Within-herd genetic and 
92–9146–678–7 phenotypic analysis of selected 
production and fertility traits 
The International Livestock Research Institute (ILRI) works to improve food and nutritional security and reduce
poverty in developing co untries through research for efficient, safe and sustainable use of livestock. of multibreed commercial 
Co-hosted by Kenya and Ethiopia, it has regional o r country offices and projects in East, South and
Southeast Asia as well as Central, East, Southern and West Africa. ilri.org  dairy cows in Kenya
  83
CGIAR is a global agricultural research partnership for a food-secure future. Its research is carried
out by 15 research centres  in collaboration with hundreds of partner organizations. cgiar.org
I L R I  R E S E A R C H  R E P O RT    
Within-herd genetic and phenotypic analysis of selected production and fertility traits of multibreed commercial dairy cows in Kenya i
Within-herd genetic and 
phenotypic analysis of selected 
production and fertility traits of 
multibreed commercial dairy 
cows in Kenya 
Chinyere Charlotte Ekine-Dzivenu1, Richard Dooso Oloo1, Raphael Mrode1,2, Julie M.K. Ojango1,  
Evans Ilatsia3 and Ally Mwai Okeyo1 
1. Animal Biosciences, International Livestock Research Institute, Nairobi, Kenya 
2. Animal and Veterinary Science, Scotland Rural College, Edinburgh, United Kingdom 
3. Kenya Agricultural & Livestock Research Organization, Nakuru, Kenya
November 2021
© 2021 International Livestock Research Institute (ILRI) 
ILRI thanks all donors and organizations which globally supports its work through their contributions to the CGIAR Trust Fund
This publication is copyrighted by the International Livestock Research Institute (ILRI). It is licensed for use under the Creative 
Commons Attribution 4.0 International Licence. To view this licence, visit https://creativecommons.org/licenses/by/4.0.  Unless 
otherwise noted, you are free to share (copy and redistribute the material in any medium or format), adapt (remix, transform, and build 
upon the material) for any purpose, even commercially, under the following conditions:
 ATTRIBUTION. The work must be attributed, but not in any way that suggests endorsement by ILRI or the author(s).
NOTICE:
For any reuse or distribution, the licence terms of this work must be made clear to others. 
Any of the above conditions can be waived if permission is obtained from the copyright holder. 
Nothing in this licence impairs or restricts the author’s moral rights. 
Fair dealing and other rights are in no way affected by the above. 
The parts used must not misrepresent the meaning of the publication.  
ILRI would appreciate being sent a copy of any materials in which text, photos etc. have been used.
Editing, design and layout—ILRI Editorial and Publishing Services, Addis Ababa, Ethiopia.
Cover photo—ILRI/K Dhanji
ISBN: 92–9146–678–7
Citation: Ekine-Dzivenu, C.C., Oloo, R. D., Mrode, R., Ojango, J.M.K., Ilatsia, E. and Okeyo, A.M. 2021. Within-herd genetic and phenotypic analysis of 
selected production and fertility traits of multibreed commercial dairy cows in Kenya. ILRI research report 83. Nairobi, Kenya: ILRI.
Patron: Professor Peter C. Doherty AC, FAA, FRS 
Animal scientist, Nobel Prize Laureate for Physiology or Medicine–1996
Box 30709, Nairobi 00100 Kenya ilri.org Box 5689, Addis Ababa, Ethiopia 
Phone  +254 20 422 3000 better lives through livestock Phone +251 11 617 2000 
Fax+254 20 422 3001 Fax +251 11 667 6923 
Email ilri-kenya@cgiar.org ILRI is a CGIAR research centre Email ilri-ethiopia@cgiar.org 
ILRI has offices in East Africa • South Asia • Southeast and East Asia • Southern Africa • West Africa
Within-herd genetic and phenotypic analysis of selected production and fertility traits of multibreed commercial dairy cows in Kenya iii
Contents
Tables             iv
Figures             iv
Acknowledgements           v
Executive summary           vi
1 Introduction            1
2 Materials and methods          2
 2.1 Geographical locations and climatic conditions       2
 2.2 Herd management and crossbreeding strategies      2
 2.3 Phenotype, genotype, pedigree data       2
 2.4 Statistical analysis         3
3 Results and discussion          5
 3.1 Mean phenotypic values for milk yield, age at first calving and calving interval   5
 3.2 Heritability and repeatability estimates        6
 3.3 Genetic merit of animals—estimated breeding value      6
 3.4 Genetic and phenotypic trend over time       7
4 Conclusions and recommendatzion        10
5 References           11
iv Within-herd genetic and phenotypic analysis of selected production and fertility traits of multibreed commercial dairy cows in Kenya 
Tables
Table 1. Least square means (LSM) for age at first calving (AFC), calving interval (CI), and milk yield (MY)  
 by breed class in the Kilifi and Naivasha herds          5
Table 2.  Mean, standard deviation (std-dev), heritability repeatability, estimates of annual phenotypic  
 and genetic changes             7
Table 3. Annual genetic gain for bulls and cows        8
Figures
Figure 1. Per cent of breed presence in crossbred cow population in the Kilifi and Naivasha herds  3
Figure 2. Milk yield of cows by proportion of exotic genes in the Naivasha and Kilifi herds   6
Figure 3. Genetic and phenotypic trend for milk yield in the Kilifi herd      8
Figure 4. Genetic and phenotypic trend for calving interval in the Nakuru herd Naivasha   8
Within-herd genetic and phenotypic analysis of selected production and fertility traits of multibreed commercial dairy cows in Kenya v
Acknowledgements
This report was financially supported by the African Dairy Genetic Gains (ADGG) project (https://africadgg.wordpress.
com/), an International Livestock Research Institute (ILRI)-led investment by the Bill & Melinda Gates Foundation (BMGF) 
that is developing and testing a multi-country genetic gains platform that uses on-farm digitally captured performance 
information and genomic data to identify and provide superior crossbred bulls for artificial insemination (AI) delivery and 
planned natural mating to smallholder farmers in Africa.
The authors would like to acknowledge and thank partners and colleagues in the ADGG project for their support and 
input in the various activities that enabled us collate information from the households in Tanzania. Our gratitude also 
goes to all the livestock keepers within the different regions of Tanzania who took time to answer the questions posed by 
our team. We would like to thank Government of Tanzania and all the National Government Officers who assisted with 
logistics in the villages, and all the performance recording agents (pras) who spent many days working tirelessly with the 
farmers.
vi Within-herd genetic and phenotypic analysis of selected production and fertility traits of multibreed commercial dairy cows in Kenya 
Executive summary
The purpose of this study was to estimate within-herd genetic and phenotypic trends of test day milk yield, age at 
first calving and calving interval in two commercial multibreed dairy herds in Kilifi and Nakuru counties of Kenya and 
to determine the optimum proportion of exotic and indigenous cattle crosses that best suit the different ecologies, 
considering the above traits. Genetic and phenotypic trends are indicative of the changes that have occurred in the herds 
because of selection and/or breed combination practices and the overall respective breeding strategies that have been 
adopted over time.
Lactation and calving records of 436 and 191 cows in both the Kilifi commercial dairy farm and Kenya Agricultural & Livestock 
Research Organization (KALRO’s) Naivasha herd in Nakuru county, were collected from 2000 to 2019 for this study. For milk 
yield in both herds, a fixed regression animal model was used to analyse the data. The model was adjusted for fixed effects of 
year and season of calving, year and season of milking, parity, age nested within parity, days in milk nested within parity and 
degree of exoticness of the cow. Legendre polynomials of order three was used to model fixed regressions of the lactation 
curves. Mean daily milk yield was 15.06±4.92 kg and 7.24±3.62 kg for the Kilifi and Naivasha herds, respectively. 
The model used for analysing age at first calving included degree of exoticness, year and season of birth, and the 
repeatability model fitted to calving interval included year and season of calving, degree of exoticness, and calving age. 
Mean age at first calving was 28.85±4.10 months for the Kilifi herd and 37.98±5.57 months for the Naivasha herd while 
mean calving interval was 447.5±96.86 days and 512.4±170.16 days for the Kilifi and Naivasha herds respectively.
Breed proportion of cows generated from the information provided in the data was used to group the animals into 
three breed classes, based on the proportion of exotic dairy breed into exotic class 1 (EC1) made up of animals having 
0–50% exotic breed proportion (n=481), exotic class 2 (EC2) with animals having between 50 and 75% exotic breed 
composition and exotic class 3 (EC3) with animals having over 75% exotic dairy breed proportion (n=1,009). Adjusted 
means of the breed classes were calculated and contrasted for the different traits. There was no significant difference 
(p < 0.05) between breed types for age at first calving and calving interval, but significant differences were observed 
for milk yield in both herds with EC2 cows having the highest milk production across breed group in both the Kilifi herd 
(9.93±0.26 for EC2 vs 8.24±0.39 and 8.67±0.4 for EC1 and EC3) and Naivasha herd (15.9±0.34 for EC2 vs 15.4±0.4 
and 14.9±0.27 for EC1 and EC3), respectively.  
These results suggest the breed makeup of EC2 cows might be more suitable for sustainable and profitable dairy farming 
under the dairy farming context of both herds since they produced more milk. 
The heritability (h2) estimates for test day milk, age at first caving and calving interval were 0.27±0.06, 0.31 ± 0.13 and 
0.14±0.12, respectively, for the Kilifi data and 0.20±0.09, 0.24±0.27 and 0.03±0.13 for the Naivasha data, respectively. 
Estimated breeding values (with accuracy of prediction) for milk yield, age at first calving and calving interval in the Kilifi herd 
ranged from -6.88 to 4.6 kg (0–88.9%), -3.92 to 5.91months (0–84.8%), -75.5 to 80.7 days (0–71.3%), respectively and 
in the Naivasha herd, it ranged from -3.61 to 1.65 kg (0-81.7%), -6.82 to 4.87 (0-74.6%) and -78.97 to 87.55 (0–72.6%), 
respectively. 
Within-herd genetic and phenotypic analysis of selected production and fertility traits of multibreed commercial dairy cows in Kenya vii
At herd level, phenotypic and genetic trends for milk yield was significantly positive (+0.97 litres per year) and negative 
(-0.26 litres per year) respectively for the Kilifi herd although both trends showed no significant change for the Naivasha 
herd. Phenotypic and genetic trend for age at first calving and calving interval for the Kilifi herd did not show significant 
change over time, and for the Naivasha herd, both trends showed no significant change over time for age at first calving 
although there was a significant favourable negative phenotypic trend for calving interval (-10.54±4.5). No significant 
increase in breeding value was observed for the sire and dam population over time in both herds for these traits.
Despite the favourable phenotypic changes which occurred in milk yield in the Kilifi herd and calving interval in the 
Naivasha herd, genetic trend for all traits analysed did not significantly improve in both herds over time indicating that 
these favourable phenotypic changes were not accompanied by improvement in breeding values for both traits, but likely 
reflect improvement in herd management practices over the past 19 years. 
Genetic improvement for these traits in the herds may be achieved and accelerated by increasing selection pressure on 
them, followed by wide-ranging use of the selected bulls and cows as sires and dams.
Within-herd genetic and phenotypic analysis of selected production and fertility traits of multibreed commercial dairy cows in Kenya 1
1 Introduction 
The fundamental aim of any animal breeding program is to increase productivity of the animals and profitability of the farm 
(Kosgey and Okeyo 2007). Increase in production efficiency could be achieved either by improving the environmental 
conditions of the animals, improving the mean breeding value of the population, or by combining of both (Katkasame 
et al. 1996). Estimating of genetic and phenotypic trends (i.e. computing the change in mean breeding values and 
phenotypic values over the years) indicates the progress made in the herd over time and helps in understanding to what 
extent selection and breeding programs have been effective (Oldenbroek and van der Waaij 2014). It conveys a quick 
assessment of how successful or not a breeder has been in achieving the intended selection goals over the generations. 
Analyses of such trends may reinforce or optimize selection and management goals established by a breeder or even 
signal the need to change/redesign the selection and/or management strategy. A comparison of genetic trends to 
phenotypic trends would help in understanding whether the progress observed could be attributed to the genetic 
improvement of the herd or to improvement in management practices (Javed et al. 2007). Continuous genetic and 
phenotypic evaluation to monitor whether desirable trends are being realized for each trait is important for sustaining 
herd improvement and profitability.
Data generated in two multibreed commercial dairy cattle herds in Kilifi county and Naivasha sub-county in Nakuru county 
of Kenya between 2000 and 2019 has not been genetically evaluated to establish if indeed progress has been made or 
is being achieved. Specifically, this assessment was to obtain estimates of genetic parameters and assess the phenotypic 
and genetic trends for total milk yield, age at first calving and calving interval and to estimate breeding values for 
comparative ranking by genetic merit of all animals in the herd for selecting superior replacement breeding stock within 
a multibreed-herd of cattle in the Kilifi and Nakuru counties of Kenya. The optimum proportion of exotic x indigenous 
dairy crosses suited to the herds for the different traits was explored. Information generated is useful for optimizing herd 
operations and to give direction for future crossbreeding programs for genetic improvement of the herds, especially 
because these two herds are significant sources of breeding bulls and heifers to neighbouring smallholder dairy farmers. 
Progress made in these herds or lack of it, has cascading consequences to the dairy sector in these regions.
2 Within-herd genetic and phenotypic analysis of selected production and fertility traits of multibreed commercial dairy cows in Kenya 
2 Materials and methods
2.1 Geographical locations and climatic 
conditions 
The Naivasha herd is in Nakuru county which lies approximately 90 km northwest of Nairobi at an altitude of 1,829–
2,330m founf above sea level. The region falls within agro-ecological zone UM 5 (Upper Midland-Livestock) and is 
defined as a semi-arid pasture-based zone. Annual mean temperatures range between 18.3 and 19.6°c with an annual 
average rainfall of 650–750 mm. The region receives bimodal rainfall with a main rainy season from March to June and a 
short rainy season from October to December (Jaetzold et al. 2010). The Kilifi herd is in the Kilifi County, which is located 
about 120 km north of Mombasa with a usually hot and humid climate all year round. It falls under the coastal lowland 
semihumid agroecological zone. The mean daily temperature ranges from a minimum of 22°c to a maximum of 29.5°c. It 
has two rainy-seasons with the long rains occurring between April to June, and short rains between October to December. 
The average rainfall ranges from 400 mm in the hinterland to over 1,200 mm along the coastal belt (Hoorweg et al. 1995).
2.2 Herd management and crossbreeding 
strategies
The Naivasha herd is kept under both extensive (controlled) and zero-grazing production systems, with the latter 
including supplementation with high-quality feeds and concentrates. Grading up and rotational crossbreeding was 
practiced in this herd where the respective indigenous cows were crossed with mainly Holstein Friesian bulls through 
artificial insemination. Indigenous and crossbred bulls were also used for mating to achieve the required exotic breed 
persentages. The Kilifi herd is made up of an intermating population of composite cattle originating from crossbred 
parents. It is managed under an extensive system supplemented with natural grass silage, giant panicum and Napier 
grass, maize stover, and fodder trees. The farm produces synthetic lines of cattle, which are continuously maintained 
when interbred, hence stabilizing new composite breeds.
2.3 Phenotype, genotype, pedigree data
The data used in this study was lifted from historical farm records for periods between 2000 and 2019 from two 
multibreed commercial dairy farms located in Kilifi and Nakuru counties of Kenya. Farm records from both herds were 
first lifted from paper copies and converted to digital records before data analysis. The data included a total of 208,634 
milk records and 824 calving records of 436 cows from the Kilifi herd and 10,160 milk records and 490 calving records 
of 191 cows from the Naivasha herd. The data was accessed for quality and cows with missing dates of birth and calving 
dates were excluded from the analysis while outlying data were replaced with trait means. Age at first calving and calving 
Within-herd genetic and phenotypic analysis of selected production and fertility traits of multibreed commercial dairy cows in Kenya 3
interval were calculated from the data received using dates of birth and calving. Age at first calving was defined as age 
from birth to first calving in months and calving interval was defined as the interval between two consecutive calving 
events in days. Age at first calving was restricted to 48 months and parity was grouped into parities 1, 2 and 3+ because 
fewer numbers of cows were observed in later parities. 
The Kilifi pedigree file had a total of 1,015 animals (754 females and 261 males) born between 1992 and 2020 while the 
Naivasha pedigree had a total of 1,496 animals (984 females and 512 males) born between 1985 and 2018. In addition, 
the Naivasha herd had 200 cows genotyped with the GeneSeek Genomic Profiler Bovine 50K chip. These were imputed 
to the Illumina HD chip using a reference population consisting of crossbred cattle from east Africa and several European 
purebred animals (Aliloo et al. 2018). After quality control, 684,183 single nucleotide polymorphism markers (SNPs) 
remained for inclusion in the analysis. Breed proportions of the cows were generated from the information provided by 
the farmers and used to group the animals into three breed classes, based on the proportion of exotic dairy breed. Exotic 
class 1 (EC1) was made up of animals having 0–50% exotic breed proportion, exotic class 2 (EC2) with animals having 
between 50 and 75% exotic breed composition and exotic class 3 (EC3) with animals having over 75% exotic dairy breed 
proportion. Available breed composition data for the Kilifi herd show that Sahiwal, Ayrshire and Holstein breeds have the 
most penetrance in the herd (Figure 1). The breed composition data for the Naivasha herd (Figure 1) showed that most 
of the animals in this herd had their breed information defined. However, due to unplanned mating during grazing, the 
sires of some animals are not known, thus represented as unknown in the chart below (Figure1). The exotic dairy breeds 
included Holstein Friesian, Jersey, Guernsey, Ayrshire, Brown Swiss, Fleckvieh, Milking Shorthorn, Meuse Rhine Issel, 
Scandinavian Redwood and Montbéliarde. Non-exotic or zebu dairy breeds in both populations included Sahiwal, Boran 
and Gir.
Figure 1. Per cent of breed presence in crossbred cow population in the Kilifi and Naivasha herds.
Kilifi herd
Sahiwal
Ayrshire
Holstein
Brown Swiss
Jersey
Gir
Fleckvieh
ScandinavianRed
Meuse-Rhine-Issel
Guernsey
Boran
MilkingShorthorn
Australian Illawarra Shorthorn
montbeliarde
Naivasha herd
Ayshire
Unknown
Holstein Fr iesian
Sahiwal
0 5 10 15 20 25 30
Per cent %
2.4 Statistical analysis
Variance components and corresponding genetic parameters and breeding values were estimated for each trait using the 
Average Information Restricted Maximum Likelihood (AIREML) algorithm of the Wombat program (Meyer 2007) by fitting 
three single trait animal models. The model used for analysing age at first calving included degree of exoticness, year and 
season of birth as fixed effects and animal effect as random while the repeatability model fitted to calving interval included 
fixed effects of year and season of calving, degree of exoticness, parity and age nested within parity with permanent 
Dairy breeds present in both herds
4 Within-herd genetic and phenotypic analysis of selected production and fertility traits of multibreed commercial dairy cows in Kenya 
environment and animal effects fitted as random. For milk yield, a fixed regression animal model was used to analyse the 
data, which was adjusted for fixed effects of year and season of calving, year and season of milking, parity, age nested 
within parity, days in milk nested within parity and degree of exoticness. Legendre polynomials of order three was used to 
model the fixed regressions of lactation curves. The pedigree-derived relationship matrix, A and the pedigree-genomic 
relationship matrix H (pedigree relationships [A] combined with genomic relationship matrix based on SNP information 
[G]) were used to describe covariance among individuals for breeding value estimation in both a classical and single step 
best linear unbiased prediction (BLUP) procedure for the Kilifi and Naivasha herds, respectively.
Estimates for genetic change in the herds and within sire and dam groups over time was obtained by averaging breeding 
values within year of birth and regressing them on years of birth and regressing least square means within year of calving 
on years of calving for phenotypic change.
Adjusted means of the breed classes were calculated and contrasted for the different traits. 
Within-herd genetic and phenotypic analysis of selected production and fertility traits of multibreed commercial dairy cows in Kenya 5
3 Results and discussion
3.1 Mean phenotypic values for milk yield, age at 
first calving and calving interval
The mean daily milk yield was 15.06±4.92 kg, mean age at first calving was 28.85±4.10 months and mean calving 
interval was 447.5±96.86 days for the Kilifi herd and for the Naivasha herd, mean daily milk yield was 7.24±3.62 kg, 
mean age at first calving was 37.98±5.57 months and mean calving interval was 512.4±170.16 days. 
There was no significant difference (p < 0.05) between breed types for age at first calving and calving interval, but 
significant differences were observed for milk yield in both herds with EC2 cows having the highest milk production 
across breed group in both herds (9.93±0.26 for EC2 vs 8.24±0.39 and 8.67±0.4 for EC1 and EC3 cows in the Naivasha 
herd and 15.9±0.34 for EC2 vs 15.4±0.4 and 14.9±0.27 for EC1 and EC3 cows in the Kilifi herd, respectively see Table 1). 
Table 1. Least square means (LSM) for age at first calving (AFC), calving interval (CI), and milk yield (MY) by breed class in the Kilifi 
and Naivasha herds
Breed classes
Herd Trait
LSM±SE
EC1 EC2 EC3
AFC 30±0.59 29.6±0.42 28.5±0.37
Kilifi
CI 445±11.9 437±10.7 450±10.2
MY 15.4±0.4ab 15.9±0.34a 14.9±0.27b
AFC 37±1.25 37.9±0.7 37.1±1.19
Naivasha CI 496±39.3 489±17.1 527±35.3
MY 8.24±0.39a 9.93±0.26b 8.67±0.4a
Note: Exotic class 1 (EC1) was made up of animals having 0–50% exotic breed proportion, exotic class 2 (EC2) with animals having between 50 
and 75% exotic breed composition and exotic class 3 (EC3) with animals having over 75% exotic dairy breed proportion.
These results suggest that since the breed groups did not differ across herds for age at first calving and calving interval; 
EC2, which are cows having between one half and three-quarters of exotic blood in their breed makeup might be more 
suitable for sustainable and profitable dairy farming under the context of both herds since they produced more milk 
(Figure 2). It is valuable for farmers to understand the types of crossbred cows they need to keep, to produce more milk 
for increased profitability.
6 Within-herd genetic and phenotypic analysis of selected production and fertility traits of multibreed commercial dairy cows in Kenya 
Figure 2. Milk yield of cows by proportion of exotic genes in the Naivasha and Kilifi herds.
Naivasha Kilifi
18
ab a
16 b
14
12 b
10 a a
8
6
4
2
0
EC1 EC2 EC3
Cow exotic class
3.2 Heritability and repeatability estimates 
Heritability summarizes how much of the variation in a trait is due to variation in genetic factors and is important in 
animal selection/genetic improvement programs because response to selection depends on it (Bourdon 2000). 
Heritability for milk yield and age at first calving in the Kilifi herd were moderate at 0.27±0.06 and 0.31±0.13, 
respectively, suggesting that breeding values for milk yield and age at first calving have a moderate influence on its 
phenotypic values. Heritability estimates for calving interval was lower at 0.14 ± 0.12. In the Naivasha herd, heritability for 
milk yield was moderate, at 0.20±0.09. Heritability estimates for age at first calving and calving interval were 0.24±0.27 
and 0.03±0.13, respectively. Low heritability estimate for calving interval suggests that non-additive genetic effects such 
as dominance or non-genetic factors such as herd management had a greater influence on this trait as is expected in 
crossbred populations. 
Repeatability is an indicator of changes in traits over time and measures consistency between repeated records for a 
trait and the tendency of animals to maintain their ranking over time (Bourdon 2000). Moderate to high repeatability 
values suggest that the first record on an animal’s performance is a good indicator of its second and subsequent records. 
Thus, an animal’s single performance for a trait is on average a good indicator of its producing ability. In the Kilifi herd, 
repeatability for milk was moderate at 0.45±0.03, and low at 0.14 ± 0.12 for calving interval. An awareness of the 
repeatability of a trait is useful in making culling decisions. In the Naivasha herd, repeatability for milk was moderate at 
0.39±0.03, and low for calving interval at 0.03±0.15. Low repeatability for calving interval might be due to the data 
structure as records of individuals for these repeated records were not consistent across parities. It might also be due to 
changes in herd management and environment over time.
3.3 Genetic merit of animals—estimated breeding 
value
Estimated breeding values (EBV) reflect the genetic potential or transmitting ability of animals. They are estimated for 
individual animals based on own performance records, that of their parents, siblings, and/or progenies (Satpal et al. 
2012) after correcting for various environmental and management factors like season of calving, age at calving, breed, 
Least square means for milk yield (kg)
Within-herd genetic and phenotypic analysis of selected production and fertility traits of multibreed commercial dairy cows in Kenya 7
etc. For a trait such as milk yield, animals with the highest breeding values are genetically superior, while for traits such 
as calving interval and age at first calving, those with least breeding values, are genetically superior and are selected as 
parents for breeding the next generation of animals to ensure faster genetic progress is made in the herd. 
All animals in the pedigree were evaluated and ranked based on their breeding values for daily milk yield, age at first 
calving and calving interval. Superior animals would have high EBVs for milk and low (negative) EBVs for age at first calving 
and calving interval. It is important to consider the accuracy of each EBV during selection as this provides a measure of 
its reliability. The accuracy of an animal’s EBV is the correlation between its true value and the estimated one (Mrode 
2014). The higher the accuracy of an EBV, the lower the likelihood that it will change as more information is gathered and 
analysed for that animal.
EBVs are reported in plus or minus values, in the same units of measurement as the recorded trait and are easy to interpret. 
For example, a bull/cow with +8 breeding value for milk means that on average the daughters of this bull/cow have a 
genetic capacity to produce 4 litres more milk per day compared to the population of daughters of all other bulls in the 
herd. A bull/cow will pass on half of its genes to its calves so its EBVs is halved to estimate the average genetic worth of its 
progeny. The absolute value of an EBV indicates magnitude and the sign indicates direction of the breeding value effect. 
A positive sign indicates increase and a negative sign, decrease. 
When selecting a bull to breed heifers and cows, positive values are preferable in certain situations such as milk yield 
or growth and negative values are preferred for traits such as birth weight or calving interval. Where bulls do not have 
a favourable combination for all traits of interest, corrective mating strategies should be employed by pairing the bulls 
and cows/heifers in such a way that weaknesses in the genetic attributes of either the bull or the cow are corrected or 
compensated for by either party or mate (Allaire 1977), this way, the overall positive breeding goals are achieved.
3.4 Genetic and phenotypic trend over time
One way of monitoring herd progress over time is by evaluating genetic and phenotypic trends for traits of interest. A 
genetic/phenotypic trend measures change in performance over time (typically years) due to change in mean breeding 
value or phenotypic value for the trait of interest in the population (Herbert and Bhatnagar 1988).
Information from trends give direction for future definition of breeding goals to support profitable and sustainable dairy 
business. Assumptions underlying trend analysis is that if animals with superior genetic merit have been used for breeding 
each year, then phenotypic and breeding value averages should increase over time. Regression coefficients of genetic 
and phenotypic trends for the entire population in both herds are represented in Table 2 and trendlines (only significant 
trends) in Figures 3 and 4. Annual genetic change was also calculated separately for bulls and cows in both herds and 
presented in Table 3. 
Table 2. Mean, standard deviation (std-dev), heritability, repeatability, estimates of annual phenotypic and genetic changes
Herd Trait No. Mean Std-dev Heritability Repeatability Δ phen Δ gen
Kilifi MY (kg) 161,322 15.06 4.92  0.27±0.06 0.45±0.03 0.97±0.07*** -0.26±0.02***
AFC 341 28.85 4.10 0.31±0.13 - -0.12±0.07 -0.005±0.02
(months)
CI (days) 295 447.5 96.86 0.14±0.12 0.14±0.12 2.72±1.77 0.67±0.40
Naivasha MY (kg) 10,160 7.24 3.62 0.20±0.09 0.39 ± 0.03 0.29±0.59 -0.008± 0.009
AFC 163 37.98 5.57 0.24±0.27 - 0.14±0.19 0.003±0.007
(months)
CI (days) 216 512.4 170.16 0.03±0.13 0.03 ± 0.15 -10.54±4.5* 0.19±0.58
(Δ), MY-milk yield, AFC- age at first calving and CI calving interval. *, **, ***mean significant at p<0.05, 0.01 and 0.001 respectively.
8 Within-herd genetic and phenotypic analysis of selected production and fertility traits of multibreed commercial dairy cows in Kenya 
Table 3. Annual genetic gain for bulls and cows
Annual genetic gain
Herds Animal
MY (kg) AFC (months) CI (days)
Kilifi Sires 0.07 0.006 0.168
Dams -0.17*** 0.008 0.030
Naivasha Sires -0.006 0.007 0.18
Dams -0.008 0.003 0.21
MY-milk yield, AFC- age at first calving and CI calving interval.  *, **, ***mean significant at p<0.05, 0.01 and 0.001 respectively.
Figure 3. Genetic and phenotypic trend for milk yield in the Kilifi herd.
16 2.5
14 2
1.5
12
1
10 0.5
8 0
6 -0.5
-1
4
-1.5
2 -2
0 -2.5
Years
adjusted average milk phenotype
average milk breeding value
Linear (adjusted average milk phenotype)
Linear  (average milk breeding value)
Figure 4. Genetic and phenotypic trend for calving interval in the Naivasha herd.
250 20
200
15
150
10
100
50 5
0
4 5 8 9 0 0
00 00 00 00 01 01
1 12 13 14 15 16 17 18
-50 2 2 2 2 2 2 20 20 20 20 20 20 20
-5
-100
-150 -10
Years
adjusted calving interval phenotype
average estimated breeding value
Linear (adjusted calving interval phenotype)
Linear  (average estimated breeding value)
Adjusted Phenotypic value for calving interval (days) Adjusted average milk yield (Kg)
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
Average breeding value for milk ( kg)
Average breeding value for calving interval (days)
Within-herd genetic and phenotypic analysis of selected production and fertility traits of multibreed commercial dairy cows in Kenya 9
In the Kilifi herd, the results indicate a significant increase in milk yield over time with a regression coefficient of 0.97 litres/
year and a decline in milk breeding value with regression coefficient of -0.26 litres/year. Figure 1 shows that phenotypic 
and genetic trends for milk yield for the Kilifi herd changed in reverse direction over the course of time suggesting that 
phenotypic increase in milk yield is mostly due to management. Phenotypic and genetic trend for age at first calving and 
calving interval did not show significant change over time. 
In the Naivasha herd, genetic and phenotypic trends for milk yield and age at first calving, showed no significant change 
over time (Table 2) but there was a favourable significant negative phenotypic trend for calving interval (-10.54±4.5) 
suggesting a reduction of about 10 days in calving interval over time in the herd although the genetic trend did not show a 
significant change over time. 
No significant increase in breeding value was observed for sires and dams over time in both herds for these traits. Since 
these are mainly crossbred animals, this result suggests that no significant genetic gains were being achieved among the 
parental population or that sires and dams of similar or worse genetic merit have been used in the herds from year to year.
These results suggest that except for significant phenotypic improvement and decline in milk yield and calving interval, 
in the Kilifi and Nakuru herds, respectively, performance traits considered in this analysis have remained relatively 
unimproved over the period of herd data analysed. This might be an indication that selection pressure on these traits is 
low in the respective herds. However, considering these trends within breed groups might reveal a different picture even 
though this is not feasible in the present analysis due to the low within group sample sizes. 
10 Within-herd genetic and phenotypic analysis of selected production and fertility traits of multibreed commercial dairy cows in Kenya 
4 Conclusions and recommendation
Herd genetic trend showed no significant genetic progress made for milk yield, age at first calving and calving interval 
over the period of study (i.e. period covered by the data used in the analysis for both herds). Also, no significant increase 
in breeding value was observed within sires and dam populations over the same period in both herds.
Routinely calculating of breeding values for both herds is recommended to provide insights for designing optimal 
breeding plans and enable restricted selection of genetically superior breeding parents (bulls and replacement heifers) to 
ensure genetic progress is made in desired directions for all the traits.
While there was no genetic progress for any of the traits in both herds, there was a significant phenotypic increase in 
milk yield in the Kilifi herd and a significant decrease in calving interval in the Naivasha herd. Significant phenotypic 
improvement in these traits outside of genetic progress suggests proper management of the herd. 
Farm records from both herds were first taken from paper copies and converted to digital records before data analysis. 
Digital record management is advised for improved efficiency of record-keeping, analysis and monitoring of animal 
performance. Digital tools eliminate time consumed in searching for historical farm records needed to make decisions and 
readily delivers accurate information needed to improve overall farm productivity.
A more complete delivering farm data management solution would involve integrating a trusted third-party platform 
which acts as a central hub in keeping farm records accessible and sharable, secure, and confidential and produces 
summaries for more efficient farm operations. The African Dairy Genetics Gain program (ADGG - https://www.ilri.org/
research/projects/african-dairy-genetic-gains) has developed a farm management data platform for medium- and large-
scale farms which stores farm data, analyses and generates reports on herd basis and on individual animal basis to support 
data driven decision-making on the farm for increased productivity and profitability. It has validation checks built in and 
can capture errors in spelling, dates and correct them in real time and flag important details which are missing at data 
entry point, thus ensuring high standards of data accuracy and consistency. The database also helps with scheduling and 
assigning tasks and activities on the farm.
Collectng of hair samples and subsequent genotyping of active bulls and their known relatives across herds in Kenya 
to optimize genetic linkages for better prediction and increased accuracy of breeding values is underway. Genotype 
information will help verify pedigrees, and farmer declared breed compositions to inform proper management of the 
animals for optimum extraction of their genetic potential.
Data management support and training will be provided to medium- and large-scale farms in Kenya as a guide on how 
to use the ADGG database and how to conduct genetic evaluation using freely available animal breeding and genetic 
software to determine genetic merit of animals for selection and breeding.
Within-herd genetic and phenotypic analysis of selected production and fertility traits of multibreed commercial dairy cows in Kenya 11
5 References
Aliloo, H., Mrode, R., Okeyo, A.M., Ni, G., Goddard, M.E. and Gibson, J.P. 2018. The feasibility of using low-density 
marker panels for genotype imputation and genomic prediction of crossbred dairy cattle of east Africa. Journal of Dairy 
Science 101(10): 9108–9127.
Allaire, F.R. 1977. Corrective mating methods in context of breeding theory. Journal of Dairy Science 60(11): 1799–1806.
Bourdon, R.M. 2000. Understanding animal breeding. 2nd ed. Hoboken, New Jersey, USA: Prentice Hall.
Herbert, S. and Bhatnagar, D.S. 1988. Genetic trends of economic traits in dairy cattle: A review. Agricultural Reviews 
9(4): 200–216.
Hoorweg, J., Foeken, D. and Klaver, W. 1995. Seasons and nutrition at the Kenya Coast. The Hague, the Netherlands: 
African Studies Centre: Avebury.
Jaetzold, R., Schmidt, H., Hornetz, B. and Shisanya, C. 2010. Farm management handbook of Kenya: Subpart B1a, 
Southern Rift Valley Province. 2010, II, 213.
Javed, K., Babar, M.E. and Abdullah, M. 2007. Within-herd phenotypic and genetic trend lines for milk yield in Holstein-
Friesian dairy cows. Journal of Cell and Animal Biology 1(4): 066–070. https://d-nb.info/1097289923/34
Katkasame, S., Tumwasorn, S., Thaninndratarn, B. and Pasanpanich, S. 1996. Trend analysis on milk production traits in 
the dairy farming promotion organization of Thailand. Agriculture and Natural Resources 30(2): 211–219.
Kosgey, I.S. and Okeyo, A.M. 2007. Genetic improvement of small ruminants in low-input, smallholder production 
systems: Technical and infrastructural issues. Small Ruminant Research 70(1): 76–88.
Meyer, K. 2007. WOMBAT—A tool for mixed model analyses in quantitative genetics by restricted maximum likelihood 
(REML). Journal of Zhejiang University Science B 8(11): 815–821.
Mrode, R.A. 2014. Linear models for the prediction of animal breeding values. Cabi.
Oldenbroek, K. and van der Waaij, L. 2014. Textbook animal breeding: Animal breeding and genetics for BSc students. 
Wageningen, the Netherlands: Centre for Genetic Resources and Animal Breeding and Genomics Group, 
Wageningen University and Research Centre. 
Satpal, D., Sivalingam, J. and Dinesh, K. 2012. Chisinau, Republic of Moldova: Genomic and phenomic tools for livestock 
genome analysis. Lap Lambert Academic Publishing GmbH KG.
Within-herd genetic and 
92–9146–678–7 phenotypic analysis of selected 
production and fertility traits 
The International Livestock Research Institute (ILRI) works to improve food and nutritional security and reduce
poverty in developing co untries through research for efficient, safe and sustainable use of livestock. of multibreed commercial 
Co-hosted by Kenya and Ethiopia, it has regional o r country offices and projects in East, South and
Southeast Asia as well as Central, East, Southern and West Africa. ilri.org  dairy cows in Kenya
  83
CGIAR is a global agricultural research partnership for a food-secure future. Its research is carried
out by 15 research centres  in collaboration with hundreds of partner organizations. cgiar.org
I L R I  R E S E A R C H  R E P O RT