Edited by: Fabyano Fonseca Silva, Universidade Federal de Viçosa, Brazil
Reviewed by: Victor Breno Pedrosa, Universidade Estadual de Ponta Grossa, Brazil; Gábor Mészáros, University of Natural Resources and Life Sciences, Austria
This article was submitted to Livestock Genomics, a section of the journal Frontiers in Genetics
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
Due to the complexity of longevity trait in dairy cattle, two groups of trait definitions are widely used to measure longevity, either covering the full lifespan or representing only a part of it to achieve an early selection. Usually, only one group of longevity definition is used in breeding program for one population, and genetic studies on the comparisons of two groups of trait definitions are scarce. Based on the data of eight traits well representing the both groups of trait definitions, the current study investigated genetic parameters and genetic architectures of longevity in Holsteins. Heritabilities and correlations of eight longevity traits were estimated using single-trait and multi-trait animal models, with the data from 103,479 cows. Among the cows with phenotypes, 2,630 cows were genotyped with the 150K-SNP panel. A single-trait fixed and random Circuitous Probability Unification model was performed to detect candidate genes for eight longevity traits. Generally, all eight longevity traits had low heritabilities, ranging from 0.038 for total productive life and herd life to 0.090 for days from the first calving to the end of first lactation or culling. High genetic correlations were observed among the traits within the same definition group: from 0.946 to 0.997 for three traits reflecting full lifespan and from 0.666 to 0.997 for five traits reflecting partial productive life. Genetic correlations between two groups of traits ranged from 0.648 to 0.963, and increased gradually with the extension of lactations number regarding the partial productive life traits. A total of 55 SNPs located on 25 chromosomes were found genome-wide significantly associated with longevity, in which 12 SNPs were associated with more than one trait, even across traits of different definition groups. This is the first study to investigate the genetic architecture of longevity representing both full and the partial lifespan simultaneously, which will assist the selection of an appropriate trait definition for genetic improvement of longevity. Because of high genetic correlations with the full lifespan traits and higher heritability, the partial productive life trait measured as the days from the first calving to the end of the third lactation or culling could be a good alternative for early selection on longevity. The candidate genes identified by this study, such as RPRM, GRIA3, GTF2H5, CA5A, CACNA2D1, FGF10, and DNAJA3, could be used to pinpoint causative mutations for longevity and further benefit the genomic improvement of longevity in dairy cattle.
Longevity is an economically important trait in dairy cattle, due to its large impact on the efficiency of dairy farming (
Since 1990s, longevity has been included in the total selection index in many countries (Interbull
By incorporating the information of genetic makers associated with the target trait into selection decisions, faster genetic progress could be achieved (
The objectives of this study were (1) to estimate heritabilities and genetic correlations for longevity traits with different definitions, including both traits representing full lifespan and traits representing partial lifespan; and (2) to identify genetic variants associated with longevity. Results of this study will assist the selection of appropriate trait definitions to be used for genetic improvement of longevity in dairy cattle.
The descriptive statistics of longevity traits in the Chinese Holstein population are presented in
Descriptive statistics of cow longevity traits in Chinese Holsteins.
Traits1 | N | Min | Max | Mean | SD | Coefficient of variation |
Lon11 | 103,479 | 1 | 365 | 330.6 | 93.2 | 28.2 |
Lon12 | 90,279 | 1 | 730 | 560.9 | 235.3 | 42.0 |
Lon13 | 82,826 | 1 | 1,095 | 680.9 | 368.2 | 54.1 |
Lon14 | 79,571 | 1 | 1,460 | 731.4 | 458.7 | 62.7 |
Lon15 | 78,144 | 1 | 1,825 | 749.2 | 510.0 | 68.1 |
ML | 78,227 | 1 | 3,836 | 738.0 | 506.8 | 68.7 |
PL | 78,227 | 1 | 4,007 | 822.3 | 575.9 | 70.0 |
HL | 78,227 | 1 | 4,825 | 1,617.7 | 588.7 | 36.4 |
Lactation | 117,491 | 1 | 13 | 2.7 | 1.6 | 60.2 |
The estimates of variance component and heritability for eight longevity traits from single-trait animal models are shown in
Estimates of additive genetic variance (
Trait1 | N | |||
Lon11 | 103,479 | 725.13 ± 64.05 | 7,333.92 ± 58.52 | 0.090 ± 0.008 |
Lon12 | 90,279 | 2,812.37 ± 292.38 | 42,117.79 ± 301.72 | 0.063 ± 0.006 |
Lon13 | 82,826 | 5,557.25 ± 610.59 | 98,433.22 ± 680.48 | 0.053 ± 0.006 |
Lon14 | 79,571 | 8,111.88 ± 903.21 | 151,005.01 ± 1,038.21 | 0.051 ± 0.006 |
Lon15 | 78,144 | 10,476.65 ± 1157.65 | 186,076.45 ± 1,309.04 | 0.053 ± 0.006 |
ML | 78,227 | 7,811.06 ± 995.05 | 187,066.34 ± 1,237.10 | 0.040 ± 0.005 |
PL | 78,227 | 9,627.98 ± 1,236.02 | 241,743.18 ± 1,574.84 | 0.038 ± 0.005 |
HL | 78,227 | 9,602.13 ± 1,232.00 | 241,404.44 ± 1,571.52 | 0.038 ± 0.005 |
Genetic and phenotypic correlations among HL, PL, and ML and among Lon11, Lon12, Lon13, Lon14, and Lon15 estimated using multi-trait animal models are presented in
Genetic (rg) and phenotypic (rp) correlations between longevity traits in Chinese Holsteins.
Trait definition group | Trait11 | Trait2 | rg ± SE | rp ± SE |
Partial lifespan traits | Lon11 | Lon12 | 0.912 ± 0.014 | 0.789 ± 0.072 |
Lon11 | Lon13 | 0.774 ± 0.031 | 0.622 ± 0.135 | |
Lon11 | Lon14 | 0.698 ± 0.039 | 0.533 ± 0.167 | |
Lon11 | Lon15 | 0.666 ± 0.042 | 0.487 ± 0.180 | |
Lon12 | Lon13 | 0.981 ± 0.004 | 0.919 ± 0.033 | |
Lon12 | Lon14 | 0.943 ± 0.011 | 0.831 ± 0.069 | |
Lon12 | Lon15 | 0.922 ± 0.015 | 0.775 ± 0.089 | |
Lon13 | Lon14 | 0.994 ± 0.002 | 0.964 ± 0.015 | |
Lon13 | Lon15 | 0.984 ± 0.004 | 0.922 ± 0.033 | |
Lon14 | Lon15 | 0.997 ± 0.001 | 0.984 ± 0.007 | |
Full lifespan traits | HL | PL | 0.951 ± 0.007 | 0.998 ± 0.001 |
HL | ML | 0.946 ± 0.008 | 0.994 ± 0.003 | |
PL | ML | 0.997 ± 0.001 | 0.996 ± 0.002 | |
Across trait groups | PL | Lon11 | 0.648 ± 0.044 | 0.462 ± 0.188 |
PL | Lon12 | 0.865 ± 0.022 | 0.707 ± 0.114 | |
PL | Lon13 | 0.938 ± 0.011 | 0.842 ± 0.065 | |
PL | Lon14 | 0.958 ± 0.008 | 0.907 ± 0.040 | |
PL | Lon15 | 0.963 ± 0.007 | 0.937 ± 0.028 |
The SNPs reached the genome-wide significant level in association with longevity traits in Chinese Holsteins are presented in
The genome-wide significant SNPs and candidate genes associated with longevity traits in Chinese Holsteins.
Chromosome | Position1(bp) | SNP | Trait2 | MAF3 | P-value | Effect size4 (%) | Candidate gene5 |
2 | 91,629,36 | rs135750821 | Lon15 | 0.43 | 1.71 E-07 | 0.49 | CALCRL |
2 | 42,630,124 | rs137712544 | Lon11 | 0.39 | 1.82 E-08 | 1.00 | RPRM |
2 | 46,201,858 | rs110628337 | Lon15 | 0.35 | 3.48 E-07 | 0.67 | – |
3 | 64,114,128 | rs42867525 | ML | 0.33 | 2.96 E-09 | 0.73 | – |
3 | 118,421,054 | rs137402734 | Lon14 | 0.39 | 1.72 E-08 | 0.65 | – |
4 | 11,190,947 | rs41613051 | Lon11 | 0.36 | 5.72 E-08 | 0.58 | TFPI2, GNGT1, GNG11, BET1 |
4 | 38,428,113 | rs136492033 | PL | 0.36 | 1.28 E-07 | 0.58 | CACNA2D1 |
4 | 85,565,338 | rs109342356 | ML | 0.18 | 1.10 E-07 | 0.50 | ING3 |
5 | 115,983,568 | rs135565406 | HL, ML, PL, Lon11, Lon15, | 0.42 | 9.96 E-10, 3.20 E-09, 7.95 E-08, 4.34 E-08, 7.01 E-08 | 0.73, 0.74, 0.55, 0.66, 0.63 | RIBC2, FBLN1, ATXN10 |
6 | 2,639,865 | rs137700260 | HL, Lon13, Lon14 | 0.07 | 3.42 E-07, 2.87 E-08, 1.35 E-07 | 0.47, 0.60, 0.54 | NPY1R, NAF1 |
6 | 6,916,170 | rs43451042 | Lon15 | 0.43 | 2.69 E-08 | 0.46 | PRSS12, NDST3 |
7 | 43,904,171 | rs29012367 | HL | 0.36 | 1.27 E-08 | 0.53 | CIRBP, FAM174C, PWWP3A, NDUFS7, GAMT, DAZAP1, RPS15, C7H19orf25, REEP6, MBD3, UQCR11, TCF3 |
7 | 54,635,988 | rs132789189 | HL | 0.45 | 5.21 E-09 | 0.65 | NR3C1 |
7 | 95,581,783 | rs109221022 | Lon15 | 0.39 | 4.11 E-07 | 0.68 | PCSK1 |
8 | 72,165,064 | rs43559099 | Lon12 | 0.26 | 6.83 E-08 | 0.49 | NEFM, NEFL |
8 | 72,275,281 | rs134248248 | Lon13 | 0.43 | 3.68 E-07 | 0.70 | NEFM, NEFL |
9 | 15,157,597 | rs42693004 | Lon12 | 0.15 | 3.46 E-07 | 0.55 | FILIP1, SENP6 |
9 | 94,882,463 | rs134672623 | Lon12, Lon15 | 0.45 | 2.59 E-07, 2.35 E-08 | 0.55, 0.95 | SERAC1, GTF2H5, DYNLT1 |
9 | 99,144,354 | rs109415769 | Lon14 | 0.24 | 9.83 E-08 | 0.44 | QKI |
11 | 14,367,967 | rs133167045 | ML | 0.17 | 8.37 E-09 | 0.53 | XDH, SRD5A2, MEMO1 |
11 | 24,568,406 | rs41592158 | PL | 0.38 | 3.61 E-07 | 0.42 | PKDCC, EML4, COX7A2L |
12 | 12,473,832 | rs43691605 | Lon13, Lon14 | 0.41 | 3.73 E-08, 2.14 E-08 | 0.48, 0.53 | – |
12 | 18,879,230 | rs109505275 | Lon11 | 0.30 | 9.14 E-09 | 0.80 | FNDC3A, CDADC1, SETDB2, PHF11 |
13 | 71,211,797 | rs109520811 | PL | 0.19 | 2.99 E-10 | 0.94 | – |
14 | 9,447,596 | rs110760208 | PL | 0.15 | 3.16 E-08 | 0.61 | – |
14 | 11,909,376 | rs110417753 | HL | 0.29 | 3.94 E-08 | 0.65 | – |
14 | 17,597,579 | rs110392124 | HL | 0.35 | 8.15 E-08 | 0.53 | – |
14 | 52,854,862 | rs133758639 | Lon11 | 0.33 | 3.45 E-07 | 0.60 | – |
15 | 23,500,595 | rs110052834 | Lon12 | 0.39 | 1.87 E-07 | 0.46 | NCAM1 |
16 | 13,142,992 | rs136880011 | PL | 0.14 | 7.51 E-08 | 0.55 | RGS18 |
17 | 39,552,453 | rs135929367 | Lon11 | 0.21 | 4.18 E-08 | 0.81 | RAPGEF2 |
17 | 70,229,172 | rs135552789 | PL | 0.48 | 1.63 E-07 | 0.52 | RNF185, LIMK2, PIK3IP1, PATZ1, DRG1, PISD |
18 | 13,532,413 | rs134612709 | Lon13, Lon14, Lon15 | 0.45 | 1.12 E-08, 1.23 E-09, 1.11 E-07 | 0.69, 0.73, 0.63 | CA5A, BANP |
18 | 62,531,696 | rs134066287 | Lon14 | 0.21 | 1.47 E-07 | 0.48 | RDH13, NCR1, FCAR, KIR2DL5A, KIR3DL1, KIR2DS1, KIR3DL2 |
20 | 30,741,701 | rs41567172 | ML | 0.15 | 4.07 E-07 | 0.43 | FGF10 |
22 | 19,689,530 | rs136814811 | PL | 0.50 | 3.85 E-07 | 0.54 | GRM7 |
23 | 40,794,975 | rs109782010 | Lon12 | 0.37 | 4.58 E-11 | 1.01 | LOC574091, GMPR, MYLIP |
24 | 4,941,502 | rs109618980 | HL, Lon14 | 0.25 | 1.28 E-07, 2.40 E-08 | 0.51, 0.59 | NETO1 |
24 | 39,596,108 | rs110544386 | PL | 0.22 | 1.39 E-08 | 0.58 | – |
24 | 50,494,672 | rs110534364 | HL | 0.10 | 1.27 E-09 | 0.66 | MRO, ME2, ELAC1, SMAD4 |
24 | 52,840,962 | rs136448037 | Lon12, Lon15 | 0.37 | 1.19 E-10, 3.27 E-08 | 0.80, 0.59 | – |
25 | 3,628,922 | rs109064423 | PL | 0.39 | 8.83 E-08 | 0.57 | TFAP4, GLIS2, CORO7, VASN, DNAJA3, NMRAL1, HMOX2, CDIP1, UBALD1, MGRN1, NUDT16L1, ANKS3, C25H16orf71 |
25 | 28,648,581 | rs109457044 | Lon15 | 0.16 | 9.62 E-08 | 0.62 | TYW1, CALN1 |
26 | 1,207,558 | rs109520495 | Lon12 | 0.15 | 2.30 E-07 | 0.51 | – |
27 | 42,337,167 | rs110187242 | Lon11 | 0.42 | 1.34 E-07 | 0.48 | UBE2E1, UBE2E2, MIR6532 |
28 | 22,970,021 | rs42105434 | Lon11, Lon12 | 0.05 | 4.97 E-08, 3.83 E-07 | 0.63, 0.50 | CTNNA3 |
X | 7,058,580 | rs134612804 | HL | 0.16 | 6.92 E-09 | 0.67 | |
X | 7,387,994 | rs135876977 | Lon11 | 0.30 | 2.58 E-12 | 1.43 | GRIA3 |
X | 7,811,883 | rs41622390 | Lon13, Lon14 | 0.08 | 4.39 E-09, 2.18 E-07 | 0.64, 0.49 | GRIA3, THOC2 |
X | 32,975,187 | rs137272895 | Lon13 | 0.41 | 6.21 E-09 | 0.96 | – |
X | 46,704,483 | rs137840659 | Lon13, Lon14 | 0.40 | 5.41 E-08, 8.61 E-09 | 0.59, 0.62 | – |
X | 55,267,455 | rs137389486 | PL | 0.40 | 3.89 E-09 | 0.85 | RNF128 |
X | 52,406,362 | rs137119176 | HL, ML, Lon13 | 0.25 | 1.47 E-08, 4.37 E-08, 9.37 E-10 | 0.81, 0.99,0.75 | TCEAL1, MORF4L2, GLRA4, PLP1, RAB9B |
X | 57,778,726 | rs41612337 | Lon11, Lon12 | 0.36 | 8.51 E-09, 9.69 E-08 | 1.19, 0.76 | GUCY2F, NXT2 |
X | 128,252,398 | rs110389261 | HL | 0.42 | 1.26 E-07 | 0.70 | PIR, VEGFD, ASB11, ASB9 |
The SNPs most significantly associated with Lon11, Lon12, Lon13, Lon14, Lon15, HL, PL, and ML were rs135876977, rs109782010, rs41622390, rs134612709, rs134672623, rs135565406, rs109520811, and rs42867525, respectively, in which rs135565406 was most significantly SNP associated with the full lifespan traits and rs135876977 was on one with the partial lifespan traits. Among the significant SNPs, the proportion of phenotypic variance explained by the top SNPs ranged from 0.73% (rs134612709 for Lon15) to 1.43% (rs135876977 for Lon11).
Manhattan plots and Q-Q plots for each longevity trait in Chinese Holsteins are presented in
Manhattan plots for 8 longevity traits in Chinese Holsteins. Lon11, the days from the first calving to the end of the first lactation or culling; Lon12, the days from the first calving to the end of the second lactation or culling; Lon13, the days from the first calving to the end of the third lactation or culling; Lon14, the days from the first calving to the end of the fourth lactation or culling; Lon15, the days from the first calving to the end of the fifth lactation or culling; PL, productive life referring the days from the first calving to culling or death; ML, milking life referring the days from the first calving to culling or death but excluding all dry periods; HL, herd life referring the days from birth to culling or death.
Quantile-quantile (Q-Q) plots of genome-wide association study for 8 longevity traits in Chinese Holsteins. Lon11, the days from the first calving to the end of the first lactation or culling; Lon12, the days from the first calving to the end of the second lactation or culling; Lon13, the days from the first calving to the end of the third lactation or culling; Lon14, the days from the first calving to the end of the fourth lactation or culling; Lon15, the days from the first calving to the end of the fifth lactation or culling; PL, productive life referring the days from the first calving to culling or death; ML, milking life referring the days from the first calving to culling or death but excluding all dry periods; HL, herd life referring the days from birth to culling or death.
Genes harboring or closest (with 200 kb distance) to the significant SNPs were suggested as potential candidate genes for longevity traits. By using this strategy, a total of 106 protein-coding genes and 1 micro RNAs were identified as candidate genes for longevity traits in Chinese Holsteins (
Our study showed that the average number of lactations in Chinese Holsteins was 2.7, based on the data of dairy cows born from 1999 to 2017. A previous study using data from 1990s showed that Chinese Holsteins usually culled or dead at 3.4 - 4.0 lactations (
In the present study, the estimated heritabilities of three full lifespan traits were relatively low (0.038 ∼ 0.040), which was within the range (0.01 ∼ 0.10,
Based on the breeder’s equation (
In current study, two significant SNPs (rs137712544 and rs110628337) on BTA2 are within the reported QTL (37,604,171-49,295,365 bp) for the total productive life in German Holsteins (
The genes GTF2H5 and CA5A are associated with more than one longevity traits and being detected by the top associated SNP for Lon15 (rs134672623) and Lon14 (rs134612709), respectively. The gene GTF2H5, participating in the interstrand adducts removal process of DNA repair, was reported to be associated with mastitis (
Among all 55 significant SNPs identified by the present study, most of them were within the reported QTL for calving traits (26 SNPs, mainly including calving ease, stillbirth, calf size, and birth weight), health traits (24 SNPs, mainly including mastitis, somatic cell count/score, abomasum displacement and ketosis), fertility traits (22 SNPs, mainly including non-return rate and gestation length), and immunity (16 SNPs, mainly including blood immunoglobulin G level). This phenomenon has also been observed in the previous GWAS for longevity in North American (
In current study, the X chromosome had the greatest number of significant SNPs (
This is the first study to investigate the genetic architecture of longevity traits representing a full or a partial lifespan simultaneously. Because of high genetic correlations with the total productive lifespan traits and higher heritability, the partial productive life measured as days from the first calving to the end of third lactation or culling (Lon13) could be a good alternative trait for early selection on longevity. The shared underlying biological processes among different longevity traits were further confirmed by the detection of shared significant variants. The genes RPRM, GRIA3, GTF2H5, CA5A, CACNA2D1, FGF10, and DNAJA3 were suggested to be candidate genes for longevity in Holsteins, which could be used to pinpoint causative mutations and further benefit genomic prediction for longevity in dairy cattle. This study proved the feasibility of genetic evaluation on longevity in Chinese dairy population, and it should be considered in selection index of Chinese dairy cattle.
The dates of birth, calving, drying, and culling or death were collected for 132,690 Chinese Holstein cows born from 1999 to 2017, which were raised in 31 herds in China, including 22 herds in Beijing, two herds in Hebei, and one herd each in Tianjin, Yunnan, Henan, Heilongjiang, Jilin, and Inner Mongolia, respectively. These electronic records of each cow were extracted from the farm management software (AfiFarm
A total of 2,629 cows born from 2003 to 2015 were genotyped with the Illumina 150 K bovine bead chip (Illumina, Inc., San Diego, CA, United States). Genotype imputation was performed using the Beagle 5.1 software (
Variance and co-variance components for eight longevity traits were estimated using the average information restricted maximum likelihood algorithm implemented in the DMU software (
where y is the vector of phenotypes for longevity traits, afc is the vector of fixed effect of age at first calving (≤22, 23, 24, 25, 26, 27, 28, 29, and ≥30 month); hy is a vector of fixed effect of herd-birth year; ys is a vector of fixed effect of birth year-season; a is the vector of additive genetic effects; and e is the vector of random residual effects. It was assumed that a ∼
The de-regressed estimate breeding values (dEBV) (
The data analyzed in this study is subject to the following licenses/restrictions: This manuscript utilizes proprietary data. Requests to access these datasets should be directed to YW,
Ethical review and approval was not required for the animal study because All phenotypic data were recorded as part of routine dairy cattle management and genetic evaluations. The DNA samples were obtained for the purpose of routine genomic evaluations in previous projects. Thus, no additional animal handling or experiment was performed specifically for this study.
HZ, GS, and YW organized the study. HZ and AL led the manuscript preparation. HZ performed the data analysis and HL did the data curation of genotype. XY and XL did the data curation of phenotype. LL provided support for collection of raw data. All authors contributed to the article and approved the submitted version.
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
We thank the Dairy Association of China (Beijing, China) for providing the pedigree.
The Supplementary Material for this article can be found online at:
Distribution of eight longevity traits in Chinese Holsteins.
The number of principal components (PC) added in GWAS model, the proportion of dEBV variance explained by these principal components and the inflation factors (λ) of single-trait GWAS for each longevity trait.
The potential candidate genes related to longevity and their functions.
Number of SNPs and genotyped animals kept for association analysis on each longevity trait in Chinese Holstein cattle.
Descriptive statistics of de-regressed estimate breeding values for longevity traits in Chinese Holsteins.
genome-wide association study
single nucleotide polymorphism
de-regressed estimate breeding value
principal component
quantitative trait locus
minor allele frequency
chromosome
base pair
the days from the first calving to the end of the first lactation or culling
the days from the first calving to the end of the second lactation or culling
the days from the first calving to the end of the third lactation or culling
the days from the first calving to the end of the fourth lactation or culling
the days from the first calving to the end of the fifth lactation or culling
productive life referring the days from the first calving to culling or dead
milking life referring the days from the first calving to culling or death but excludes all dry periods
herd life referring the days from birth to culling or death
somatic cell count
somatic cell score.