Abstract: 【Objective】We herein integrated digital gene expression profiling and genome-wide association study in a White Duroc × Erhualian F₂ resource population to identify candidate genes for hematological traits.【Method】The White Duroc × Erhualian F₂ resource population were slaughtered at 240 ± 3 days. Blood was collected in anticoagulation tubes. A set of hematological parameters were measured using a whole blood analyzer. The 1 020 F₂ pigs were genotyped using an Illumina porcine 60K SNP chip. Individuals with genotype-missing rates > 10% and Mendellian errors > 5% were removed. SNPs with a call rate < 95%, minor allele frequency < 5%, P value < 5 × 10^-6 for Hardy Weinberg equilibrium (HWE), and the autosomal SNPs that were linked to sex chromosome were excluded. The liver samples of the 502 F₂ pigs were then performed digital gene expression profiles sequencing on Illumina GA II. The raw tags were filtered to obtain clean tags. The clean tags were uniquely mapped to the reference gene sequences and were defined as unambiguous clean tags. The number of unambiguous clean tags was normalized to represent the expression level of each transcript. The expression level of each transcript was further transformed to lg2 value. The transcripts that expressed less than 20% of individuals were rejected. The traits of phenotype and gene expression were adjusted for sex, batch and kinship using polygentic function of GenABEL in R package. The correlations between gene expressions and phenotypic traits were evaluated using the residuals by Spearman’s correlation coefficient with a conservative threshold P < 0.0005. Positions of detected eQTL were plotted against the positions of the genes for which that eQTL were found. We also searched the eQTL within the 5.0 Mb region of the peak SNP of GWAS and performed an integrated analysis of eQTL and GWAS. Gene Ontology & KEGG pathway enrichment analysis was implemented by DAVID online tools and gene co-expression network was constructed by GeneMANIA online tools. 【Result】A total of 20 108 liver transcripts of 502 F₂ pigs achieved the quality control requirements. We obtained 259 transcripts strongly associated with hemoglobin (HGB), red blood cell count (RBC), hematocrit (HCT), mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH) and white blood cell count (WBC) respectively with a conservative threshold P < 0.0005. 34 liver transcripts were significantly associated with two or more phenotypic traits. Three hundred and four loci (eQTL) were identified to affect the transcription levels of these genes including 35 cis-eQTL and 120 trans-eQTL with P < 10^-5. Each transcript was mapped to one to six eQTL. MCH and MCV shared the same cis-eQTL located on SSC8. The largest number of eQTL were located on SSC7 and most of them were trans-eQTL. KIT was identified as a candidate gene by eQTL analysis. Gene Ontology & KEGG pathway enrichment analysis allowed us to prioritize five candidate genes such as for KIT, PSEN2,and TFRC for RBC and THBS1, CYR61 for WBC. RPS10 was also identified as the candidate gene for WBC by the integration of eQTL, GWAS and gene co-expression network. 【Conclusion】In this study, we identified KIT, PSEN2,TFRC as the candidate genes for RBC, THBS1, CYR61, RPS10 as the candidate genes for WBC by integrating gene expression profiling and genome-wide association study in the White Duroc × Erhualian F₂ resource population.

Key words: pig, hematological traits, gene expression profiling, expression quantitative trait loci, genome-wide association study, candidate gene