The In Situ RNA Detection And Spatial Transcriptome Analysis Of Gene Expression And Alternative Splicing

BackgroundGene expression and alternative splicing(AS),as an integrated system,regulates physiological and pathophysiological processes at the temporal(cell division,circadian rhythm,development,aging,etc.)and spatial dimension(organs,tissues,cells,organelle,etc.).They together affect the structures and functions of proteins.AS is a highly regulated post-transcriptional process whereby a single gene can produce multiple m RNA isoforms from one gene.It dramatically diversifies the transcriptome and proteome in eukaryotes,with more than 95%of human genes undergoing alternative splicing.A large amount of whole-genome and transcriptome sequencing data provides sufficient resources for exploring the evolution of transcriptome diversity and intra-/inter-species complexity.AS regulates multiple processes in a spatio-temporal specific manner during cell differentiation and organ development,including the development of preimplantation embryos,the differentiation of megakaryocytes and erythroid lineage,the formation of axons,and the differentiation of neural progenitor cells to neurons during the development of the cerebral cortex.Dysregulation of AS pattern causes tumorigenesis and other pathological processes.Therefore,how to better study the spatial location and expression level of genes and splicing is one of the important fileds of transcriptome research.Recently,the advances of several biotechnologies have greatly promoted the research of spatial transcriptomes,including single-cell RNA sequencing(sc RNA-seq),multiplex RNA Fluorescence in situ hybridization(RNA-FISH)technology,and the next generation sequencing(NGS)based spatial transcriptome and third generation full-length sequencing(TGS)technology.They have their own characteristics and are complementary.First,sc RNA-seq facilitates the studies of rare cell types and heterogeneity using single-cell expression profiling.However,the current popular single-cell sequencing technologies are based on the premise that the tissue is dissociated into single cells,and therefore,the spatial information of the microenvironment and tissue morphology are lost.Second,multiplex RNA-FISH technology introduces a coding scheme based on traditional RNA-FISH,increasing the number of fluorescence channels of microscope through multiple rounds of hybridization and imaging.By integrating fluorescent signals of multiple rounds of imaging,decoding can be achieved,and therefore,the number of target RNAs that can be detected by RNA-FISH exponentially increased and potentially can be applied to the whole transcriptome profiling.It can not only provide information on the distribution of gene expression profiles at the tissue level,but also defining cell types using gene expression patterns of cell populations,meanwhile,observing the gene expression patterns of cell development microenvironments such as hematopoietic microenvironment,tumor microenvironment,or tumor heterogeneity using differentially expressed genes,which can be integrated with single-cell sequencing data to study physiological and pathological processes.However,due to the transcriptome-wide RNA-FISH at single molecule/single cell resolution technology is still in its infancy,special equipment(such as liquid handling robots that automatically change liquids between multiple rounds of hybridization,super-resolution fluorescence microscopes,etc.)or experimental skills(the reproducibility of the experimental process is relatively poor)have certain requirements,making this technology relatively difficult to be widely used.In contrast,the spatial transcriptome technology based on NGS greatly simplified the workload of transcriptome RNA-FISH and lowered the threshold for the use of the technology so that it can be extended to a wide variety of applications,is currently the most widely used whole-transcriptome sequencing technology that can retain tissue morphological characteristics for solving more biological problems.Although the second-generation short-sequence sequencing technology has the advantage of high accuracy,however,as the fragments interrupting during the library preparation,only the 3’-end or 5’-end sequence remains intact,resulting in the lost of the full-length alternative splicing and sequence information.Finally,most of studies using high-throughput RNA-FISH and spatial transcriptome analysis have focused on gene expression,while the alternative splicing has not been largely unexplored.Third-generation full-length sequencing technology is ideal tool for the detection of splicing,gene fusion and structural variation such as large insertion and deletion.Oxford Nanopore Technology(ONT)has become more and more popular due to its increased sequencing accuracy,throughput,flexibility,and cost-efficiency.However,the characteristics of ONT sequencing determine that its sequencing accuracy and throughput need to be improved.Therefore,due to method limitations of each method,it is still difficult to analyze the spatial distribution of alternative splicing.Here,we explore the specific solutions for the following two difficulties:(1)There are no specific probes designing tools of RNA-FISH assay for splicing sites(usually between two exon junctions).The signal amplification effect conducted by rolling-circle amplification mediated by a DNA polymerase(such as phi29)harboring strand displacement characteristic,triggered by padlock probe is considered to be a trade-off between the detection efficiency,fidelity,and technical difficulty in a wide variety of high through-put RNA-FISH technologies.However,due to the unique probe structure,there is currently a lack of rapid and high-throughput padlock probes designing tool equipped with specificity(at the transcriptome level)checking for splicing sites or circ RNA(reverse splicing).This project integrates the existing multiple RNA-FISH detection methods,develops padlock probes designing software,and provides a set of relatively comprehensive solution from probe designing and sample preparation to fluorescence signal recognition.(2)During the research of spatial transcriptome,such as the 10×Genomics Visium spatial transcriptome,which relies on the capture of poly A~+m RNA,adding cell barcode and molecular tag(Unique Molecular Identifiers,UMI)NGS technology,most of the alternative splicing information is lost.The third-generation full-length sequencing technology can sequence the full-length transcriptome.However,the accuracy of ONT sequencing is relatively low,making it difficult to match the cell/feature barcodes and molecular tag(UMI)from single-cell sequencing libraries or spatial transcriptome sequencing libraries.We used the consensus methods that have been reported so far,in addition to four methods of double-stranded DNA self-looping to conduct a systematic evaluation to determine the experimental method with the most data output and the best self-correction effect in single cell/spatial transcriptome library.This thesis focuses on the improvement of in situ detection and the accuracy of the third-generation ONT sequencing,and provide a comprehensive analysis of genes and alternative splicing in single cell/spatial transcriptome of mouse brain.Methods and Results(1)Integrate the published RNA-FISH experimental methods,determine a set of"PFA fixation-heat shock-permeabilization-padlock probes directly hybridize to the target RNA-padlock probe nick sealing-rolling circle amplification-amplified product acrylic modified fixed-fluorescent probe hybridization detection"multiple RNA-FISH program,at the same time design and development of padlock probe high-throughput designing tool,comprehensive consideration of up to ten probe hybridization conditions to give candidate probes binding region,with the assistance of high-throughput comparison software,the candidate probes are screened,and the probe binding region with higher specificity is reserved for lock- type probe construction and mfold software for secondary structure evaluation.In addition,it also integrates a set of software that uses the Difference of Gaussian(Do G)method to perform spot detection on fluorescence in situ hybridization signal points and multiple rounds of image registration between hybridizations(based on the b Unwarp J algorithm)to realize the detection of fluorescence.Recognition,quantification and image reconstruction of in-situ hybridization signals,and can reconstruct the identified fluorescent in-situ hybridization signal points onto the HE stained picture of the same slice.(2)Five methods,including Gibson recombination,stem-loop structure joint,splint-mediated lock probe ligation,blunt-end double-strand self-loop method,and Circ Ligase-mediated single-strand Self-looping and rolling-circle amplification,are used to form long tandem repeats,and the sequence is corrected by splitting each other,so as to systematically compare the effects of five looping methods on sequence self-correction.Furthermore,the Gibson recombination method and the splint-mediated lock-probe ligation method were performed and compared using the mouse brain spatial transcriptome full-length c DNA library.(3)Using the 10×Genomics Visium spatial transcriptome followed by second-generation sequencing and ONT sequencing,we match the barcodes and analyze the spatial distribution of gene expression and AS in mouse brain.ConclusionsStudies of transcriptional and post-transcriptional regulations require better experimental and bioinformatic methods with spatial information.This study explores the following three aspects:(1)The experimental method for single-molecule RNA-FISH detection has been systematically optimized,and the high-throughput design of lock-type probes at both gene expression and splicing levels has been developed.The method of in situ hybridization signal spot pattern recognition and registration,especially to reduce the design difficulty of in situ detection at the splicing level;(2)Given the relatively low accuracy of third-generation ONT sequencing,we systematicly compared five potential sequences and found that the advantages of one method,providing an potentail solution for the application of ONT sequencing in single cell and spatial transcriptome;(3)Integrating the 10×Genomics Visium spatial transcriptome,in situ detection and ONT sequencing,we explore the gene expression at the transcriptome level and alternative splicing and RNA editing at the post-transcriptional level for mouse brain slices.In summary,this research provides a complete solution for the spatial analysis of gene expression and alternative splicing,including padlock probe design,image processing,and the accuracy improvement of ONT sequencing,providing new tools for the study of gene expression and alternative splicing regulations in the variety of physiology and pathology settings.