinical diagnosis.