Abstract: Single nucleotide polymorphism(SNP)has very broad prospects in the research fields of population genetics of aquacultural species,molecular marker-assisted breeding and biological evolution.The development of SNP markers is normally obtained through genome sequencing,by sequence comparison.However,for the species for which no large-scale genome sequencing has been carried out,using EST database is often an important way for the development of SNP markers.In this study,the SNP genotyping assays and development of SNP markers for blue mussel(Mytilus galloprovincialis)were conducted through expressed sequence tags(ESTs)database mining.Some 19 709 EST sequences of blue mussel from the GenBank were clustered into 2 486 contigs,of which 963 contained four or more ESTs.After manual quality filtering,4 833 putative SNPs were identified from these SNP-containing contigs.The average putative SNP frequency was one per 129.2 bp of contig sequences.C/T and A/G with high frequency account for 28.8% and 27.4%,respectively.31 of the putative SNPs were chosen for validation by allele specific PCR with melting temperature(T_m)-shift analysis,and 14(47%)of them were polymorphic with the minor allele frequency ranging from 0.083 to 0.446.The observed heterozygosity and expected heterozygosity were distributed from 0.166 7 to 0.615 4 and 0.155 4 to 0.503 2,respectively.Six primers(20%)amplified no any product and 10(33%)were monomorphic.BLASTX showed that significant hits for all 14 genotyped SNP-containing contigs,12 of which were located in coding regions and all resulted in a synonymous substitution.The result of present study shows that ESTs could provide effective means for SNP identification in species with limited genome sequence resources,and T_m--shift analysis is a simple,efficient and reliable SNP genotyping method for non-model organisms.