Development Of Smartphone-based Mobile Digital PCR System And Its Application

Digital polymerase chain reaction(dPCR)circumventing the external calibration and potentially providing absolute quantification of nucleic acids has become an increasingly popular manifestation of PCR in biological researches.However,currently reported or commercial dPCR devices are not suitable for applications in laboratories or zones with limited infrastructures,due to low function integration and high cost.Herein,in order to enable accurate DNA quantitative analysis in such situations,a smartphone-based mobile dPCR device was developed.Furthermore,to extract proper signals from dPCR fluorescence images in chip-based digital assays automatically,an image processing method based on machine learning was also proposed in this work.The specific work and results of this paper mainly include the following aspects:(1)A miniature thermocycling control system suitable for handheld digital PCR instrument was developed,in which hardware and temperature control algorithm were both designed and optimized.The performance of this system was also validated;(2)Based on the detection principle of the Taqman probe,a fluorescence detection system suitable for the handheld digital PCR device was constructed through hardware selection and software development,and a good fluorescence imaging effect was achieved;(3)A handheld dPCR device was successfully developed by integrating thermocycling control system,fluorescence detection system and microfluidic chip,in which all the function units were automatically controlled using a custom Android software.The device is approximately 90mm × 90mm × 100mm in size and about 500 g in weight.It only costs about ￥2000 except the smartphone,much lower than commercial dPCR instruments.Through the dPCR validation,this device showed the ability to absolutely quantify nucleic acids with low copy numbers,as well as a comparable analytical accuracy with the QuantStudio TM 3D dPCR system launched by Life Technologies Corporation;(4)In order to overcome the deficiencies of traditional threshold segmentation based method in fluorescence image processing,a new method using machine learning for automated image analysis applied to chip-based digital assays was proposed.And the detection accuracy can achieve 97.78%.The dPCR device developed in this work is the smallest one up to now,which is simple in structure,low-cost with high detection accuracy.Combining with the newly developed machine learning based fluorescence image processing algorithm,the performance of the device will be further improved.Therefore,it has a great potential in Point-of-care(POC)applications.