Genome-wide Association Mapping Of Alternative Splicing Variation In Maize Kernel

Altermative splicing(AS)is prevalent in eukaryotes,which plays important roles in regulating gene expression and enhances proteome diversity.Therefore,identifying the genetic factors regulating AS variation will not only help us understand its roles in transcriptional and post-transcriptional regulation but also reveal its contribution to phenotypic variation.Although nuoerous studies on gene expression regulation have been reported,studying genetic variation regulating AS has received little attention in plants.Here,a genome-wide association study(GWAS)was performed based on RNA-Seq data of immature maize kernels from natural population consisting of 368 diverse maize inbred lines.We aim to dissect the genetic architecture of AS and investigate its regulatory mechanisms and contribution to phenotypic variation.The major results are as follows:1.Reference-based transcript assembly was performed at population level using high-throughput RNA-Seq data.A total of 49,132 novel transcripts were assembled.As a result,the proportion of genes undergoing alternative splicing increases from 32%to 50%;2.Based on 1.25 million high density SNPs(MAF>0.05),GWAS on AS was performed.Using a genome-wide significance cutoff(P<4E-08),a total of 8,293 splicing QTLs(sQTL)involving 5,028 genes were detected,of which cis-sQTL predominates overtrans-sQTL.Functional annotation showed that genes detected with sQTL were significantly enriched in genes related to protein and RNA metabolism,spliceosome and response to stimulus;3.Most of sQTLs(72%)showed small isoform usage changes without involving major isoform switch between genotypes.By characterizing sQTL-associated AS events,we found that intron retention is the most frequent AS type,followed by alternative 3' and 5' splice sites,and exon skipping;4.Nonsense-mediated mRNA decay(NMD)and peptide interference(PEPi)are two important regulatory mechanisms that AS mediate.We found that 4,228 sQTL regulated transcripts were potential targets by NMD,of which 1,552 sQTL affected transcripts carry premature termination codon(PTC).Furthermore,we found that 207 sQTL and 783 sQTLs were potentially affected by PEPi mechanism;5.The comparison analysis of cis-sQTL and cis-eQTL identified previously showed that only 13.3%sQTs simultaneously affect mRNA level,suggesting that AS appears to be under relatively independent genetic control from overall mRNA level.Further analysis on the variant type of cis-sQTL and cis-eQTL showed that cis-sQTL are significantly enriched in UTR,intron,splice sites and splice regions,while cis-eQTL are significantly enriched in UTR,indicating that different cis-sequences are preferentially used in the cis-regulation of AS and mRNA level;6.Through analysis of the genomic distribution of trans-sQTL,we identified 49 putative transacting splicing factors.These splicing factors include genes encoding enzymes,transcription factors,splicing factors and ribonucleoprotein complex.As a representative example,we conducted further analysis for ZmGRP1;7.Through the colocalization analysis of sQTL,mQTL and pQTL,we determined that 280 sQTL colocalize with the previous GWAS signals for downstream phenotypes,most of which occurred without changes in overall mRNA level,underscoring the importance of AS in phenotypic variation.