Application Study Of Trace Sample Preparation Based On Microfluidic Technology For High Throughput Sequencing

With the development for more than a decade,high-throughput sequencing has become an important tool to comprehensively understand genome,also has been widely applied in many core research in biology and genetics such as the construction of animal and plant genome,investigation of the genetic mechanism of disease,exploration of microbe or environmental organisms and so on.However,the complexity of the genome is not limited to its grand scale,but more importantly,the mechanism of genome functionalization in the biological progress is unknown and complex.Therefore,we need information of more high quality and from different observing dimension to further predicting the mechanism of the genomic events.Whole genome amplification is a landmark technique for sample preparation in high throughput sequencing,which expands the detection threshold of high-throughput sequencing in input sample quantity,enabling us to observe the change of genome structure at an actual single-cell spatial resolution and this information tends to be annihilated in the averaged conventional sequencing data.However,current whole genome amplification techniques are not perfect,and according to the principle of their biochemistry,they are affected by different types and characteristics of errors,mainly in the following aspects: coverage,uniformity,repeatability,chimera and base substitution,etc.Most of the improved whole genome amplification methods require either complicated and expensive instruments or tedious experimental procedures and reagents which may interfere with the amplification.Response to these shortcomings,this paper aims to develop a new whole genome amplification protocol and related microfluidic device to realize the high-efficienct,high-performance,low-cost and convenient amplification of the trace samples.In this paper,we proposed a novel whole genome amplification approach,microchannel multiple displacement amplification,?cMDA,which substitutes a quasi-linear reaction space for the 3D reaction space of conventional MDA.Due to the limited diffusion of the macro molecules,such as the templet-polymerase complexes and amplicon clusters,reaction units are distantly distributed across the slender space and the interactions only occur among the few neighboring units.The continuous overamplification of preferential amplicon clusters is,therefore,restricted because of the exhaustion of local reactants and MDA's major technical obstacle of the systematically non-uniform amplification is suppressed.We use this method to amplify the normal diploid YH-1 genome and the aneuploid K562 cancer genome.From the starting biochemical experiment to the final sequencing data analysis,we validated the performance advantages of this new method in a number of important parameters,such as: sequencing coverage of 30× YH-1 data reached 92.18%,much larger than that of MALBAC,but sligtly lower than the experimental theoretical limit.In terms of the effectiveness of sequencing,10× data of this new method can cover the same size of regions with 25× conventional sequencing data,which means 60% reduction of the sequencing cost.In amplification uniformity,we use coefficient of variation,lorenz curve,the correlation analysis and the power spectral density methods to verify the broadened coverage is the result of significantly more uniform amplification.In addition,we also discussed the benefits of this method in engineering,such as there requirs no extra instrument and reagents;the consumables are available and inconvenient;the operations are simple which reduces the probability of contamination.We also speculated on the principle model of the new method,and carried out the control variable experiment to verify the correlation between the inner diameter and the amplification uniformity.Subsequently,we applied the micro-channel multiple displacement amplification in two specific detections in high-throughput sequencing,single nucleotide variation and copy number variation detections.In the single nucleotide variation detection,due to the high fidelity,high coverage and high uniformity advantages,both ~27% more homozygous and heterozygous single nucleotide variations can be detected out with a sensitivity increase of ~11%.In addition,better results were obtained in the detection defects such as allele deletion,detection error and false positive estimation.In the copy number variation detection,this method mainly plays its characteristics of amplification uniformity and GC-pereference suppression.In the simulation of copy number variation detection in the diploid YH-1 genome,this new method provides averagely ~20% higher detection rate for the copy number variations from 300 kb to 2 Mb in size.In detecting the real copy number variations in the K562 cancer genome,we reduced the sample input to the single-cell level,and sucessfuly detected out 250-kb tiny copy number variations under 50-kb variable binsize.This new method provides better results in both the final and phased copy number variation patterns of high selfsimilarity and high consistency with the bulk reference,implying better repeatability and accuracy.In this paper,we have also designed a microfluidic device compatible with our new method.The chip contains multiple sets of parallel channels to significantly reduce the injecting pressure while remaining as large capacity as the classical long channel.This design additionally expands the use of this device for multiple sample preparations.In the sample-preparation related microfluidic droplet manipulations,we realized the the generation of multiple sets of monodisperse droplets,low-voltage droplet fusion and dynamicly adjustable droplet division,laying the foundation for the subsequent development of barcoding-based parallel sample amplification and sequencing.