Presented by R. Mrode, J. Ojango, Ekine Chinyere, John Gibson and Okeyo Mwai at the Strategic Interest Research Group Meeting on Genetic Improvement of Livestock II, IITA, Ibadan, 2-3 September 2019

1. Innovative digital technology and genomic
approaches to dairy cattle genetic improvement in
developing countries
R. Mrode, J. Ojango, Ekine Chinyere, John Gibson and Okeyo Mwai
Strategic Interest Research Group Meeting on Genetic Improvement of
Livestock II, IITA, Ibadan, 2-3 September 2019

2. Outline
• Innovative and digital tools for data capture
• National level
• Herd level
• Genomic tools and approaches
• Dairy systems based primarily on importation of pure breeds
• Dairy systems based primarily on cross breeding
• Conclusions

3. Structure of most genetic Improvement Program in developed
countries
Data
Collection
and storage
(PI)
Analytical tools (PPP)
Breeding Programs or Delivery
of appropriate genetics to
farmers (PI)
Herd management information
Genetic & Genomic evaluations Farmers
Foreign Genetics
Good Foundation for sustainability or continuous genetic progress

4. Structure of most genetic Improvement Program in developing
countries
Breeding Programs or
Delivery of appropriate
genetics to farmers (PI)
Farmers
Foreign Genetics
Indigenous
breeds or
crosses
No Foundation for sustainability or continuous genetic progress
No Training
No feedback

5. What is ADGG
Platform for African Dairy Genetic Gains (ADGG) is a multi-country,
multi-institution ILRI-led pioneering proof of concept which is:
developing and testing a multi-country genetic
gains platform that uses on-farm performance
information and basic genomic data for identifying
and proving superior crossbred bulls for AI delivery
and planned natural mating for the benefit of
smallholder farmers in Africa

6. ADGG: Approach and Objectives
Innovative application of existing &
emerging technologies
1. To establish National Dairy Performance Recording Centers
(DPRCs) for herd and cow data collection, synthesis, genetic
evaluation and timely farmer-feedbacks
2. To develop & pilot an ICT platform (FFIP) to capture herd, cow
level & other related data & link it to DPRCs (feeds back key
related herd/cow summaries, dairy extension & market info.
etc.).
3. To develop low density genomic chip for breed composition
determination & related bull certification systems for
crossbred bulls

7. Cow
Calendar
AI/Animal Health
Technician
Genomic
chip
Increased
efficiency and
profitability
Platform
Platform

8. 8
Test day records Tanzania: 47,807 records from 11,438 cows
Over 6.2 Million text messages on farmer feedback & training

9. Data collection - Sustainability
• Several options are being explored
• Training individuals to undertake a bundle of services:
• AI services, supply inputs to farmers and data collection or
verification
• Easy to use phone based ICT tools
• farmers can input data. Data is only verified by personnel
providing bundle services occasionally during farm visits.
• well supported with very good feedback that farmers can see
benefits of data collection

10. Innovation and data to optimize productivity:
herd level
• Factors optimizing productivity improve efficiency
achieve genetic potential
• What data and innovation for each stage?
• Calf – Heifer management
• Adequate size & age at first calving between 22 and 24 months.
Growth data & growth profile charts alerts
• Reproductive efficiency
• Cow calendar - Mobile App/digital tool (Icow)
• Heat detection innovations:
• pedometers
• Cow-side, progesterone based lateral flow test for pregnancy diagnosis

11. Innovation and data to optimize productivity:
herd level
• Managing the lactation curve
• Daily milk yield --- ODK, Mobile Apps
• Drop in yield a good indicator of animal facing a disease challenge
- Lactation curve profile - alerts
• Managing Cow replacement
• Milk yield, disease incidence & treatment cost – ODK/ Mobile Apps
• Herd reports with simple ranking simple database
• Heifers brought in are above herd average & analytic tool

12. Quick wins from genomics in dairy systems
in developing countries systems
• With lack of systematic data and pedigree, quick
wins from genotypic data includes
• Reduces the need for accurate pedigree recording as
genomic relationship can easily be computed
• Parentage discovery using SNP data
• Determining breed composition
• The use of few herds with detailed recording for difficult to
measure traits to undertake genomic prediction applicable to
the whole population
• Example :Genomic Information Nucleus (Ginfo) in Australia :103 herds –
fertility traits, diseases (mastitis) traits
• Requirements – herds connected to the whole population

13. Developing a low density SNP Chip for
breed composition
• Low density snp assay can be developed that gives accurate
estimates of dairy proportion

14. Low density assay for breed composition
• Based on DGEA data
• A low density assay of about 200 SNPs have been
identified for determination of breed composition
• If parentage verification is included, assay expands to
400 SNPs
• Piloting its utilization among farmers for certification
of bulls, heifer and cows.

15. Genomics tools in Dairy Systems in
developing countries
• Systems based on importation – purebred
• Genomic predictions to screen heifers for replacement and male
calves as possible breeding bulls on GEBVs
• Quick win: simple tool for converting foreign evaluations to
importing country scale ( which country ? which breed?)
• Steps
• Identification of animals and data collection in local country
• Genotyping of cows using of low density (3K -10k)
• Most sires in exporting countries are genotyped – need for
collaboration to get genotypes for imputation or public databases
• Imputation and genomic prediction (>= 1000 cows)
• Genomic predictions to screen heifers and bulls

16. Genomics tools in Dairy Systems in
developing countries
• Requirements:
• Maintain International identities of imported animals
• Obtain figures of their genetic merit & pedigree in exporting
countries
• Data pooled from several farms operating this system
• Initially simple tool for conversion can be based on
prediction equations involving yield deviations of bull
daughters in local country and foreign proofs of bulls

17. Genomics tools in Dairy Systems in
developing countries
• Systems based on cross breeding
• Optimizing productivity and adaptation
• Interaction between various agro-ecologies and genotype
• Genomic tools
• Determine breed composition - Low density SNP assay
• Genomic prediction
• Genomic algorithms to optimizing productivity and adaptation
• SNP effects
• Selection of crossbred bulls on GEBVs
• For screening foreign bulls in terms of their performance in
crossing in developing countries

18. ADDG example :Genotyping strategy
• GeneSeek Genomic Profiler (GGP) Bovine 50K used
for genotyping:
• 47843 SNPs returned
• About 1500 considered to have private content and not
distributed
• About 5600 animals genotyped from Tanzania using
hair samples
• After QA, 40581 SNPs remained and these were
imputed to HD

19. PCA
First 3 PCs explain 69.96, 5.7 and 2.71 % of the genetic variance
First PC separates exotic dairy from Nellore, with N’Dama about
half way between the two major Taurus groups

20. Average breed proportions of 4614
Tanzanian crossbred samples
Dairy Ayrshire Friesian Guernsey Holstein Jersey N’Dama Nelore
Average 0.785 0.176 0.296 0.089 0.133 0.091 0.04 0.176
SD 0.171 0.173 0.196 0.111 0.177 0.117 0.032 0.143
MAX 0.99998
1
0.99994 0.99994 0.985 0.991 0.99994 0.295 0.902
MIN 0.00005 0.00001 0.00001 0.00001 0.00001 0.00001 0.00001 0.00001

21. Genetic Parameter estimation
• Subset of 1930 cows with 9193 test days records with
genotypes were used
• Cows classified to 4 breed classes based on proportion of
exotic genes : >0.875,0.61-0.875,60-36 , and < 0.36
• G matrix used for all analysis
• Models Examined
• Fixed Regression model : Fixed ward, age nested parity, test-
year-season, fixed curves with Legendre polynomials nested
within breed classes by parity interaction plus random herd
animal and pe

22. Genetic Parameter estimation
• Fixed Regression model + plus dominance as random effect
and D matrix computed from genotypes
• Random regression model with Legendre polynomials of
order 2 for animal and Pe effects
• Bayesians models (Bayes A, B, and C) using YD deviations
from fixed regression model with weights - function on the
number of records each cow has and the variance of YD

23. Genetic parameters
Parameters FRM FRM +
Dominance
RRM
Heritability 0.14 ± 0.04 0.14 ± 0.04 0.26
Variance due to Pe 0.10 ± 0.04 0.08 ± 0.08 0.16
Variance due to herd 0.26 ± 0.02 0.26 ± 0.02 0.24
Variance due to dominance 0.03 ± 0.08

24. Validation
• Cross-classified validation excluding records of a
particular breed class (FRM)
• Forward validation: 254 cows born after 2014

25. Cross –Validation results
Class of
cows
N GBLUP GBLUP +
Dominance
Corr Reg Corr Reg
>0.875 705 0.28 1.7 0.27 1.7
0.61-
0.875
942 0.28 1.6 0.27 1.6
0.36 –
0.60
239 0.38 1.7 0.37 1.8
<0.36 43 0.44 2.6 0.42 2.6

26. Forward validation results
Model
Corr Reg
FRM 0.31 1.2
FRM +
Dominance
0.30 1.2
RRM 0.43 1.1
BayesA 0.11 0.08
BayesB 0.21 0.21
BayesC 0.10 1.4

27. Genomic prediction in Tanzania
Two approaches : GBLUP
• About 2000 cows with genotypes and data
• 530 Males with only genotypes
• 1537 cows with only genotypes
• FRM used for parameter estimation
• Single step
• Data as in GBLUP
• Plus cow with sire or dam or both identified
• 638 cows with 2787 test day records were included

28. Example of some Cow GEBVs
AnimalTag HairsampleID noTDs MeanMilk MeanYD GEBV StdGEBV
TZN000000000001 00001 10 20.400 1.319 3.560 129
TZN000000000002 00002 4 18.375 2.490 2.983 124
TZN000000000003 00003 3 20.833 3.029 2.799 123
TZN000000000004 00004 10 16.400 0.927 2.697 122
TZN000000000005 00005 1 30.000 9.100 2.651 121
TZN000000000006 00006 4 17.500 1.796 2.639 121
TZN000000000007 00007 6 18.333 1.189 2.556 121
TZN000000000008 00008 6 17.667 0.926 2.524 120

29. Optimizing Adaptation and productivity
• Use of frequencies computed across all cross-bred data in
GBLUP tends to produce top cow lists to be dominated with
cows of high degree of exotic genes
• However in G markers are weighted by their expected
variances - solely function of their allele frequencies
• Crossbred data - some animals with >87% exotic genes and
others less than 36%.
• Allele frequencies differ among these categories of animals
• Usually more cows with >50 exotic genes in the data and
dominate allele frequencies estimates.

30. Optimizing Adaptation and productivity
.
• Can we optimize the use of across-breed frequencies in terms
of accuracy? We can examine approaches that
• Standardizes allele frequencies of markers and therefore equalizes
their relative contribution
• Weights markers on the basis of their effects on traits of interest

31. Optimizing Adaptation and productivity
• Examine genomic models attempting to optimize the use of
across breed frequencies with aim of improving accuracies
and ranking.
• GBLUP
• Regular G matrix
• Gstd from standardized allele frequencies
• G0.5 with allele frequencies set to intermediate (0.5)
• Weighted G matrices
• GwtA -- G weighted by SNP effects from BayesA from all data
• GwtA-exo -- weighted with effects from only animals with > 0.65
exotic genes (669 cows)
• GwtA-ind -- weighted with effects from only animals with < 0.65
exotic genes (335)

33. Correlation among allele frequencies for
different cows of different breed proportion.
Categor
y of
cows
No. All > 0.875 0.61-0.875 0.36-0.60 <0.36
All 1034 1.00
>
0.875
304 0.97 1.00
0.61 -
0.875
457 1.00 0.98 1.00
0.36
– 0.60
212 0.96 0.86 0.94 1.00
< 0.36 61 0.85 0.67 0.82 0.95 1.00

34. Accuracy of Genomic Predictions –GBLUP
and standardizing allele frequencies
frequen

35. Accuracy of Genomic Predictions-BayesA
and weighted Analyses

36. • Ranking of indigenous cows in the top 40% increased by 14%
with using SNP effects from indigenous or combined
• Looking at methods that determine breed origin of alleles or
haplotypes for individuals
• Need for broad based index of selection considering
production and fitness traits
Weighted analysis using BayesA

37. • Existing and emerging digital tools offers us the opportunity
to explore innovative data capture cheaply in dairy systems
in developing countries.
• Genomics offers diary systems in developing countries some
quick wins and initial prediction results are encouraging
• Given data structure, fine tuned genomic methodologies
may be needed to optimize productivity and adaptability
• In order to benefit from large numbers and diverse genetic
backgrounds, regional programs pooling data together may
be useful
Conclusions

38. Dairy Farmers & Farmer
organizations
National/regional
Institutions/govts.
Acknowledgements

39. This presentation is licensed for use under the Creative Commons Attribution 4.0 International Licence.
better lives through livestock
ilri.org
ILRI thanks all donors and organizations who globally supported its work through their contributions
to the CGIAR system