1226
INTRODUCTION
In dai y p oduc ion, gene ic change and imp o emen s
in milk p oduc ion pe o mance a e ealized when he
pa en s o he nex gene a ion o animals a e accu a ely
chosen. Fo a dai y he d, his means choosing he si es and
dams on he basis o hei es ima ed gene ic me i o
ma ing o p oduce po en ial eplacemen hei e s ha ha e
high expec ed gene ic me i . The e o e, accu a e es ima ion
o he gene ic me i o dai y animals has long been he
subjec in many s udies and g ea ad ances ha e been made
in he las decades.
In mos opical en i onmen s, milk p oduc ion is an
impo an pa o li es ock a ming (de Leeuw e al., 1999).
As a esul , imp o ing he gene ic po en ial o dai y cows
has been aken as one o he op ions o inc ease milk
p oduc i i y in local he ds. Howe e , ac o s ela ed o lack
o pe o mance eco ding, small he d size, insu icien
a i icial insemina ion se ices and lack o clea ly de ined
b eeding objec i es ha e been he main p oblems (Ke ena
e al., 2011). Fu he mo e, he e is s ill a gap in he use o
ad ances o he es ima ion o he gene ic me i o animals
especially applying mode n me hodologies. Pa icula ly, in
dai y ca le, he selec ion o milk yield in mos coun ies is
based on he use o he adi ional 305-d lac a ion eco ds
(Hammoud and Salem, 2013; Goshu e al., 2014). In his
me hod, incomple e lac a ions o pa lac a ions a e
ex ended o 305-d lea ing a oom o in oduc ion o some
e o s. Mo eo e , he a bi a y s anda diza ion o lac a ion
yields o 305-d and he simple compila ion o es -day (TD)
eco ds in o 305-d lac a ion eco ds, as p ac iced in mos
coun ies, su e s om lack o co ec ion o sho e m
Open Access
Asian Aus alas. J. Anim. Sci.
Vol. 28, No. 9 : 1226-1234 Sep embe 2015
h p://dx.doi.o g/10.5713/ajas.15.0173
www.ajas.in o
pISSN 1011-2367 eISSN 1976-5517
Gene ic Analysis o Milk Yield in Fi s -Lac a ion Hols ein F iesian in E hiopia:
A Lac a ion A e age s Random Reg ession Tes -Day Model Analysis
S. Mese e , B. Tami , G. Geb eyohannes1, M. Lidaue 2, and E. Negussie2,*
Depa men o Animal P oduc ion S udies, CVMA, Addis Ababa Uni e si y, Deb e Zei P.O. Box 34, E hiopia
ABSTRACT: The de elopmen o e ec i e gene ic e alua ions and selec ion o si es equi es accu a e es ima es o gene ic
pa ame e s o all economically impo an ai s in he b eeding goal. The main objec i e o his s udy was o assess he ela i e
pe o mance o he adi ional lac a ion a e age model (LAM) agains he andom eg ession es -day model (RRM) in he es ima ion o
gene ic pa ame e s and p edic ion o b eeding alues o Hols ein F iesian he ds in E hiopia. The da a used consis ed o 6,500 es -day
(TD) eco ds om 800 i s -lac a ion Hols ein F iesian cows ha cal ed be ween 1997 and 2013. Co- a iance componen s we e
es ima ed using he a e age in o ma ion es ic ed maximum likelihood me hod unde single ai animal model. The es ima e o
he i abili y o i s -lac a ion milk yield was 0.30 om LAM whils es ima es om he RRM model anged om 0.17 o 0.29 o he
di e en s ages o lac a ion. Gene ic co ela ions be ween di e en TDs in i s -lac a ion Hols ein F iesian anged om 0.37 o 0.99. The
obse ed gene ic co ela ion was less han uni y be ween milk yields a di e en TDs, which indica ed ha he assump ion o LAM may
no be op imal o accu a e e alua ion o he gene ic me i o animals. A close look a es ima ed b eeding alues om bo h models
showed ha RRM had highe s anda d de ia ion compa ed o LAM indica ing ha he TD model makes e icien u iliza ion o TD
in o ma ion. Co ela ions o b eeding alues be ween models anged om 0.90 o 0.96 o di e en g oup o si es and cows and ma ked
e- ankings we e obse ed in op si es and cows in mo ing om he adi ional LAM o RRM e alua ions. (Key Wo ds: B eeding
Values, Dai y Ca le, Gene ic Pa ame e s, Random Reg ession Tes -Day Model, Tes -Day Reco ds)
Copy igh © 2015 by Asian-Aus alasian Jou nal o Animal Sciences
This is an open-access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion Non-Comme cial License (h p://c ea i ecommons.o g/licenses/by-nc/3.0/),
which pe mi s un es ic ed non-comme cial use, dis ibu ion, and ep oduc ion in any medium, p o ided he o iginal wo k is p ope ly ci ed.
* Co esponding Au ho : E. Negussie. Tel: +358-403-547-208,
E-mail: enyew[email p o ec ed]
1 Minis y o Ag icul u e, Addis Ababa, P.O. Box 62347, E hiopia.
2 Biome ical Gene ics, Na u al Resou ces Ins i u e (LUKE),
31600 Jokioinen, Finland.
Submi ed Feb. 27, 2015; Re ised Ap . 28, 2015; Accep ed May 18, 2015
Mese e e al. (2015) Asian Aus alas. J. Anim. Sci. 28:1226-1234
1227
en i onmen al e ec s (Schae e e al., 2000). Thus, wi h
305-d lac a ion a e age models (LAM) sho e m changes
in en i onmen du ing lac a ion a e usually igno ed, and a
simple he d-yea -season e ec is o en used o accoun o
he a e age o en i onmen al e ec s on each TD (Bilal and
Khan, 2009). In mos cases, he p ojec ion ac o s used o
ex end incomple e o pa lac a ions assume a s anda d
shape o he lac a ion cu e o a cow o a pa icula b eed
and lac a ion numbe . Wi h such an assump ion, cows ha
ha e g ea e pe sis ency could gene ally be unde es ima ed,
whils hose ha a e less pe sis en could be o e es ima ed
which would cause a bias in si e e alua ion (Bilal and Khan,
2009).
Recen ly, eco ds om single and ea ly lac a ion TDs
ha e been used in animal e alua ions which enable a me s
o make an ea lie selec ion decisions (Negussie e al.,
2008). The use o TD eco ds di ec ly as opposed o 305-d
accoun s o all he ac o s a ec ing milk yield on each TD,
which imp o es he accu acy o gene ic e alua ion and
p o ides be e modeling and ex ending o pa lac a ion is
no longe needed. I also a oids he use o ac o s o ex end
pa ial lac a ion eco ds (Wiggans and Godda d, 1996). A
gene ic analysis based on LAM does no u ilize all
in o ma ion in he da a, as i does no allow simul aneous
es ima ion o s age o lac a ion e ec s (Odega d e al.,
2003). E alua ion o he gene ic me i o milk yield may
bene i om analyses based on he TD models.
So a , in E hiopia selec ion o dams and po en ial bull
cal es ha e been based on 305-d lac a ion milk yields.
Meanwhile a dai y he d pe o mance eco ding sys em is
es ablished in E hiopia which will allow gene ic e alua ions
ha u ilizes all TD da a om he ds unde eco ding. So a ,
Geb eyohannes (2013) wo ked on E hiopian mul i b eed
dai y ca le popula ion as i s s ep o he applica ion o
andom eg ession TD model (RRM) o he es ima ion o
gene ic pa ame e s based on TD eco ds. Howe e , s ill
he e is limi ed in o ma ion on he applicabili y o TD
model e alua ion me hods in he opical dia y p oduc ion
sys ems and pa icula ly es ima es o gene ic pa ame e s
and b eeding alues o E hiopian Hols ein F iesian i ing
TD models a e in gene al lacking. The use o accu a e
model de ini ions in gene ic analyses and accu a e es ima es
o pa ame e s con ibu es o inc eased e iciency o
selec ion p og ams. The e o e, he objec i es o his s udy
we e o assess he ela i e pe o mance o he adi ional
lac a ion a e age and he andom eg ession animal model
in es ima ing gene ic pa ame e s and p edic ing he gene ic
me i o Hols ein F iesian in E hiopia.
MATERIALS AND METHODS
Da a
Da a o his s udy consis ed o TD milk yield eco ds o
i s -lac a ion Hols ein F iesian cows ha cal ed om 1997
o 2013 and belonged o wo di e en he ds. The da a we e
ex ac ed om he ecen ly es ablished E hiopian na ional
dai y ca le milk eco ding da abase. Reco ds om all o he
he ds we e s ill oo ew o be included in o he analyses. Fo
his s udy, wo di e en da a se s we e p epa ed and he i s
da a se was he 305-d lac a ion eco ds and he second was
TD eco ds. The s anda d 305-d milk yield o each animal
was es ima ed om TD milk yield eco ds using es
in e al me hod as desc ibed by Sa gen (1968) as ollows:
nn
n1-n
1-n
32
2
21
110
MI
2
MM
I
2
MM
I
2
MM
I+MI yieldmilk d-305






























Whe e,
M1, M2 ... Mn = TD milk yield (kg);
I1, I2 … In-1 = he in e als be ween eco ding da es
(days);
I0 = he in e al be ween he lac a ion pe iod s a da e
and he i s eco ding da e (days) and
In = he in e al be ween he las eco ding da e and he
305 h lac a ion (days).
Wi h he in e al me hod, 305-d lac a ion da a om 800
cows wi h eco ds we e p epa ed wi h an a e age milk yield
o 3,396.9±1,021.7 kg (Table 1). To keep consis ency he
TD da a was also p epa ed by ollowing ce ain da a edi ion
ules. The TD da a was edi ed in such a way ha eco ds
p io o days in milk (DIM) 5 and a e DIM 305 and cows
wi h less han 5 TD eco ds we e excluded o he
es ima ion o gene ic pa ame e s. In addi ion, eco ds o
cows wi h age a cal ing less han 20 mon hs o g ea e han
54 mon hs we e excluded. Age a cal ing was g ouped in o
i e classes (in mon hs). These included cows less han 27,
28 o 33, 34 o 39, 40 o 45 and abo e 45 mon hs o age a
i s cal ing. The cal ing season was di ided in o h ee
Table 1. Desc ip ion o s a is ics o he 305-d lac a ion and es -
day (TD) milk yield da a se s
305-d
lac a ion
TD
Obse a ions
800
6,850
Cows wi h own eco ds
800
800
Si es
149
149
Numbe o animal in he pedig ee
1,779
1,779
Cal ing seasons
3
3
Cal ing yea
17
17
He d es mon h
-
316
A e age milk yield (±SD) (kg)
3,396.9(1021.7)
11.1(3.9)
SD, s anda d de ia ion.
Mese e e al. (2015) Asian Aus alas. J. Anim. Sci. 28:1226-1234
1228
dis inc seasons o long d y (Oc obe o Feb ua y), sho
ainy (Ma ch o May) and long ainy (June o Sep embe ).
The TD da a included 6,850 milk yield eco ds om 800
cows wi h an a e age o 8.5 TD milk eco ds pe cow. The
a e age TD milk yield was 11.1±3.9 kg (Table 1). The inal
da a se used in he s udy included 800 cows which we e
daugh e s o 149 si es. The pedig ee ile con ained 1,779
animals. De ailed desc ip ion o he da a is p esen ed in
Table 1.
Models
Lac a ion a e age model (LAM): The LAM is a single
ai animal model, which is based on 305-d lac a ion milk
yield eco ds. In his s udy, LAM was used o compa e i s
pe o mance agains he RRM in he e alua ion o he
gene ic me i o si es and cows.
The desc ip ion o he LAM used o he analysis o
305-d milk yield was:
Whe e, = lac a ion milk yield eco d;
= ixed e ec o he d;
= ixed e ec o cal ing season;
= ixed e ec o age a cal ing;
= andom e ec s o si e×cal ing yea in e ac ion;
al = andom animal gene ic e ec and
= esidual e ec
Random eg ession es -day model (RRM): The RRM
o he analysis o TD milk yield was selec ed because o i s
abili y o model co ec ly changes in mean and dispe sion
wi h ime (Meye , 2003). In his analysis, he pe manen
en i onmen al and gene ic animal e ec s we e modeled by
Legend e polynomials o o de wo. The main eason o
his was ha p elimina y compa a i e analysis in ol ing
se e al di e en TD models ha e shown ha o he da a
and popula ion in ques ion a RRMs wi h second o de
Legend e polynomial o pe manen en i onmen al and
addi i e gene ic e ec s we e ound o be he bes by mos
model selec ion c i e ia.
In his analysis, he ixed lac a ion cu e o he TD
model on DIM (d) was modeled by a combina ion o
Legend e polynomial and Wilmink unc ion (Wilmink,
1987). Wilmink unc ion wi h exponen ial e m –0.05 along
wi h combina ions o o hogonal Legend e polynomials o
milk yield ai ha e also been used by se e al au ho s
(Lidaue e al., 2003; Negussie e al., 2008; San os e al.,
2013).
The RRM used o he analysis o he TD da a can be
desc ibed as:
Whe e,
ijklmo
y
= milk yield eco ds on TD o;
= ixed e ec o he d;
= ec o wi h ixed eg essions coe icien speci ic o
cal ing season subclass j and measu ed on DIM (d);
= ixed e ec o age a cal ing;
= andom si e×cal ing yea in e ac ion;
= andom he d es mon h e ec ;
= ec o wi h andom pe manen en i onmen al
andom eg ession coe icien s speci ic e ec s o cow m;
= ec o wi h addi i e gene ic andom eg ession
coe icien s speci ic o he animal e ec o cow m and
= esidual e ec
The a iance s uc u e o he andom e ec s o he
model was as ollows:
Whe e, I is he iden i y ma ix, σs2 and σh2 is he
a iance o he andom sy and h m e ec , espec i ely, A is
he ma ix o addi i e gene ic ela ionships among animals,
is he K onecke p oduc , P and G a e co a iance
ma ices o pe manen en i onmen al and addi i e gene ic
e ec s, espec i ely, R is he diagonal ma ix o he o m
Iσe2, and σe2 is he esidual a iance.
Es ima ion o gene ic pa ame e s
Va iance componen s o bo h LAM and RRM we e
es ima ed by A e age In o ma ion Res ic ed Maximum
Likelihood me hod using DMU p og am (Madsen and
Jensen, 2013).
He i abili y
Lac a ion a e age model (LAM): Es ima e o
he i abili y o a 305-d milk yield was calcula ed as a a io
o gene ic a iance ( o o al pheno ypic a iance
( which is he sum o addi i e gene ic ( and esidual
a iances ( .
Mese e e al. (2015) Asian Aus alas. J. Anim. Sci. 28:1226-1234
1229
Random eg ession es -day model (RRM): In he RRM
he addi i e gene ic a iance o DIM d was
es ima ed as:
Whe e, G is he co a iance ma ix o he andom
addi i e gene ic eg ession coe icien and d is DIM.
Simila ly, he pe manen en i onmen al a iance o
DIM d
was es ima ed as:
Whe e, Pe is he co a iance ma ix o he andom
pe manen en i onmen eg ession coe icien and d is DIM.
He i abili y o a pa icula DIM d in lac a ion we e
calcula ed by di iding he es ima ed gene ic a iance
by he sum o pe manen en i onmen al a iances
(
gene ic a iances ( and esidual a iances
( o ha pa icula DIM.
Gene ic and pheno ypic co ela ions: The gene ic
co ela ion be ween wo days in lac a ion and was
calcula ed by di iding he addi i e gene ic co a iance
be ween days and by he p oduc o he squa e oo
o he gene ic a iances o he days and .
Simila ly, he pheno ypic co ela ion was calcula ed
di iding he pheno ypic co a iance be ween days and
, di ided by he p oduc o squa e oo o pheno ypic
a iances o day and .
Whe e, P is he co a iance ma ix o he pheno ypic
eg ession coe icien
Es ima ion o b eeding alue
Fo he es ima ion o b eeding alues o bo h models,
mixed model equa ions we e sol ed by he p econdi ioned
conjuga e g adien me hod wi h i e a ion on da a echniques
as shown in S andén and Lidaue (1999). Solu ions o
addi i e gene ic (â) e ec s we e hen used o o m
es ima ed b eeding alue (EBV) co esponding o 305-d.
Fo he LAM, EBVs o 305-d o animal l was
calcula ed as:
Fo he RRM, EBVs o animal l was calcula ed as:
Compa ison o model pe o mances
The pe o mance o LAM and RRM was compa ed in
e ms o EBVs and e alua ion o he gene ic me i o
b eeding animals. The EBVs om he wo models we e
analyzed o unde s and he ac ual di e ences be ween LAM
and RRM in assessing he gene ic me i o b eeding
animal's. The analysis was done in e ms o s anda d
de ia ion (SD) o EBVs, co ela ion be ween EBVs and
also by assessing he di e ence be ween he models in he
anking o op si es and cow. The analyses o b eeding
alues in ol ed wo g oups o si es and one g oup o cows.
The wo g oups o Hols ein F iesian si es conside ed we e: i)
si es wi h less han 15 daugh e s and ii) si es wi h g ea e o
equal o 15 daugh e s. In he analyses o b eeding alues o
cows o milk yield, cows bo n a e 2007 we e used.
RESULTS AND DISCUSSION
A e age milk yield and es -day p oduc ion ends
The o e all a e age TD milk yield du ing i s -lac a ion
o Hols ein F iesian cows in E hiopian he ds was 11.1
(±3.9) kg. On he o he hand, he a e age milk yield o
305-d calcula ed wi h he in e al me hod was 3,396.9
(±1,021.7) kg (Table 1). P e ious s udies on pa o he
same da a ha e epo ed a lac a ion mean o 3,084 and
3,661 kg by Goshu e al. (2014) and Ayalew (2014),
espec i ely. The sligh di e ence in he mean lac a ion
yield be ween he cu en s udy and he abo e epo s could
be ela ed o he ype and size o da ase , cal ing yea and
he me hods and unc ions used o adjus ing he pheno ypic
305-dmilk yield.
The pheno ypic end o TD milk yield showed ha
du ing he beginning o lac a ion milk yield was lowe and
peaks up immedia ely a DIM 30 o 35. A e peak lac a ion
milk yield showed a g adual bu consis en gen le decline
Mese e e al. (2015) Asian Aus alas. J. Anim. Sci. 28:1226-1234
1230
un il he end o lac a ion pe iod (Figu e 1). A mo e o less
simila end has been epo ed o Hols ein ca le by
Shadpa a and Yazdanshenas (2005) and Abdullahpou e
al. (2010). One in e es ing esul was he peak o he
lac a ion pe iod a ained e y ea ly in lac a ion as compa ed
o o he popula ions. Fo ins ance, in Hols ein cows in
B azil peak lac a ion we e a ained in he second mon h o
lac a ion (San os e al., 2013).
Es ima es o gene ic pa ame e s
He i abili ies: He i abili ies o daily TD milk yields a e
p esen ed in Table 2. The es ima es o he i abili ies o TD
milk yield had mo e o less a simila end wi h addi i e
gene ic a iance ac oss he di e en s ages o lac a ion
(Figu e 2). Es ima es in gene al we e lowe a he beginning
o lac a ion and hen i inc eased consis en ly owa ds he
end o lac a ion be o e s a ing o decline in la e lac a ion.
The main easons o he sligh ly lowe he i abili ies
obse ed a bo h ends o he lac a ion ajec o y could be
due o he highe es ima es o pe manen en i onmen al
e ec s (Figu e 2). Such a end o lowe he i abili y alues
a bo h ends o he lac a ion ajec o y has been epo ed by
D ue e al. (2003) o F ench Hols ein ca le, Negussie e al.
(2008) o No dic Red ca le and Abdullahpou e al. (2010)
and Cobuci e al. (2011) o I anian Hols ein ca le and
B azil Hols ein ca le, espec i ely. Howe e , Cobuci e al.
(2005) and Geb eyohannes (2013) wo king on B azil
Hols ein and E hiopian mul i b eed ca le popula ion,
espec i ely, epo ed an inc easing end o he i abili y o
milk yield om he s a o he end o lac a ion. On he
o he hand, Shadpa a and Yazdanshenas (2005) and
Abdullahpou e al. (2013) ha e epo ed no end o
he i abili y es ima es wo king on i s -lac a ion Hols ein
ca le.
In li e a u e, some di e ences be ween he i abili y
es ima es o LAM and RRM ha e been epo ed. In he
Table 2. Es ima es o pe manen en i onmen al (
), addi i e gene ic ( ), and esidual a iances ( ) and he i abili y (h2) om bo h
lac a ion a e age and andom eg ession es -day models
Model
Days in milk
h2
Random eg ession es -day model
5
5.45
1.6
2.33
0.17
35
3.74
1.28
2.33
0.17
65
2.98
1.21
2.33
0.19
95
2.71
1.3
2.33
0.21
125
2.62
1.46
2.33
0.23
155
2.5
1.61
2.33
0.25
185
2.28
1.71
2.33
0.27
215
2.01
1.74
2.33
0.29
245
1.88
1.69
2.33
0.29
275
2.18
1.56
2.33
0.26
305
3.35
1.41
2.33
0.20
Lac a ion a e age model
305-d milk yield
305-d
-
197,320.6
452,672.9
0.30
Figu e 1. A e age es -day milk yield (kg) a di e en s ages in i s -lac a ion Hols ein F iesian cow.

Mese e e al. (2015) Asian Aus alas. J. Anim. Sci. 28:1226-1234
1231
cu en s udy, he es ima e o i s -lac a ion milk yield om
he LAM was 0.30 whils es ima es om he RRM anged
om 0.17 o 0.29 (Table 2). The es ima e o 0.30 om
LAM is sligh ly highe han he es ima es by Akbaş e al.
(1999), Shadpa a and Yazdanshenas (2005), San os e al.
(2013) and Goshu e al. (2014). In gene al, in he cu en
s udy, he he i abili y o milk yield om LAM was sligh ly
highe han he es ima es om he TD model. S abel and
Szwczkowski (1997) and Kim e al. (2009) epo ed
he i abili y es ima es o milk yield ha we e sligh ly lowe
han es ima es om a compa able TD model whils Akbaşe
al. (1999), Lidaue e al. (2003) and Shadpa a and
Yazdanshenas (2005) o Hols ein and San os e al. (2013)
o Guze a ca le epo ed highe he i abili ies o LAM
han o he TD model. The main easons con ibu ing o
hese di e ences could be di e ences in he da a se , ypes
o unc ions, numbe o obse a ion and da a edi ion c i e ia.
Gene ic and pheno ypic co ela ions: Es ima es o
gene ic and pheno ypic co ela ions o selec ed DIM om
he RRM a e p esen ed in Table 3. Gene ic co ela ions
be ween TD milk yield o i s -lac a ion Hols ein F iesian in
E hiopian he ds anged om 0.37 o 0.99 whils he
es ima es o pheno ypic co ela ions anged om 0.29 o
0.71. In gene al, i was obse ed ha gene ic co ela ions
be ween TD ha we e close o each o he we e highe
compa ed o hose TD ha we e u he apa . The ac ha
he gene ic co ela ions in i s -lac a ion is less han uni y
indica ed ha milk yield a di e en s ages o lac a ion a e
clea ly di e en ai s implying ha hey a e con olled by
di e en se s o genes and should he e o e be ea ed as
di e en ai s. In iew o his ac , he combining o he
di e en TD eco ds in o one single alue as p ac iced wi h
he LAM will lead o a less accu a e e alua ion o he ac ual
gene ic me i o animals o milk yield. The gene al end o
bo h he gene ic and pheno ypic co ela ions obse ed was
in his s udy a e in line wi h he es ima es epo ed by
Table 3. Gene ic (below he diagonal) and pheno ypic co ela ions (abo e he diagonal) be ween selec ed days in milk o Hols ein
F iesian om he andom eg ession es -day model
Days in milk
5
35
65
95
125
155
185
215
245
275
305
5
0.71
0.62
0.52
0.43
0.36
0.32
0.29
0.29
0.3
0.31
35
0.96
0.67
0.61
0.54
0.49
0.45
0.42
0.39
0.36
0.33
65
0.84
0.96
0.66
0.62
0.58
0.55
0.51
0.47
0.41
0.33
95
0.69
0.87
0.97
0.66
0.64
0.61
0.58
0.53
0.45
0.34
125
0.56
0.77
0.92
0.99
0.67
0.65
0.62
0.57
0.48
0.36
155
0.46
0.7
0.87
0.96
0.99
0.67
0.64
0.59
0.51
0.39
185
0.4
0.65
0.83
0.94
0.98
0.99
0.66
0.62
0.55
0.43
215
0.37
0.62
0.81
0.92
0.97
0.99
0.99
0.64
0.59
0.5
245
0.37
0.61
0.8
0.91
0.96
0.98
0.99
0.99
0.63
0.57
275
0.39
0.62
0.8
0.9
0.95
0.97
0.98
0.99
0.99
0.65
305
0.44
0.65
0.81
0.89
0.92
0.94
0.95
0.96
0.97
0.99
Figu e 2. Es ima ed o addi i e gene ic, pe manen en i onmen al and esidual a iances ac oss i s -lac a ion o Hols ein F iesian om
he andom eg ession es -day model.
Mese e e al. (2015) Asian Aus alas. J. Anim. Sci. 28:1226-1234
1232
Lidaue e al. (2003), El Fa o e al. (2008), Negussie e al.
(2008) and Cobuci e al. (2011). Howe e , El Fa o e al.
(2008), wo king on Ca acu ca le popula ion epo ed a
much lowe co ela ion be ween he di e en DIM
pa icula ly owa ds he end o lac a ion. In gene al, he
eason o his kind o end o lowe o nega i e
co ela ions be ween dis an TDs could be due o he ype o
unc ion used o he pauci y o a ailable in o ma ion
owa ds he end o he lac a ion pe iod.
Analyses o b eeding alues
The analysis o EBVs om he wo models showed ha
EBVs om LAM anged om –585 o 686 kg, whils hose
om RRM anged om –680 o 1,109 kg. The SD o EBVs
o i s -lac a ion milk yield om LAM and RRM a e
p esen ed in Table 4. Fo all g oups o si es and cows
analyzed, he SD o EBVs om he RRM was ound o be
highe han LAM. Highe SD o EBVs o TD models
compa ed o LAMs ha e been epo ed o soma ic cell
sco e (Negussie e al., 2006) and milk yield (Lidaue e al.,
2003) o he No dic Red ca le. The inc ease in he SD o
EBVs in mo ing om LAM o a RRM was ela i ely
highe o he si e g oup wi h ≥15 daugh e s han o o he
si e and cow g oups. Negussie e al. (2006) wo king on
bo h models using da a om he No dic Red ca le
concluded ha he inc ease in he SD o EBVs by RRM
o e and abo e ha om LAM could be an indica ion o
be e u iliza ion o in o ma ion in TD eco ds by e ealing
mo e gene ic a ia ion and would enable he selec ion o
supe io si es o cows.
A close look a he es ima es o co ela ions be ween
EBVs om he wo di e en models o he di e en g oups
o si es and cows would enable o judge he magni ude o
changes in animal e alua ions in cases o mo ing om he
adi ional LAM o he RRM. In his espec , he
co ela ions be ween EBVs o he wo models we e
calcula ed o he di e en g oups o si es and cows (Table
5). The esul showed ha co ela ions be ween he EBVs
om wo models anged om 0.90 o 0.96 o he di e en
g oups o animals. In compa ison, he co ela ion be ween
EBVs om LAM and RRM we e sligh ly highe o he
g oups o si es han o he g oup o cows. The co ela ions
be ween EBVs o he g oups o si es we e 0.95 and 0.96
whils o he cows g oup i was 0.90. In gene al, he
co ela ions be ween si es and cows EBVs es ima ed in his
s udy om LAM and RRM was sligh ly highe han hose
epo ed in li e a u e (Lidaue e al., 2003). Basically, a high
co ela ion be ween si es and cows EBVs om any wo
di e en models would indica e equal abili y in e alua ion
o he gene ic me i o b eeding animals, pa icula ly o
si es wi h la ge numbe o daugh e s and cows wi h la ge
numbe o obse a ions. The esul s om his s udy showed
sligh ly lowe co ela ions be ween EBVs om he LAM
and RRM indica ing a possible e- anking and ank changes
o si es and cows. Pa icula ly, he sligh ly lowe
co ela ions o he cows g oups could be explained by he
a ailabili y less in o ma ion o cows compa ed o si es.
To assess he e ec s o he ela i ely lowe co ela ions
be ween he EBVs om he LAM and RRM on e alua ion
o si es and cows, ank changes in he op 20 and 50 si es
and cows g oups we e assessed. The esul ob ained showed
ha he e we e ma ked e- ankings among he op si es and
cows. This is qui e expec ed and is in line wi h he
co ela ions obse ed be ween he wo models. When si es
we e anked wi h espec o hei EBVs om LAM and
RRM, 5 di e en si es appea ed in he op 20 and 6
di e en si es in he op 50. Simila ly, when cows we e
anked, 8 di e en cows appea ed in he op 20 cows and 14
di e en cows in he op 50. The pe cen o si es on he op
20 g oup ha a e common and a e on bo h LAM and RRM
lis s was 75% and 88% o he op 50 g oup o si es. On he
o he hand, o he cows i was 60% o he op 20 cows
g oup and was 72% o he op 50 cows g oup. The esul
om he p esen s udy indica es ha he anking o cows
was much mo e a ec ed in case o mo ing om LAM o
RRM. In gene al, ou esul in e ms o he pe cen o si es
and cows on bo h lis s we e sligh ly lowe han hose
epo ed by Akbaş e al. (1999). Lidaue e al. (2003) also
epo ed negligible di e ence on he anking abili y o he
RRM and LAM o milk yield ai wo king on p oduc ion
ai s using da a om he No dic Red ca le.
CONCLUSION
E alua ion o he gene ic me i o dai y cows equi es
accu a e es ima es o gene ic pa ame e s and bes models
Table 4. S anda d de ia ions o es ima ed b eeding alues (EBVs)
om lac a ion a e age model (LAM) and andom eg ession es -
day model (RRM) o g oup o si es and cows
G oups
Model
No si es/cows
LAM
RRM
Si es wi h
<15 daugh e s
129
125
140
≥15 daugh e s
20
307
349
Cows
Bo n a e 2007
402
243
260
Table 5. Co ela ions be ween es ima ed b eeding alues om
lac a ion a e age and andom eg ession es -day models o
g oups o si es and cows
G oups
Co ela ions
Si es wi h
<15 daugh e s
0.95
≥15 daugh e s
0.96
Cows
Bo n a e 2007
0.90
Mese e e al. (2015) Asian Aus alas. J. Anim. Sci. 28:1226-1234
1233
o he p edic ion o b eeding alues. The es ima es o
gene ic pa ame e s o i s -lac a ion milk yield showed ha
gene ic co ela ions be ween TD milk yields anged om
0.37 o 0.99. This indica ed ha milk yield a di e en TDs
a e indeed di e en ai s. The e o e, combining hem in o
a single 305-d lac a ion yield as p ac iced wi h he
adi ional LAM may lead o bias in he e alua ion o he
gene ic me i o dai y animals. The compa ison be ween he
LAM and RRM showed ha he use o TD models o he
gene ic e alua ion o animals esul ed in mo e e icien use
o a ailable in o ma ion as e idenced wi h he highe SD
EBVs o he di e en g oups o b eeding animals. The
co ela ion be ween b eeding alues om he wo models
was sligh ly lowe and anged om 0.90 o 0.96. This is
he e o e an indica ion ha , depending on he SDs and
co ela ions be ween he EBVs om he wo models, some
changes in he anking o op si es and cows a e expec ed.
IMPLICATIONS
Accu a e e alua ion o he gene ic me i o animals is
he mos impo an s ep in making selec ion decisions. The
compa ison be ween he adi ional LAM and RRM showed
he ela i e pe o mance o hese models in e alua ing he
gene ic me i o animals. The TD model has shown be e
quali ies o e he adi ional LAM. The use o TD models
a oids ex ension o pa o incomple e lac a ions and
p o ides be e co ec ion o sho e m en i onmen al
e ec s. This, by sho ening he gene a ion in e al would
help o maximize gene ic p og ess. The e o e TD models
would be one o he bes op ions o he accu a e e alua ion
o he gene ic me i o dai y cows in opical en i onmen s.
ACKNOWLEDGMENTS
The au ho s would like o acknowledge he E hiopian
Minis y o Ag icul u e, he Na ional A i icial Insemina ion
Cen e o p o iding he da a. The au ho s a e also g a e ul
o he Na u al Resou ces Ins i u e Finland, Depa men o
Biome ical Gene ics o hei suppo in he da a analyses.
Addis Ababa Uni e si y and Wolai a Sodo Uni e si y a e
acknowledged o hei inancial suppo .
REFERENCES
Abdullahpou , R., M. M. Shah babak, A. Neja i-Ja a emi, and R.
V. To shizi. 2010. Gene ic analysis o dai y milk, a and
p o ein pe cen age o I anian i s lac a ion Hols ein ca le.
Wo ld Appl. Sci. J.10:1042-1046.
Abdullahpou , R., M. M. Shah babak, A. Neja i-Ja a emi, R. V.
To shizi, and R. M ode. 2013. Gene ic analysis o milk yield,
a and p o ein con en in Hols ein dai y cows in I an:
Legend e polynomials andom eg ession model applied.
A chi . Tie zuch . 56:497-508.
Akbaş, Y., İ. Kaya, C. Uzmay, Ç. Yakupoğlu, and Y. Ün e . 1999.
B eeding Value Es ima ion o Dai y Ca le Using Tes Day
Milk Yields. Tu kish ag icul u e and na u al scien is s,
Ge man-Tu kish Ag icul u al Resea ch. 6 h Symposium on 27
Sep embe - Oc obe 2, 1999 a he Uni e si y o Giessen,
Giessen, Ge many.
Ayalew, W. 2014. Es ima ion o Gene ic Pa ame e s o P oduc i e
and Rep oduc i e T ai s o Hols ein F iesian Dai y Ca le
Fi ing Mul i a ia e Animal Model a Hole a Bull Dam Fa m,
E hiopia. M.Sc. Thesis. Jimma Uni e si y, Jimma E hiopia.
Bilal, G. and M. S. Khan. 2009. Use o es -day milk yield o
gene ic e alua ion in dai y ca le: A e iew. Pakis an Ve . J. 29:
35-41.
Cobuci, J. A., C. N. Cos a, J. B. Ne o, and A. F.de F ei as. 2011.
Gene ic pa ame e s o milk p oduc ion by using andom
eg ession models wi h di e en al e na i es o ixed
eg ession modeling. R. B as. Zoo ec. 40:557-567.
Cobuci, J. A., R. F. Euclydes, P. S. Lopes, C. N. Cos a, R. A.
To es, and C. S. Pe ei a. 2005. Es ima ion o gene ic
pa ame e s o es -day milk yield in Hols ein cows using a
andom eg ession model. Gene . Mol. Biol. 28:75-83.
deLeeuw, P. N., A. Omo e, S. S aal, and W. Tho pe. 1999. Dai y
p oduc ion sys ems in he opics. In: Smallholde Dai ying in
he T opics (Eds. L. Fal ey and C. Chan alakhana). ILRI
(In e na ional Li es ock Resea ch Ins i u e), Nai obi, Kenya.
pp. 19-44.
D ue , T., F. Ja ezic, D. Boicha d, and V. Duc ocq. 2003.
Modeling lac a ion cu es and es ima ion o gene ic
pa ame e s o i s lac a ion es -day eco ds o F ench
Hols ein cows. J. Dai y Sci. 86:2480-2490.
El Fa o, L., V. L. Ca doso, and L. G. de Albuque que. 2008.
Va iance componen es ima es o es -day milk yield applying
andom eg ession models. Gene . Mol. Biol. 31:665-673.
Geb eyohannes, G. 2013. Gene ic E alua ion o Lac a ion Pa e n
and Milk P oduc ion T ai s o Dai y Ca le in a Mul i B eed
Popula ion in E hiopia. PhD Thesis, Kase sa Uni e si y,
Bangkok, Thailand. 146 Pages.
Goshu, G., H. Singh, K. J. Pe e sson, and N. Lundeheim. 2014.
He i abili y and co ela ion among i s lac a ion ai s in
Hols ein F iesian cows a Hole a Bull Dam S a ion, E hiopia.
In . J. Li es . P od. 5:47-53.
Hammoud, M. H. and M. M. I. Salem. 2013. The gene ic
e alua ion o some i s lac a ion ai s o Hols ein cows in
Egyp . Alex. J. Ag ic. Res. 58:1‐8.
Ke ena, E., W. Zewdie, D. Tadelle, and H. Aynalem. 2011. Gene ic
and en i onmen al ends in he long e m dai y ca le gene ic
imp o emen p og ammes in he cen al opical highlands o
E hiopia. J. Cell Anim. Biol. 5:96-104.
Kim, B., D. Lee, D. Lee, J. Jeon, and J. Lee. 2009. Es ima ion o
gene ic pa ame e s o milk p oduc ion ai s using a andom
eg ession es -day model in Hols ein cows in Ko ea. Asian
Aus alas. J. Anim. Sci. 22:923-930.
Lidaue , M., E. A. Män ysaa i, and I. S andén. 2003. Compa ison
o es -day models o gene ic e alua ion o p oduc ion ai s in
dai y ca le. Li es . P od. Sci. 79:73-86.
Madsen, P. and J. Jensen. 2013. DMU A Package o Analysing
Mul i a ia e Mixed Models. Ve sion 6, elease 5.2. Cen e o
Quan i a i e Gene ics and Genomics Dep . o Molecula
Biology and Gene ics, Uni e si y o Aa hus Resea ch Cen e
Mese e e al. (2015) Asian Aus alas. J. Anim. Sci. 28:1226-1234
1234
Foulum Box 50, 8830 Tjele Denma k.
Meye , K. 2003. Random eg ession models o analysis o
longi udinal da a in animal b eeding. In: P oceedings 54 h
Session o he In e na ional S a is ical Ins i u e, 2003, Be lin,
Ge many.
Negussie, E., I. S anden, and E. A. Män ysaa i. 2008. Gene ic
associa ion o clinical mas i is wi h es -day soma ic cell sco e
and milk yield du ing i s lac a ion o FinnishAy shi e cows. J.
Dai y Sci. 91:1189-1197.
Negussie, E., M. Koi ula, E. A. Män ysaa i, and M. Lidaue . 2006.
Gene ic e alua ion o soma ic cell sco e in dai y ca le
conside ing i s and la e lac a ions as wo di e en bu
co ela ed ai s. J. Anim. B eed. Gene . 123:224-238.
Ø dega d, J., J. Jensen, G. Kleme sdal, P. Madsen, and B.
He ings ad. 2003. Gene ic analysis o soma ic cell sco e in
No wegian ca le using andom eg ession es -day models. J.
Dai y Sci. 86:4103-4114.
San os. D. J. A., M. G. Peixo o, R. R. Bo quis, R. S. Ve neque, J.
C. Pane o, and H. Tonha i. 2013. Compa ison o andom
eg ession models o es ima e gene ic pa ame e s o milk
p oduc ion in Guze a (Bosindicus) cows. Gene . Mol. Biol. 12:
143-153.
Sa gen , F. D., V. H. Ly on, and O. G. Wall. 1968. Tes in e al
me hod o calcula ing dai y he d imp o emen associa ion
eco ds. J. Dai y Sci. 51:170-179.
Schae e , L. R., J. Jam ozik, G. J. Kis emake , and J. Van
Doo maal. 2000. Expe ience wi h a es -day model. J. Dai y
Sci. 83:1135-44.
Shadpa a , A. A. and M. S. Yazdanshenas. 2005. Gene ic
pa ame e s o milk yield and milk a pe cen age es day
eco ds o I anian Hols ein cows. Asian Aus alas. J. Anim. Sci.
18:1231-1236.
S abel, T. and T. Szwczkowski. 1997. Addi i e gene ic and
pe manen en i onmen al a iance componen s o es day
milk ai s in Black-Whi e ca le. Li es . P od. Sci. 48:91-98.
S andén, I. and M. Lidaue . 1999. Sol ing la ge mixed linea
models using p econdi ioned conjuga e g adien i e a ion. J.
Dai y Sci. 82:2779-2787.
Wiggans, G. R. and M. E. Godda d. 1996. A compu a ionally
easible es day model wi h sepa a e i s and la e lac a ion
gene ic e ec s. P oc. 56 h New Zealand Soc. Anim. P od.
Annu.M g., Hamil on, New Zealand. pp. 19-21.
Wilmink, J. B. M. 1987. Adjus men o es -day milk, a and
p o ein yield o age, season and s age o lac a ion. Li es .
P od. Sci. 16:335-348.