Detection Of Candidate Genetic Markers Of Polled Phenotype And Mining Of Functional Gene In The Sichuan Cattle

Horns,one of most crucial phenotypic traits of Bovidae family,have a significant influence and impact on surviving in competitive wild environment.However,the hornless natural mutant has been observed and noticed in cattle population from Europe and Africa since 8000 years ago.Early researcher has already proved that it is dominant heredity.Even though there are a large population of cattle in Sichuan province,they were all horned in the past.In recent decades,in order to improve the indigenous cattle,Simmental and Holstein have been introduced into Sichuan province,leading to the fact that during the process,hornless alleles have also been introduced into cattle genes pool,which results in a phenomenon that some polled cattle have showed up.Nowadays,research proved that there are three genetic markers related to hornless phenotype and one of each mutating will cause to polled phenotype.However,it is also unclear that which marker has been introduced into cattle population in the Sichuan province,which limits the selection of those molecular makers.In order to address those issues,we conducted our study from the following aspects:(1)Screened polled genetic makers among cattle.We collected ear tissue samples from 48 polled and 16 horned cattle and they are unrelated with each other.Then we designed PCR primer to amplify three markers,including P2021D?P80kbID?P2191D,which relate with hornless phenotype and four markers,containing P51D?PG1855898A?PC1768587A?PG1654405A,which unrelate with polled phenotype.In the end,the results showed that among 48 polled cattle,there are 39 cattle contain P202ID mutations;among those mutations,there are 36 heterozygous mutations and 3 mutations are homozygous.4 cattle have P80kbID mutation and they are all homozygous.Besides,5 cattle contain P2191D mutation,which all are heterozygous.In total,the mutation frequency of P219ID,P80kbID and P2191D are 43.75%,8.33%and 5.68%respectively.However,one of those has no mutation at all and all horned cattle are wild homozygote.Moreover,the linkage analysis found that the positions among P5ID?PG1855898A?PC1768587A and PG1654405A are 100%linked;in contrast there is no absolute relation among hornless phenotype.Further,we sequenced DNA control region of mitochondria and result showed that 65.6%cattle are Bos taurus,the left 34.4%are B.indicus.However,there is no difference on the rate polled cattle between these two populations.(2)A machine learning method based on multiple genetic markers to predict POLLED phenotypes.Because of the complete or incomplete correlation between multiple independent genetic markers and the POLLED phenotype,it is necessary to establish a more complete model to select the genetic markers completely related to the POLLED phenotype.constructed three models,such as logistic regression,support vector machine and decision tree model.The genotypes of 7 genetic markers were used as the model's feature input to predict the POLLED phenotype,Among them,80%samples were randomly selected as training data and 20%samples as test data.And the F-measure(fl-score)and receiver operating characteristic curve(ROC)were used to measure the performance of the model The results showed that the F-measure of logistic regression model was 93%and the AUC of ROC was 95%,while the two evaluation indexes of support vector machine and decision tree model were 100%.(3)Mining of un-annotated expression signals based on Transcriptome sequencing data in POLLED locus.To solve the problem,that functional genes leading to POLLED phenotype in cattle have not been identified,30 transcriptome data from 4 breeds and 20 tissue samples were downloaded from NCBI database in this study.After quality filtration,high quality reads were mapping to the reference genome sequence,and transcripts were assembled using StringTie software.Finally,eight genes were detected in the 381Kb target region(1.668 Mb-2.049 Mb)of chromosome 1,including six annotated genes(ITSN1,CRYZL1,DONSON,SON,GART,DNAJC28)and two un-annotated genes(LOC104970777 and LOC112447103).Homologous comparison showed that the homologous gene of LOC104970777 was DNAJC28 in sheep,while the homologous gene of LOC112447103 was CRYZL1 in Bubalus bubalis,Capra hircus,Ovis aries,Pantholops hodgsonii and bisonAnalyzed the POLLED of the phenotypic-related genetic markers in the Shuxuan cattle,and established a machine learning method based on multiple genetic markers to predict POLLED phenotypes.Moreover,preliminarily explored the functional genes that may lead to POLLED phenotypes in the cattle.The results can be used to establish a marker-assisted selection method for the breeding of sichuan cattle.