Serviços Personalizados
Journal
Artigo
texto em 
Inglês (pdf)
Artigo em XML
Referências do artigo
Tradução automática
Enviar este artigo por emailIndicadores
-
Citado por SciELO -
Acessos
Links relacionados
-
Similares em
SciELO
Compartilhar
Permalink
Revista bio ciencias
versão On-line ISSN 2007-3380
Revista bio ciencias vol.8 Tepic 2021 Epub 04-Out-2021
https://doi.org/10.15741/revbio.08.e1052
Original articles
Minor allele frequency in genomic prediction for growth traits in Braunvieh cattle
1 Universidad Autónoma Chapingo, Posgrado en Producción Animal, Departamento de Zootecnia, Estado de México, México.
The present research aimed to compare the effect of minor allele frequency (MAF) threshold during the quality control process of genotypes, using single nucleotide polymorphisms (SNPs) to predict genomic values of growth traits in Braunvieh cattle. The analysis was performed including 28,973 and 18,994 phenotypic records of birth and weaning weights, 12,835 single nucleotide polymorphisms and 300 animals. The traits were analyzed with the single-step genomic evaluation approach. The criteria for comparison of the minor allele frequency thresholds compared were: 1) ranking of genomic values: Spearman and Pearson correlation coefficient estimates were obtained for genomic values predicted with standard (0.05) and lower levels of minor allele frequency; 2) prediction ability of models: cross- validation with four replicates was performed, average of correlation between real and predicted phenotypes was obtained; and 3) regression coefficient estimates: the genomic values obtained with standard and lower levels of minor allele frequency were used as independent and dependent variables, respectively. The effect of minor allele frequency threshold on the ranking of genomic values was not important (correlation estimates greater than 0.99, <2.2x10-16). The prediction ability model was similar: around 0.7. the regression coefficient estimates were close to one (0.99, <2x10-16). The results suggest that a minor allele frequency between 0.0 and 0.05, does not influence the genomic value prediction in small populations. However, these results should be regarded as preliminary and susceptible to changes if the study is repeated with a greater number of markers and animals.
Keywords: Single nucleotide polymorphism; minor allele frequency; genotype quality control; prediction of genetic values; Braunvieh
En el estudio se compararon umbrales de frecuencia del alelo menor (MAF) durante el control de calidad de los genotipos, utilizando polimorfismos de nucleótido simple (SNPs) para predecir valores genómicos de características de crecimiento en bovinos Suizo Europeo. El análisis se realizó con 28,973 y 18,994 registros de pesos al nacimiento y destete, 12,835 polimorfismos de nucleótido simple y 300 animales. Las características se analizaron bajo el enfoque de evaluación genómica en un solo paso. Los criterios de comparación fueron: 1) jerarquización de valores genómicos: se estimaron coeficientes de correlación Spearman y Pearson entre valores genómicos predichos con frecuencia estándar del alelo menor (0.05) y niveles menores; 2) habilidad de predicción de los modelos: se realizó validación cruzada con cuatro repeticiones y se obtuvo el promedio de los coeficientes de correlación de los fenotipos reales contra predichos; y 3) coeficiente de regresión simple: los valores genómicos obtenidos con frecuencia del alelo menor estándar e inferiores, fueron las variables independiente y dependiente, respectivamente. El efecto del nivel de frecuencia del alelo menor en la jerarquización de los valores genómicos no fue importante (correlaciones mayores que 0.99, <2.2x10-16). La habilidad predictiva del modelo fue similar: 0.7 promedio. Los coeficientes de regresión fueron cercanos a uno (0.99, <2x10-16). Los resultados sugieren que una frecuencia del alelo menor entre 0.0 y 0.05, no altera las predicciones de valores genómicos en poblaciones pequeñas. Sin embargo, estos resultados son preliminares y susceptibles a cambios si el estudio se repite con un número de marcadores y animales mayor.
Palabras clave: Polimorfismo de nucleótido simple; frecuencia del alelo menor; control de calidad de genotipos; predicción de valores genómicos; Pardo Suizo
Introduction
The discovery of large numbers of single nucleotide polymorphisms (SNPs) markers; the availability of high throughput technology to genotype animals at thousands of SNPs in a cost-effective manner and advances in statistical methodologies for analysis, helped to generalize the use of genomic selection (GS) in the world (Meuwissen et al., 2016). Currently, the additive effect of molecular markers in the prediction of genetic values (GV) of traits of economic interest in animals is being studied (De los Campos et al., 2013).
The success of exploiting SNP markers in GS lies in achieving accurate predictions of GVs for the traits studied. According to Meuwissen et al. (2016) individuals with reliable phenotypic information and as many SNPs as possible are required. To improve the accuracy of GV predictions, variants of the innovative methodology of (Meuwissen et al., 2001), such as semi-parametric (Gianola et al., 2006) or non-parametric (Gianola et al., 2009) approaches, have been proposed. In addition, a wide number of studies on methodological modifications and contributions have been proposed, such as the use of K-means clustering for cross- validation against the common methodology (Saatchi et al., 2011). Single-step genomic BLUP evaluation (ssGBLUP) is a novel statistical analysis in GS that allows the use of genomic and pedigree information, with superior accuracy compared to other multi-step methods (Wang et al., 2014). The ssGBLUP analysis requires the construction of several matrix arrays: matrix A (of additive genetic relationships), matrix G (of genomic relationships) and matrix H (of relationships between genotyped and non-genotyped animals) (Pértile et al., 2016).
Regardless of the selected statistical methodologies for the analysis of the information, it is required the provision of high-quality information. Anderson et al. (2010) detail the methodology to perform a good genotype quality control (GQC), one of the steps is the elimination of markers with very low Minor Allele Frequency (MAF). MAF is defined as all those alleles, in the DNA sequence, whose frequency ranges from 0.01 to 0.2 (Chouraki & Seshadri, 2014). MAF has also been related to the prevalence of rare genetic diseases (Kido et al., 2018). These authors reported that 50 % of MAF is associated with complex diseases, such as Alzheimer’s disease, diabetes, cancer, and schizophrenia.
In the field of genetic edition, the elimination of MAF is intended to increase the frequency of alleles that have an important effect on the traits of interest. To ensure genotype quality in genetic value predictions, markers with MAF less than 5 % or those up to 1 % are removed. For example, Minozzi et al. (2012) worked on the GQCs of bovine genotypes, eliminating markers with MAF less than 1 %. Anderson et al. (2010) indicated that failure to remove low-frequency alleles results in the use of false information, for two reasons 1) the associations observed in these SNPs are small because they are addressed by the genotypes of few individuals; and 2) they may come from genotyping errors in markers that are actually monomorphic in the population.
For growth traits in cattle, the effect of the MAF threshold on the accuracy of predictions in small populations is not completely clear, although it is understood that increasing genotype quality will also increase the accuracy of GV predictions, even when using high MAF values implies the elimination of several SNP marker information. Therefore, the present research aimed to compare the effect of different MAF thresholds in GQC editing using SNPs for single-stage genomic prediction of growth traits in SwissEuropean cattle.
Material and Methods
Phenotypic information source
Phenotypic information was obtained from the database of the Asociación Mexicana de Criadores de Ganado Suizo de Registro (Mexican Association of Breeders of Swiss Cattle of Registration). The pedigree records, identification and birth (BW) and weaning weights (WW) of the animals were taken. The pedigree consisted of 184,788 animals born between 1901 and 2016.
For the genomic information, hair samples were collected from 300 animals, 236 females and 64 males, randomly selected from cattle ranches located in the states of Colima, Jalisco and Veracruz, Mexico. The samples were sent to GeneSeek (Lincoln, NE, USA) for genotyping. The chip used was Genomic Profile Bovine LD, of 30,000 and 50,000 SNP markers for 150 animals each.
Phenotypes
For BW, records of individuals with a BW value outside the mean ± three standard deviations interval, without herd or maternal age information, were eliminated. Contemporary groups were defined by combining the effects of sex, herd, year and time of birth. Season of birth were defined considering the Julian day: 80-171, spring; 172-264, summer; 265-354, fall; 355-366, winter; and 1-79, winter. Information from animals in contemporary groups with less than two individuals and animals in groups of three or more individuals with zero variance were also deleted. Finally, 28,973 phenotypic records were obtained for further statistical analysis.
For WW, records with inconsistent weaning ages that were three standard deviations above or below the mean, with no information on management, herd, or age of dam, were discarded. Contemporary groups were defined by combining the effects of management, sex, herd, year, and time of birth. Management for WW was defined in three ways: 1) calf feeding with dam’s milk; 2) dam’s milk plus balanced feed; and 3) dam’s milk and a wet nurse plus balanced feed. The season of birth was defined in the same way as for BW, as well as the criterion for elimination within contemporary groups. Finally, 18,994 phenotypic records were obtained for subsequent analysis.
The phenotypes selected for analysis with the ssGBLUP method were from the genotyped animals and their ancestors five generations back. The amount of data on which these analyses were performed is shown in Table 1.
Genotypes
Genotype information was edited as follows: genomic information: SNPs in common were taken from the 30K and 50K chips, yielding 12,835 SNPs from the 300 animals; recoding: additive effects were recoded as AA=0, AB=1 and BB=2; imputation: missing genotypes were imputed within chip size using the marginal distribution samples of marker genotypes; and genotype quality: monomorphic SNPs and with different MAF thresholds were removed. Table 2 shows the numbers of SNP markers available once the MAF thresholds studied were applied.
Calculation of matrix H
The matrix of relationships between genotyped and non-genotyped animals (H) was obtained in a similar way as described by Christensen & Lund (2010) and Legarra et al. (2009). Matrix H, which includes pedigree and genomic information, is defined as:
where Ann, Agn and Agg are submatrices of A containing the relationships between non-genotyped animals, between genotyped with non-genotyped and between genotyped, respectively.
Ga is a fitted relationship matrix obtained from the matrix G described above:
Ga=βG+ and are obtained by solving the system of equations:
In the first equation the average of the diagonal
elements of G is equal to the average of the diagonal elements of Agg. In the second equation the average of all elements of G is equal to the average of all elements of Agg . For more details on the calculation of H consult (Christensen et al., 2012).
Statistical analysis
The BW and WW growth characteristics were analyzed under the one-step genomic analysis approach, modification of the Genomic Best Linear Unbiased Prediction method (GBLUP) proposed by (VanRaden et al., 2009), the most important change being the permutation of matrix G, by matrix H. In matrix notation the model can be written as:
where y is the vector of phenotypes; Z1 and Z2 are incidence matrices relating phenotypic measurements to fixed effects and random effects and ; and is the vector of residual error random effects. The genetic, contemporary group and residual additive variances were assumed:
respectively.
The model for BW (1) included linear (EM) and quadratic (EM2) maternal age covariate fixed effects; the contemporary group (CG) effect was considered random. The model for WW (2) included the same fixed effects as (1) plus that of the covariate age at weaning (ED).
Cross-validation
The predictive ability of the models used was tested by cross-validation. The population was divided into five groups, one of which served as training and the rest as test groups. The cross-validation was repeated four times, with each of the test data groups, and finally the arithmetic mean of the results of each iteration was obtained, obtaining a single result for each analysis.
Criteria for comparison of analyses MAF Effect on ranking of the animals
To determine the effect of MAF on the degree of similarity between animal ranking according to genomic estimated breeding values (GEBVs) for BW and WW, Spearman and Pearson correlation estimators were obtained between the GEBVs obtained with MAF = 0.05 and with the corresponding GEBVs obtained with MAF = 0, 0.02 and 0.04.
Predictive ability of the models
To evaluate the effect of MAF on the predictive ability of the models, a cross-validation was performed. The data were divided into training and validation sets, using the validation set to evaluate the predictive ability of the trained model. Twenty percent of the phenotypic values of the BW and WW growth traits were randomly removed. Subsequently for each model with MAFs of 0, 0.02, 0.04 and 0.05, the GEBVs for those traits were obtained and the average of four replicates of the correlation coefficients of the deleted phenotypes with the predicted ones was obtained.
Simple linear regression coefficient
Simple linear regression coefficient estimators were used to determine the effect of MAF on changes in magnitude of GEBVs, using GEBVs with MAF 0.05 as independent variables and genomic estimated breeding values predicted with MAF 0, 0.02, and 0.04 as dependent variables. The null hypothesis tested was that the regression coefficient estimator was equal to zero.
R software (RStudio Team, 2019) was used for all statistical analyses.
Results and discussion
Ranking of the animals
Spearman’s and Pearson’s correlation coefficients between GEBVs with the analyses considering different MAF thresholds are shown in Table 3. The estimators in both types of correlations were greater than 0.999 (p ≤ 2x10-16), there were minor changes in the animal ranking according to the GEBVs obtained with different thresholds tested. This indicates that the effect of MAF was not significant in the ranking of GEBVs.
Table 3 Estimation of Pearson and Spearman correlation coefficients, in parenthesis significance level for the hypothesis test of estimate equal to zero between genomic values predicted with minor allele frequency (MAF) = 0.05 versus MAF = 0, 0.02, and 0.04 for birth (BW) and weaning (WW) weights in Braunvieh cattle.
| MAF | BW | WW | ||
|---|---|---|---|---|
| Pearson | Spearman | Pearson | Spearman | |
| 0 | 0.99993 (<2×10-16) |
0.99980 (<2×10-16) |
0.99976 (<2.2×10-16) |
0.99960 (<2.2×10-16) |
| 0.02 | 0.99990 (<2×10-16) |
0.99976 (<2×10-16) |
0.99974 (<2.2×10-16) |
0.99957 (<2.2×10-16) |
| 0.04 | 0.99991 (<2×10-16) |
0.99978 (<2×10-16) |
0.99983 (<2.2×10-16) |
0.99968 (<2.2×10-16) |
MAF = threshold for minor allele frequency; BW = birth weight; WW = weaning weight.
These results suggest that for purposes of selecting animals as parents, when evaluations are done using SNP information, any MAF threshold between 0.0 and 0.05 can be used in quality control during genotype editing, although 0.05 is the standard in genomic evaluations performed with animals according to (VanRaden et al., 2009). The accuracies obtained by correlation estimators were slightly higher for BW than for WW, this may be due to the longer time of environmental influence on animals for WW and that may affect the predictions.
In genotyping quality control in Australian HolsteinFriesian cattle, using the Illumina Bovine SNP50TM chip and 798 animals, MAF <2.5 % was eliminated (Hayes et al., 2009). These authors obtained live weight accuracies for GEBVs of 50 to 67 %, similar to those obtained in the present research. Wiggans et al. (2009) eliminated records using MAF less than 2 % in genotypes of 941 Jersey cattle and 344 Braunvieh bulls, the two populations of animals used in these studies cannot be considered large according to VanRaden et al. (2009). These authors suggest using MAF no less than 5 %, although in large populations it can be decreased without affecting accuracy.
Wiggans et al. (2009) argue that the number of animals available per breed in their study meant that their results were considered preliminary, as the sample size did not allow definitive results to be obtained. This may explain the results obtained in the present investigation, as the number of animals used is not considered a large population compared to the population sizes used in other studies (VanRaden et al., 2009).
Predictive ability of the model with different MAF thresholds
Table 4 shows that setting a higher MAF in the quality control in genotype editing improves the prediction accuracy of GEBVs, although the difference in accuracy, for BW, between using MAF = 0.0 versus MAF = 0.05 is very low, 0.0027, and even lower for WW where the difference was 0.0007. This means that, in the GQC of a sample of 300 animals with 12,835 SNPs, the MAF threshold to be used of up to 0.05 is not significant in the prediction of GEBVs for BW and WW.
Table 4 Averages for Pearson correlation coefficient between genomic values predicted analyzing different thresholds for minor allele frequency (MAF) and the actual phenotype values for weights at birth (BW) and at weaning (WW) of four-fold cross validation.
| MAF | BW | WW |
|---|---|---|
| 0 | 0.7130 | 0.6444 |
| 0.02 | 0.7127 | 0.6447 |
| 0.04 | 0.7136 | 0.6451 |
| 0.05 | 0.7136 | 0.6451 |
MAF = threshold for minor allele frequency; BW = birth weight; WW = weaning weight.
Linck & Battey (2019) conducted a study with simulated SNP information, using MAF thresholds between 0.017 and 0.25; these authors reported that using a MAF greater than 0.03 implies a change in the structure of the data distribution due to the drop in the total size of the matrix after filtering and generates an increase in the accuracy of the predictors. This may explain why, in the present study, using MAF = 0.05 for BW implies slightly better predictions.
The predictive ability of the models for BW and WW for each of the thresholds studied does not represent significant changes. The above can be explained by the small number of SNPs removed given the information used compared to the information used in other studies such as that performed by (Zhu et al., 2017) with 82,594 markers by (De la Cruz & Raska, 2014), 40,000 in the study by (Linck & Battey, 2019), and 600,000 in the bird study by (Abdollahi-Arpanahi et al., 2014). These studies reported a significant effect of MAF on genomic prediction of the studied traits. An additional aspect is the low number of animals used in the present study, which does not allow conclusive results. A broader analysis should be carried out using a larger animal population.
Estimator of regression coefficients
Table 5 shows that the regression coefficients are similar to each other; however, the coefficient for WW decreases slightly as the MAF for quality control in genotype editing increases. As the estimators of the regression coefficients are different from unity, the difference in the prediction of GEBVs with the analysis alternatives in comparison is greater (Mäntysaari et al., 2010), all the estimators of the regression coefficients are close to unity consequently the differences between the effects of the MAF thresholds tested are minimal.
Table 5 Estimation of regression coefficients (± SE), independent variables were the genomic values predicted with minor allele frequency (MAF) = 0.05 and dependent variables were the genomic values predicted with MAF = 0, 0.02, and 0.04; in parenthesis significance levels for hypothesis test of estimate equal to zero for birth (BW) and weaning (WW) weights in Braunvieh cattle.
| MAF | BW | WW |
|---|---|---|
| 0 | 0.99756±0.00072 (<2×10-16) |
0.99323 ± 0.00136 (<2×10-16) |
| 0.02 | 0.99605±0.00076 (<2×10-16) |
0.99255 ± 0.00141 (<2×10-16) |
| 0.04 | 0.99578 ± 0.00072 (<2×10-16) |
0.99400±0.00114 (<2×10-16) |
MAF = threshold for minor allele frequency; BW = birth weight; WW = weaning weight.
The results of the present study are similar to those reported by (Edriss et al., 2012), who using information from 5,500 Holstein and Jersey animals with 54,000 SNPs, estimated for fertility, milk protein and mastitis traits, the correlations between observed and predicted values using MAF thresholds: 0.075, 0.275, 0.3, 0.325 and 0.35, their results showed minimal differences, for example, in Holstein the correlations for fertility had a maximum difference of 0.014; concluding that the results of their study suggest that for a small reference population, SNP markers with low MAF do not impair genomic prediction, a statement that also applies to the findings of the present study, but for growth traits.
As reported by VanRaden et al. (2009) and Wiggans et al. (2009), in relation to the number of animals in the population and the number of markers, the results obtained in this study should be taken with caution, and as mentioned by Hayes et al. (2009), they should be considered preliminary and susceptible to changes that may occur in the future when repeating the study with a larger number of animals and markers, since the effect of MAF is due to the drop in the total matrix size after filtering in the GQC, a drop that may not be significant if the amount of information is limited (Linck & Battey, 2019).
Conclusion
The minor allele frequency threshold to use in genomic evaluation could be in the range of 0 to 0.05, without significantly altering the prediction of genomic values, as it does not have a significant effect on the ranking of predicted genomic values and the ability of the model. In this sense, it is concluded that 0.05 may be a recommendable threshold to use in the evaluation of small populations and with a limited number of markers.
REFERENCES
Abdollahi-Arpanahi, R., Nejati-Javaremi, A., Pakdel, A., Moradi-Shahrbabak, M., Morota, G., Valente, B. D., Kranis, A., Rosa, G. J. M. and Gianola, D. (2014). Effect of allele frequencies, effect sizes and number of markers on prediction of quantitative traits in chickens. Journal of Animal Breeding and Genetics, 131(2): 123-133. https://doi.org/10.1111/jbg.12075 [ Links ]
Anderson, C. A., Pettersson, F. H., Clarke, G. M., Cardon, L. R., Morris, A. P. and Zondervan, K. T. (2010). Data quality control in genetic case-control association studies. Nature Protocols, 5, 1564-1573. https://doi.org/10.1038/nprot.2010.116 [ Links ]
Christensen, O. F. & Lund, M. S. (2010). Genomic relationship matrix when some animals are not genotyped. Genetics Selection Evolution, 42(2): 1-8. https://doi.org/10.1186/1297-9686-42-2 [ Links ]
Christensen, O. F., Madsen, P., Nielsen, B., Ostersen, T. and Su, G. (2012). Single-step methods for genomic evaluation in pigs. Animal 6(10): 1565-1571. https://doi.org/10.1017/S1751731112000742 [ Links ]
Chouraki, V. & Seshadri, S. (2014). Chapter five - Genetics of Alzheimer’s disease. Advances in Genetics, 87, 246-278. https://doi.org/10.1016/B978-0-12-800149-3.00005-6 [ Links ]
De la Cruz, O. & Raska, P. (2014). Population structure at different minor allele frequency levels. BMC Proc, 8, S55. https://doi.org/10.1186/1753-6561-8-s1-s55 [ Links ]
De los Campos, G., Hickey, J. M., Pong-Wong, R., Daetwyler, H. D. and Calus, M. P. L. (2013). Whole-genome regression and prediction methods applied to plant and animal breeding. Genetics, 193(2): 327-345. https://doi.org/10.1534/genetics.112.143313 [ Links ]
Edriss, V., Guldbrandtsen, B., Lund, M. S. and Su, G. (2012). Effect of marker-data editing on the accuracy of genomic prediction. Journal of Animal Breeding and Genetics, 130(2): 128-135. https://doi.org/10.1111/j.14390388.2012.01015.x [ Links ]
Gianola, D., de los Campos, G., Hill, W. G., Manfredi, E. and Fernando, R. (2009). Additive Genetic Variability and the Bayesian Alphabet. Genetics, 183(1): 347-363. https://doi.org/10.1534/genetics.109.103952 [ Links ]
Gianola, D., Fernando, R. L. and Stella, A. (2006). Genomic-Assisted Prediction of Genetic Value with Semiparametric Procedures. Genetics, 173(3): 1761-1776. https://doi.org/10.1534/genetics.105.049510 [ Links ]
Hayes, B. J., Bowman, P. J., Chamberlain, A. J. and Goddard, M. E. (2009). Invited review: Genomic selection in dairy cattle: Progress and challenges. Journal of Dairy Science, 92(2): 433-443. https://doi.org/10.3168/jds.2008-1646 [ Links ]
Kido, T., Sikora-Wohlfeld, W., Kawashima, M., Kikuchi, S., Kamatani, N., Patwardhan, A., Chen, R., Sirota, M., Kodama, K., Hadley, D. and Butte, A. J. (2018). Are minor alleles more likely to be risk alleles? BMC Medical Genomics, 11, 3. https://doi.org/10.1186/s12920-018-0322-5 [ Links ]
Legarra, A., Aguilar, I. and Misztal, I. (2009). A relationship matrix including full pedigree and genomic information. Journal of Dairy Science, 92(9): 4656-4663. https://doi.org/10.3168/jds.2009-2061 [ Links ]
Linck, E. & Battey, C. J. (2019). Minor allele frequency thresholds strongly affect population structure inference with genomic data sets. Molecular Ecology Resources, 19(3): 639-647. https://doi.org/10.1111/1755-0998.12995 [ Links ]
Lopez-Cruz, M., Crossa, J., Bonnett, D., Dreisigacker, S., Poland, J., Jannink, J.-L., Singh, R. P., Autrique, E. and De los Campos, G. (2015). Increased prediction accuracy in wheat breeding trials using a marker × environment interaction genomic selection model. Genes|Genomes|Genetics, 5(4): 569-582. https://doi.org/10.1534/g3.114.016097 [ Links ]
Mäntysaari, E., Liu, Z. and VanRaden, P. (2010). Interbull validation test for genomic evaluations. INTERBULL BULLETIN No 41. https://journal.interbull.org/index.php/ib/article/view/1134 [ Links ]
Meuwissen, T. H. E., Hayes, B. J. and Goddard, M. E. (2001). Prediction of total genetic value using genome-wide dense marker maps. Genetics, 157(4), 1819-1829. https://www.genetics.org/content/157/4/1819#:~:text=A%20genome%20of%201000%20cM,marker%20spacing%20of%201%20cM.&text=It%20was%20concluded%20that%20selection,to%20shorten%20the%20generation%20interval [ Links ]
Meuwissen, T., Hayes, B. and Goddard, M. (2016). Genomic selection: A paradigm shift in animal breeding. Animal Frontiers, 6(1): 6-14. https://doi.org/10.2527/af.2016-0002 [ Links ]
Minozzi, G., Williams, J. L., Stella, A., Strozzi, F., Luini, M., Settles, M. L., Taylor, J. F., Whitlock, R. H., Zanella, R. and Neibergs, H. L. (2012). Meta-analysis of two genome-wide association studies of bovine paratuberculosis. PLOS ONE, 7(3): e32578. https://doi.org/10.1371/journal.pone.0032578 [ Links ]
Pértile, S. F. N., Silva, F. F., Salvian, M. and Mourão, G. B. (2016). Seleção e associação genômica ampla para o melhoramento genético animal com uso do método ssGBLUP. Pesquisa Agropecuária Brasileira, 51(10): 1729-1736. https://doi.org/10.1590/s0100-204x2016001000004 [ Links ]
RStudio Team. (2019). RStudio: Integrated Development for R. RStudio, Inc., Boston, MA. https://www.rstudio.com/ [ Links ]
Saatchi, M., McClure, M. C., McKay, S. D., Rolf, M. M., Kim, J., Decker, J. E., Taxis, T. M., Chapple, R. H., Ramey, H. R., Northcutt, S. L., Bauck, S., Woodward, B., Dekkers, J. C. M., Fernando, R. L., Schnabel, R. D., Garrick, D. J. and Taylor, J. F. (2011). Accuracies of genomic breeding values in American Angus beef cattle using K-means clustering for cross-validation. Genetics Selection Evolution, 43, 40. https://doi.org/10.1186/1297-9686-43-40 [ Links ]
VanRaden, P. M., Van Tassell, C. P., Wiggans, G. R., Sonstegard, T. S., Schnabel, R. D., Taylor, J. F. and Schenkel, F. S. (2009). Invited Review: Reliability of genomic predictions for North American Holstein bulls. Journal of Dairy Science, 92(1): 16-24. https://doi.org/10.3168/jds.2008-1514 [ Links ]
Wang, H., Misztal, I., Aguilar, I., Legarra, A., Fernando, R. L., Vitezica, Z., Okimoto, R., Wing, T., Hawken, R. and Muir, W. M. (2014). Genome-wide association mapping including phenotypes from relatives without genotypes in a singlestep (ssGWAS) for 6-week body weight in broiler chickens. Frontiers in Genetics, 5, 134. https://doi.org/10.3389/fgene.2014.00134 [ Links ]
Wiggans, G. R., Sonstegard, T. S., VanRaden, P. M., Matukumalli, L. K., Schnabel, R. D., Taylor, J. F., Schenkel, F. S. and Van Tassell, C. P. (2009). Selection of single-nucleotide polymorphisms and quality of genotypes used in genomic evaluation of dairy cattle in the United States and Canada. Journal of Dairy Science, 92, 3431-3436. https://doi.org/10.3168/jds.2008-1758 [ Links ]
Zhu, B., Zhang, J., Niu, H., Guan, L., Guo, P., Xu, L., Chen, Y., Zhang, L., Gao, H., Gao, X. and Li, J. (2017). Effects of marker density and minor allele frequency on genomic prediction for growth traits in Chinese Simmental beef cattle. Journal of Integrative Agriculture, 16(4): 911-920. https://doi.org/10.1016/s2095-3119(16)61474-0 [ Links ]
Cite this paper Trujano-Chavez, M. Z., Valerio-Hernández, J. E., López-Ordaz, R., Ruíz-Flores, A. (2021). Minor allele frequency in genomic prediction for growth traits in Braunvieh cattle. Revista Bio Ciencias 8, e1052. doi: DOI: https://doi.org/10.15741/revbio.08.e1052
Received: August 29, 2020; Accepted: March 22, 2021
Este es un artículo publicado en acceso abierto bajo una licencia Creative Commons
