Droplet Nucleic Acid Two-Dimensional Barcoding For Parallel Sequencing

With the development of single-cell and trace nucleic acid sequencing,the traditional barcoding methods are difficult to meet the needs of labelling a large number of samples.Besides,conventional sequencing library construction methods require a large amount of labor,which limits large-scale sample processing,and further hinders high-throughput parallel sequencing.Droplet microfluidics,which is an important branch of microfluidic technology,provides an opportunity for large-scale labelling.Droplet microfluidics could use water-in-oil structure to separate sample into different droplets for independent reaction.Current researches in large-scale labelling for genome sequencing based on droplet microfluidics take use of multiple random bases as barcode.Although this method improves barcoding capacity,it may influence sequencing quality and decoding deviation due to the improper GC content and undesired secondary structure.Meanwhile,the unknown barcodes make it difficult to decode and correct error after mismatching during PCR.This thesis aims to develop a large-scale barcoding method on droplet microfluidic platform,which could be decoded easily and further applied to parallel sequencing of massive samples.The main contents of this thesis are shown as follow:1.Proposing a‘zipper’method for pairing and merging of two droplets with P7/P5 adaptorSamples were labeled from two dimensions（2D）in this thesis and the barcoding capacity was improved by random combination of two barcodes consisting of known sequence.A‘barcode pairing’droplet merging method was established based on droplet microfluidics,which is the key technique of this project.In this method,the two kinds of droplets with P7/P5 adaptor were staggered like a‘zipper’and further merged in pairs.By analyzing fluorescence images and droplet images during process,the strategy achieved a success rate of 53%.After eliminating ineffective droplets such as uncoded droplets and single coded droplets,the accurate 2D barcoded droplets accounted for 97%of all effective droplets.2.Establishing the protocol for droplet sequencing library construction using Tn5 transposaseA protocol for sequencing library construction was adopted,in which full-length adaptors were encapsulated in Tn5 transposase.The protocol,which simplified sequencing library construction process,made reaction conditions in droplets mild,and avoided the demulsification and mismatch problems caused by droplet PCR.The process consisted of four steps:transposome construction,tagmentation,nick translation and PCR enrichment,fragment size-selecting.The key elements,including template amount,Mg²⁺concentration,fragmentation method and PCR initial amount,were optimized to improve the fragment length.The protocol was verified from three aspects of agarose gel electrophoresis detection,cloning sequencing and second-generation sequencing.The droplet reaction demulsification rate of the protocol was only 0.824%,which was 38 times lower than methods based on PCR under the same condition.3.Parallel sequencing with droplet 2D-barcode labeling and construction of analyzing processDroplet two-dimensional barcoding and parallel sequencing were performed by combining nucleic acid 2D barcode pairing with Tn5 sequencing library construction process.Data normalization method was established according to the two-dimensional barcoding characteristics.Using the strategy,MDA product of cell line GM12878 was labelled with 2D barcodes and sequenced.The result showed that all combinations of 2D barcodes were inclusive,and the amount of data under each barcode was at the same level,with a standard deviation of 0.06,which was 2/5（based on i7）and 1/7（based on i5）of the standard deviation of based on single-index normalization.Compared with the sequencing data in PCR tube transposition,this method had better performance in uniformity and GC content decreased by 2.3 percentage points,which could probably improve the preference of high GC content to some extent.