i’., JfwD$‘,&$8 ESTIMATES OF DIRECT,...
i’., JfwD$‘,&$8
ESTIMATES OF DIRECT, MATERNAL AND GRANDMATERNAL GENETIC EFFECTS
FOR GROWTH TRAITS IN GOBRA CATTLE
ABSTRACT
Estimates of genetic parameters for birth (N = 3909), weaning (N = 3425), yearling (N = 2764) and final (N = 2144) weights
were obtained from the records of Gobra cattle collected at the Centre de Recherches Zootechniques de Dahra, Senegal.
Three animal models were fitted to obtain estimates by REML using an average information (Al) approach. Mode1 1 consid-
ered random direct, materna1 genetic and maternai permanent environmental effects. In mode1 2, a general grandmaternal
effect was added to the random effects considered in mode1 1, and in mode1 3, the general grandmaternal effect was divided
into grandmaternal genetic and grandmaternal permanent environmental effects. All models allowed covariances among
genetic effects. The inclusion of grandmaternal effects in models 2 and 3 did not change the estimates of the genetic param-
eters compared to mode1 1. Variantes attributable to grandmaternal effects became negative and were set close to zero, except
for yearling weight for which grandmaternal heritability was 0.03 + 0.03. The estimates for direct and materna1 heritabilities were,
respectively, 0.08 + 0.03 and 0.03 k 0.02 for birth, 0.20 k 0.05 and 0.21 + 0.05 for weaning, 0.26 + 0.07 and 0.16 + 0.07 for yearling
and 0.14 rfr 0.06 and 0.16 i 0.06 for final weights. The estimates of the genetic correlation between direct and materna1 effects for
birth, weaning, yearling and final weights were -0.17 f 0.40, -0.58 f 0.32, -0.52 k 0.34 and -0.34 f 0.37, respectively. For yearling
weight with grandmaternal heritability estimated to be only 0.03, mode1 3 gave estimates of the genetic correlation between direct
and grandmaternal effects and between materna1 and grandmaternal effects of 0.28 f 0.48 and -0.33 f 0.67, respectively.
Estimates of direct and materna1 heritabilities were unchanged when grandmaternal effects were not included in the model.
INTRODUCTION
out that such models may be suboptimal and need to be
improved. The objective of this study was to evaluate the
Growth traits of beef cattle are known to be influ-
influence of including grandmaternal effects in a mode1
enced by materna1 effects. Early reports (Koch and Clark,
for estimating genetic parameters for growth traits in Gobra
1955: Willham, 1963) mentioned the possible influence
cattle for which materna1 effects have previously been
of grandmaternal effects. Falconer ( 1965) considered the
shown to be important.
materna1 effect as a linear function of a mother’s pheno-
typic value intluenced by a11 materna1 ancestors. The so-
MATERIAL AND METHODS
called ‘fatty udder syndrome’ is an example of a grand-
materna1 effect: a granddam with a high materna] ability
The records of weight at birth, 6, 12, and 18 months
may over-feed her daughter and thereby have a negative
were obtained from the Gobra herd at the Centre de
intluence on her daughter’s materna1 ability by inhibiting
Recherches Zootechniques de Dahra, Senegal. The pro-
development of her udder tissue (Totusek ef ~1.. 1971).
duction environment is characterized by a dry tropical cli-
The importance of materna1 effects in beef cattle has been
mate with two distinct seasons: a long dry season from
widcly reported (Koch. 1972; Baker, 1980: Robison, 198 1;
October to June and a short rainy season from July to Sep-
Meyer, 1992). Studies on Af’rican cattle populations have
tember. The mean annual rainfall from 1964 to 1988 was
reported a strong materna1 influence on growth traits
360 mm. The mean annual temperature is 28°C. The veg-
(Tawah cf ~11.. 199.3: Khombe Côt ~11.. 1995: Diop and Van
etation is described as savanna type dominated by Acaciu
Vleck. 1998).
sp. and annual grasses. Annual biomass production is
Animal models used to estimate materna1 effects
closely related to the amount of rainfall the area receives
typically include direct and materna1 genetic effects and
each year.
rnaternal permanent environmental effects and consider the
Native pasture is the main source of feed. The quan-
covariance between genetic effects. Meyer ( 1992) pointed
tity and quality of the pasture vary considerably during the
year. During the rainy season, pastures are of good quality
and abundant. During the dry season. the nutritive value
of the forage decreases and supplemental feeding with
ground nutcake or cottonseed is provided. especially for
suckling cows and weaned calves.
Breeding females were randomly assigned to sires
(30-50 cows/sire) for a breeding season from December
to Mat-ch. Cows that were not pregnant 3 months after the
breeding season started were reassigned to a different bull.
.
364
Diop <‘/ tri.
Over the years, thc si/e of’ III~ hcrd Iras Iluctua~cd around
300 CO~S. Males and fèrnales wcrc sclec~ed bas~cl ou
weight at 6 months (wraning). Malca werc again sclccted
based on their wcight a( 1 X months aftcr which the IO bcsr
males underwcnl ii growth pcrt’ormancc
(est hct’orc~ final
selection for the two or thrce bulls to bc uscd as rcplacc-
ment bulls. Replacemen~ I’cIllalcs werc sclcctcd a( 24
months based on Ihcir wcighl. About 5’4 of thc malch and
80% of the femalcs selectcd aficr wcaning \\vcrc LIWI as
replacements.
Culling of thc c’ows was bnscd on polar rc-
productive performance (ion,0 calving iiiler\\,al OI’ failurc
to calve after two breeding aensons~ or pour pwth pcr-
formance of offspring.
The data for the present anaIysc\\ consistcd of the
records of animais h-n I’ronI 1963 IO 19X9. C’cmG~ttwcy
rcspccti\\.cIy. a1111 N ix thc numbcr of records. A is Ihc nu-
checks werc performcd upon tho idcntifica~ion I)I‘ ihc ani-
mcrator rclationshi p rnali is anlong aninlals in the pedi-
mals and their pedigrees. Records of thc progcny OI sires
frcc Iï Ic. ancl thc 1 malriccs arc identily matrices of speci-
with less than five pro~eny wcrc delctcd from I~IC ;maly-
Iïcd ordcr. Modclx I und 3 rcquirc cach iminltll with il record
sis. The PROC MIXEIl) option of SAS (SAS. 1093) was
lo havc a trandtl;~ln. Thcrclore. “du1~111y” granddam identi-
used to lest the significancc of thc fixed cfli’c~~ 01’ ~ncmth
Iïc~alic>ll \\\\‘a$ ;(S\\i cncd I~)I- records \\\\rith unkno\\\\sn granddam
of birth, year of birth. scx and parity in a modcl with Grc
in@;inp thc assomption lh;lI ~hcsc ~randdams are unrelated
considered to be a random cffcct.
;lnll~ll~ ~hcIll~t!~~~~IS il1111 1lnlL’bX~ t0 thosc \\Vith know’n idell-
For estimation of thc gcnctic parmnctc~~s. thrcc ani-
lilïcalion Tilljlt’ 1 sui~ini;iri/cs thC d3til LlSCd in thC analyses.
mal models wcrc fitted for the four traits consitlcrc4 :
i’omponcnts of \\ ;Iri;\\llcc were cstimated b> re-
strictcd ~naximunl likelihood tREM1.) usin? a n average
inli)rnialicm alyjrirhnl (.lohns~~n XI~ Thompson. 19%). The
Mode] 2: y = Xp + Z, a + ZI 111 + Cc’, ç + \\?‘? y +- c
alyorithin is a Newton ~nc~hoci w h i c h LI~~S first and WC-
Mode] 3: y = Xp + Z, a + Zz m + x3 g + \\Y, c + \\Y? q + e
ond rlci i\\.alivc\\ of thc iogilrilhm 01’ the likclihood given
thc tlat;l t0 iïnd cstirllntc~~ or \\ iliiilWC Cl~~llpO~lC~ltS
ht
where y is an N x I vector of ohsl‘rvafions,
p is lhc \\ cc101
lllilXiIlli/r~ lhc likitli hood l’unclion.
of fixed effects of ycar of birth. month of birth. WL and
A pclinlillilry
ilKll?\\iS OI‘ \\~ilriallcc ShoWed that the
parity of dam. a, m and g iII’C \\:eclors of hrcc‘tliirg \\ alues
fixcd cl‘li.*~*l\\ of iiioiilh and !c;lr of birth. scx. illld parity
for direct. malcrnal and LLrilnC11ii;ltcl.Il;II gcncilic ~~~I~~CI~, ç
wcrc 4gnil‘ic.ant. (‘c)tlxc’titlt’tlllv. ~hc\\e eff’cc~s wcre included
and q are vectors of matcrnal pcrmancnt
cn\\,ironiircnlal
in ail thrcc nll.ltlcls.
and grandmaternal
permanenl
en\\‘ironrncnt~~~ cl’fticlb and
f:xlim;itc\\ of rcIali\\‘c’ \\ arianccs attributaI.+ to ad-
e is an N x 1 vector of rcsidual random cffccis. ;IIKI X, %,,
dilivc ci~rect #~II:). nlaternal th,:.) and _«randmaterna1 (11;)
Z,, Z,, W, and W, are known incidence matriccx rclating
(~cnctic Cl’lCcts \\‘1crc cillc~ll;ltctl a4 riltiC)s of the esiimates
observations to their respective fixed and randonl cll&..~s.
Of adltil i\\re diret.t CC?). nlatcrnill gcnctic (of,,) and grand-
Matrices Z,, Z2 and Z, were augmenled for animais with-
matcrnal genctic. c$) varianccs. rcspectively. to the phe-
out records that were included in Ihe relalionship matrix.
notypic varimlcc CC$>. Thc cliI.cct-mnternal (ri,,,)). direct-
For the most completc mode1 (model 3). E (y) =
grandma~erna1 (r,?) and matcln;1l-=randmaternal
(r,,,<) ge-
Xp and
nctic correlalionx were estimated as ratios of the direct-
nlaternirl (<r,,,,,), I,lirccr-eranclnlrlternal (o,~J and maternal-
a
grandrnaternal (cF,,,,) genetic covariances t6 the square roots
of ~hc products of CJ~~ and of,, 0: and 05, and oi,. and 0:.
m I
respcctivcly. ‘Thc estimalcs of‘ relative variantes attribut-
ahle to nlatcrnal pcrnlancnt environmental effects (c’) and
g
<rr;iiidiil;ilcrii~it
pcrnianenr en~~ii’oiimenta1 effects (q’) were
V
=
ZillCllliltCd ;lb th! ratios of thC CstillliltêS of maternal (0:)
C
illld ~r~indiiialcrii31
((ri) permanent cnvironmental variancés
to thc phcnotypic variancc ($1.
RESULTS
e
The \\2riilncc components iltld estimates of genetic
where Nd and Ng are numbers 01‘ dams :III~ grand(ia~ns.
parainclcrs
(IdiOS aiitl correlations)
for birth weight are
prehcnted in Table II. The direct (hi) and maternai (hf,, )
Thc cstimates for yearling ueight (Table III)
heritabilities were 0.08 -t 0.03 and 0.03 k 0.02. respec-
sl~owctl that the inclusion of grandmatcrnal genetic effects
tively. for the three modela. The estimates were similar for
in model 3 slightly incrcased thc additive direct genetic
a11 effects included in a11 models. The cstimates of vuri-
v:ukmce wirh an estimate of grandmuternal genetic heri-
ances attributable to grandmatcrnal Fenctic and/or grimd-
tahility (hi‘) of 0.03 + 0.03. The estimated values of the
materna1 permanent environmentu1 effkcts werc close to
<Fcnctic correlations bctwern the direci 2nd grandmater-
2
zero in models 2 and 3. The estimate of the corrclation
nal genetic effects (r,_) and between the materna1 and
between direct and materna1 genetic effects L;IS -0. I7 +
grandmaterna cl‘fects i r,,,?) were small to moderate with
0.41 . The estimate of& was 0.0-i -t 0.02 for ail modelh. As
large srandard errors (O.% rt 0.38 and -0.33 tr 0.67, re-
expected from the estimates ashociated with grandtmtter-
spcctivcly). Thc eatimates of variante components attrib-
na1 effects. the three models resulted in c\\scntiidly the siune
utahlc to direct and marna1 effects and their correlation
value for the log likelihood.
uc‘re similar for the three rnodels. The estimates of hi
The pattern of the estimates for weaning wright
and hf,, were 0.36 k 0.07 2nd 0.16 + 0.07, respectively.
(Table II) was similar to tllat pattcnl of’ estimates l(n
Thc cstimate of r.,,,, was -05.5 IL 0.37 for mode1 3. The
birth weipht. i.e., the estimates of thc variance and co-
cstimates of c’ ~+ercf 0.09 k 0.03 for mode1 1 and 0.07 t-
variante components for includcd effects were 01‘ the
0.0-l for modcl 3.
same magnitude in the different ~~~odels. Variancc\\ at-
Thr estimatc of bariance for final weight attribut-
tributable to grandmaternal cffects [vert: small. with
ahlc to ~landtnatcrnnl el’fccts wx nrar zero (Tilble III).
values near zero. The estimatcs of direct nnd maternai
The estimates of direct (hi) and maternai (ht),) heritabilities
heritabilities were 0.20 $r 0.05 and 0.2 1 rt 0.05. rcspec-
wcrc O.I3 31 0.06 and 0. I6 + 0.06, respectively. The esti-
tively. The direct-materna1 genetic corrclation ~~1s -0.58
mate of c’ W;IS 0.04 $r 0.05. Thct correlation between the
IL 0.33 and c? wus 0.15 4 0.0-I.
direct and maternai genztic effects was -0.34 * 0.37.
4.
.
366
Table 111 - Estimates of çompon~~ts of variancc and gcnetic paramctcr\\
for )ç;~rlin~
and final weights for Gohru cxttle.--_
-
-
Yearling weight
Final wcirhl
-
-
-
-
ParameteW
Mode1 1
Model 2
Mode1 3
Modcl I
Moclcl 2
hlodel
3
s
169.7
169.4
171.3
I20.3
129.1
120.1
cl:,
10X.7
106.0
107.X
144.9
110.5 IN.5
7:
2 0 . 8
0.0
CT .ll/>
-70.4
-70.0
-76.6
-46.7
45.3
-15.3
cT.,g
16.X
0.0
on1p
-15.4
0.0
ls,f
5 7 . 1
53.2
4x.5
42.9
3.5.
I
35.2
s
5 . 2
0 . 0
10.3 10.7
0;
39X.4
398.5
394. I
6 3 Z . 0
632.3
633.1
0,:
6635
663.2
663.2
w3.3
~~01.0 902.4
11;
0.26 k 0.07
0.26 k 0.07
0.26 f 0.07
0.13 !: 0.06
0.14 7t O.Oh 0.11 i 0 . 0 6
hi,
0.16 k 0.05
0.16 + 0.0
0.16 f 0.07
0.16 -r 0.06
0.16 t 0.06 0.16 t O.Oh
h;
0.03 Ek 0.03
0.00 k 0.00
r.,,.
-0.52 k 0.34
-0.52 + 0.35
-o.ss * 0.37
-0.3-I i 0.27
- 0 . 7 4 -f 0 . i7 -0.34 k 0..37
r.lis
0.28 k 0.48
0.(10 i 0.00
hg
-0.33 k 0.67
0.00 r 0.00
$
0.09 f 0.03
0.08 f 0.04
0.07 Ii 0.04
0.05 + 0.04
0.03 i 0.0s 0.04 d. 0.05
Y?
0.01 -I 0.02
0.00 i 0.00
O.OI zk 0.03 O.OI f 0.03
-2 log L”
2033s.2
20.135. I
20334.2
IOJ4.5.6
I h.J-li.4
I h-115.-I
-_--.-
%si. Direct genctic variante:
o:,, materna1 genetic variante:
0:. gnuntlmnlcrnnl
selletic \\ ;ui;mcc: (T,,?. rlirec~l and malcr-
nal genetic covariance: CT,,,,, direct and grdmatunül genctlrc covariancc:
<T,,,L. m;~tcrnal and ui-;tndiiiatcr1lal
gcnctk
r
covariance:
of, maternul permanent environmentid
variancc: a. ~r;intlln;ltcrn;II
pcrm;mcnt
cm ilrulmcnt;ll
\\ ariaicx:
0:. temporary
environmental variante:
$, phcnotypic
variancc:
ha. clirccl hcritahility:
II,:.. nlalt’mal
hcrilahilit!:
h:
grandmaternal heritability; r,,,,,, direct-malcrnal
gcnetic corrclntion:
r,,_. tlirCCt-~r;llltliIl;ltCrIl:Il
pcnclic ckWrcl;ition:
r
/on
malemal-frandinaternal
genctic corr&tion:
c’= a’io;: cl?= c$k$,. “-2 lop 1. = Minus 1% icc tlic Io? ~1 lihclih(~~d $\\cn thc tlat;l.
DISCUSSION
MacNeil. M.D.. and Van VIe&. I,.D., unpublished results).
For yearling wcight. thcre was a small. positive
The inclusion ofgrandmaternal effects in the mode1
correlation bctwccn direct and ~randmaternal genetic ef-
did not change the estimates of the direct and materna1
fects (r,,r) and a modcrate. negative correlation betwYen
heritabilities or of the genetic correlation between the di-
the mat&nal ancl grandmaternal gcnetic effects (r,,iP). The
rect and materna1 genetic effects for the four traits ana-
standard en-ors associated Lvith thcse estimates uerk large
lyzed. However, for yearling weight. the inclusion of grand-
which means they arc not significantly different from zero.
materna1 genetic effects in mode1 3 slightly increased the
Dodenhoff et ol. (1998) rcported small to large positive
additive direct genetic variante. The estimates of variante
estimates of rZ,E at weaning for differcnt lines of Hereford
attributable to grandmaternal effects were very small or
cattle. The estimates for r,,,, were negative and large for
zero, suggesting that these effects did not play an impor-
weaning wcight in these Ii&. The negative correlation
tant role in these traits in this population of cattle which
between direct and materna1 and between materna1 and
exhibits large differences in materna1 genetic effects. Simi-
grandmaternal genctic effccts may bc explained as the
lar results were found with native Korean cattle (Lee et
negative influence of dams on thc materna1 ability of theil
al., 1998) which also exhibit large differences in 4 month
female offspring through ovcrfeeding ( Koch. 1972 ). Tawah
and 6 month weights as a result of materna] genetic ef-
et al. ( 1993) suggested that these negati\\,e correlations may
fects. Dodenhoff et a/. (1998) reported significant varia-
retlect adaptation 01‘ the animais to a dry tropical environ-
tion in weaning weight attributable to grandmaternal ef-
ment where l’ood rcsourccs are scarcc. In such an environ-
fects in a population of Hereford cattle with moderate ex-
ment, small cows tend to meet their nutritional reyuire-
pression of materna] effects and a large negative correla-
ments for maintenance and growth of their calves more
tion between direct and materna1 genetic effects. These
easily than larger cows. The latter would produce smaller
authors noted that when estimates of grandmaternal heri-
calves, especially at weanin,.
(y than would smaller cows of
tability were not zero. the estimates of materna] herita-
similar age.
bility increased, whereas direct heritability was not af-
Materna1 heritability was important in all cases.
fected. A similar pattern was also observed in another
except birth weight, with estimates of the same magnitude
population of Hereford cattle that had undergone long-
as for direct heritability. The estimates for weaning weight
term selection in Montana (USA) (Ferreira, G.B.,
agreed with values reported for Hereford cattle (Meyer.
1 YY2 ), Nclorr: cattle tElu ut CI/.. 1995). Makwa ;mtl GuJali
0.26 L 0,07 e 0.1 h k 0.07 piIra ]XSO COI11 1 iUl0 tZ O,I4 + 0,06 e
cattlc (Tawah or (11.. 1993). and Mashona cattlc (Khombc
0. Ih -t 0.06 para o peso final. As estimativas da correlaç5o
pt ll/.. 19%). For yearling and final weights. the estimatcs
gcn6tica cntrc os clc‘itos direto e materna1 para os pesos a0
nn5cimcnto. ao desmame. com 1 an0 e final Ioram -0.17 f 0.40.
of direct. but not matrrnal. heritability were consistent with
-0.5X f 0.32. -0,52 + 0.34 e -0.34 rt 0.37. respectivamente. Para
most published estimates. Materna1 effccts would he ex-
o IW\\O com I ano com herdabilidade relativa à av6 estimada
pected t0 diminish ilt thcse agcs bccWsc the animais Il0
coma scndo apcnus 0.03. o modela 3 deu estimativas da
longer clcpend on thcir mothers. Relatively high estimalcs
corrclnc;50 gen6tica entre os cleitos direto e relative à av6 e entre
of hi, were also reported hy Eler cif rrl. ( I 995) and
05 cltiitos matemal c relative :I avd de 0,28 + 0.48 e -0,33 IL 0X17,
Mockinnon rl nl. ( 199 1 ) af yearling and later ages. reslxc-
qxxctivamente. As estimatives de herdabilidade direta e mater-
tivcly. Thcsc authorx both also suggested that for animais
nnl Cio se altcrkma~~~ quündo os efeitos relativos à av6 ii5o foram
ritised 011 pasture with Little or no feed suppletnent. the letgth
incluidos no modela.
of time between wcaning and yearling ages may not be
etlot~gh to buffer tn:ttertliJ effccts present a~ we:ming.
REFERENCES
The cstimate of thc ratio of matemal permanent
environnlental variance to the phcnotypic v:tri>mcc. c‘. wus
Bakw. R.I.. ( IWO). The rolc of tuu~crnal cllècti, on thc efficicncy OF seleçtion
larger for weaning weight than for the other traits. as also
in hcef’catllc>. A rcvicw. I’lr~c,.
N. Z. Snc.. A~IN~. Prod. 40: 2X5-303.
Diop. %l. antl
reported hy Eler ef tri. ( 199.5) and Meycr ( I993). Perrna-
Vdll \\‘leîk, Id. ( 1
WX). Ehtirnates of genetic parama% for
gn~lh traits of Gobra cattle. A~ini. Sci. 66: 349-35.5.
nent cnvironmental effects rcsult from incidents that af-
Dodenhoff. J., Van Vleck, L.D., Kachman, S.D. and Koch, R.M. (199X).
fect nll progeny OI’ the same CIW. The cllccts before wcan-
f’ammclcr
estiiiiatcz
fi,r direct. materna1
and jrantlmaiernal
genetic
ing niay be caused by sequels of diseases or accidents to
cl’fcct\\ for hirlh wcight and seaninp
Wright in Hereford cattle. J.
,Irrirx S<~i. 7fl: 321-2527.
the ul&r. which Will affect thc rnilk production ofthc dam.
Klcr. J.l?, Van Vlrçk, L.D., Ferraz, J.R.S. and Lôbo. R.R. (19%). Esti-
whcrcas the estimate of c’ at later ages may rcllect il carry-
mation OF \\al-iancch
dur to direct and materna1
cffects for growth
over et’fect on weaning
weight.
Iralts 01 Nclorc catlle. .1. ,Irrirrr. SU. , . __
‘?. 375L32SX.
Falconrr. D.S. ( 1965 I. Matcrnal
effects antl sclcction
response. In: Gcrlef-
il2 ‘Todri~. f’~r~rwc/irrg.s o/ thc XI Inietmrfior~trl
Coqwss of’ Gmrf-
CONCLUSIONS
ic 1. ‘l’hc 1 Iague. The Nelhcrlands.
Vol 3. p. 763.
Johnson, D.I.. antl Thompson.
R. ( 1995). Restricted
maximum likelihood
Grandniaternal
ellècts do net play an important rolc
estimation
of \\ariance
components
for univariate
animal modçls us-
in thc weight of Gobra cattle at birth, weaning (6 months).
ing \\parsc matrix tcchniclues
und avt:rage information. .I. ZIclir;v :,ci.
7s: 149~4.56.
yGl~l~ll~ ( 13 months) and 18 months of age. In contrast.
Khoml)e, C.T., Hales, J.F., Cue, R.I. ml WI&, K.M. ( 1995~. Estimation
niuternal etlècts are important, even at 18 nionths of ilgt!.
r>f direct atlditivc and maternal
additive genetic effects for weaning
~Modcls l’or the genetic evaluation of Gobra cattlc could
vzight in MAona cattle of Zimbahwr usine an individual animal
sakly ignore grandmaternal
efkcts but should include di-
modcl. i\\rlirU. I’lTl<I. 60: -l I--IX.
Koch, R.M. t 1972). The rolc of matrrnal
effecth in animal hreeding.
VI.
rect ml materna1 penetic effects with the corresponding
M;~lcrnal cffcct\\ in hccf car~le. J. A/rGn. .W. .<S: 1.3 lb- 1.323.
dlrcct- ninternal genetic covariance.
Korh. R.M. UKI Clark, R.T. ( IYSS). Genctic and cnvironmental rclation-
\\hip\\ ;IIIIOII~ cconomic
charactcr\\
in hcef caltle. III. Evaluating
TXI-
RESUMO
iernal cnvironmcnt.
./. Anirn. Sci. 14: 979.996.
IN, J.W., Chai, S.B. and Van Vleck, L.D. ( lY98). Lack of grandmaternal
genetic effect\\ for birth wcight, weaning
weight, and six month weight
Forain obtidas as estimativas dos pesos uo nascimento
in Korcan
native cattle. P,r~ccwlin,q
of’ th 8th Wdtl Congres.~ on
(N = 3909). ao desmame (N = 342.5). com 1 ano de idade (N =
A~ir& Pro~lrrctior~,
Seoul National University, Seoul, Korea. June
2764) e final (N = 2114) a partir dos registres de gado C;obra
2X-July 4. 199X. Vol. 2. pp. 630-631,
coletados no Centro de Pesyuisns em Zootecnia de Dahra,
Mackinnnn, M.J.. Meyer, K. and Hetzel, J.D.S. (199 1). Genetic variation
Senegal. Três modelos animais fora-n adaptados para ohtel
and covariation
for growth. parasite resistance
and hent tolerance
in
tropical cattle. /.ii~I. Prml. sci. 27: 1 os- 122.
estimativns por REML usando uma abordagem dc informa@
Mryw, K. ( lYY3). Variance components
due to direct and materna1 effects
média (AI). 0 modela 1 con4itlerou os cfeitos amhientais
for srowth traits of Australian
heef cattle. Lil~,sr. Pnjr/. Scj. 3I: 179-
alcatdrios dircto, gcntitico maternai e matemal permanente. No
2 0 4 .
modclo 2. uni efeito geixl relative %s üvds foi adiçionado ;IOS
Robison, 0.W. ( 19X I ). Thc influence of materna1 effects on the efficiency
efeitoh aleatcirios considerados no modela 1. e no modelo 3 0
of selection:
a rcview. Lii~r. Pr&. S?i. 8: 121.137.
cl’eito gcrul relative às av6s foi dividido em eteitos amhientais
SAS Institute Il 9Y2). .SAS/sïA7: K&tr.sr 6.05’. SAS Institute Inc., Cary, NC.
genético c permanente. Todos os modelos admitiram covariâncias
T’zwah, C.I.., Mbah, D.A., Rege, J.E.O. and Oumate, H. (1993). Genetic
evalua~ion of hirth and weaning
wcights of Gudali and two-breed
entre 0% eI‘eitos genéticos. A inclus%) dos efeitos relatives 5s
\\ynthclic
Wakwa hecfcattle
populations under selection
in Cameroon:
av6s nos modelos 2 t’ 3 n5o alterou as estimativas dos parâmetroh
pcnetic ;wd phenotypic
parametera.
Atrim. Prori. 57: 73-79.
gcnéticos comparados com o modela 1. As vnrifincins atribuiveis
‘fotusek, R., Stephens, O.F.. Krnpp, J.R., Holloway, S.W., Knosi, J.W.
aos elèitos relatives às av6s tornaram-se negativai e i!xam
and Whiteman, J.V. (1 Y7 1). Milk production of range CO~S. 0X-la.
posicionadas pr6ximas a zero, exceto para 0 peso coiil 1 ano,
Aqrk. E\\t. Srtr. Mi.SC~. Pllh. x5.
para 0 qtial ii herdabilidade relativn à avo foi 0,O.J 5~ 0.01. As
Willham, R.L. ( 1963 1. The covariance between relatives for characters com-
estimativas para as herdahilidades
direta e materna1 foram.
po\\ed of componcnts
contrihuted
hy related individuals.
Riomrtri<~.s
/Y: 1X-27.
re\\peclivamente.
0.08 f 0,03 c 0,03 tr 0.02 para peso ao
nnscimento, 0.20 * 0.05 e 0.2 1 f 0.05 para peso ii0 desmame.
ESTIMATES OF DIRECT, MATERNAL AND GRANDMATERNAL GENETIC EFFECTS
FOR GROWTH TRAITS IN GOBRA CATTLE
ABSTRACT
Estimates of genetic parameters for birth (N = 3909), weaning (N = 3425), yearling (N = 2764) and final (N = 2144) weights
were obtained from the records of Gobra cattle collected at the Centre de Recherches Zootechniques de Dahra, Senegal.
Three animal models were fitted to obtain estimates by REML using an average information (Al) approach. Mode1 1 consid-
ered random direct, materna1 genetic and maternai permanent environmental effects. In mode1 2, a general grandmaternal
effect was added to the random effects considered in mode1 1, and in mode1 3, the general grandmaternal effect was divided
into grandmaternal genetic and grandmaternal permanent environmental effects. All models allowed covariances among
genetic effects. The inclusion of grandmaternal effects in models 2 and 3 did not change the estimates of the genetic param-
eters compared to mode1 1. Variantes attributable to grandmaternal effects became negative and were set close to zero, except
for yearling weight for which grandmaternal heritability was 0.03 + 0.03. The estimates for direct and materna1 heritabilities were,
respectively, 0.08 + 0.03 and 0.03 k 0.02 for birth, 0.20 k 0.05 and 0.21 + 0.05 for weaning, 0.26 + 0.07 and 0.16 + 0.07 for yearling
and 0.14 rfr 0.06 and 0.16 i 0.06 for final weights. The estimates of the genetic correlation between direct and materna1 effects for
birth, weaning, yearling and final weights were -0.17 f 0.40, -0.58 f 0.32, -0.52 k 0.34 and -0.34 f 0.37, respectively. For yearling
weight with grandmaternal heritability estimated to be only 0.03, mode1 3 gave estimates of the genetic correlation between direct
and grandmaternal effects and between materna1 and grandmaternal effects of 0.28 f 0.48 and -0.33 f 0.67, respectively.
Estimates of direct and materna1 heritabilities were unchanged when grandmaternal effects were not included in the model.
INTRODUCTION
out that such models may be suboptimal and need to be
improved. The objective of this study was to evaluate the
Growth traits of beef cattle are known to be influ-
influence of including grandmaternal effects in a mode1
enced by materna1 effects. Early reports (Koch and Clark,
for estimating genetic parameters for growth traits in Gobra
1955: Willham, 1963) mentioned the possible influence
cattle for which materna1 effects have previously been
of grandmaternal effects. Falconer ( 1965) considered the
shown to be important.
materna1 effect as a linear function of a mother’s pheno-
typic value intluenced by a11 materna1 ancestors. The so-
MATERIAL AND METHODS
called ‘fatty udder syndrome’ is an example of a grand-
materna1 effect: a granddam with a high materna] ability
The records of weight at birth, 6, 12, and 18 months
may over-feed her daughter and thereby have a negative
were obtained from the Gobra herd at the Centre de
intluence on her daughter’s materna1 ability by inhibiting
Recherches Zootechniques de Dahra, Senegal. The pro-
development of her udder tissue (Totusek ef ~1.. 1971).
duction environment is characterized by a dry tropical cli-
The importance of materna1 effects in beef cattle has been
mate with two distinct seasons: a long dry season from
widcly reported (Koch. 1972; Baker, 1980: Robison, 198 1;
October to June and a short rainy season from July to Sep-
Meyer, 1992). Studies on Af’rican cattle populations have
tember. The mean annual rainfall from 1964 to 1988 was
reported a strong materna1 influence on growth traits
360 mm. The mean annual temperature is 28°C. The veg-
(Tawah cf ~11.. 199.3: Khombe Côt ~11.. 1995: Diop and Van
etation is described as savanna type dominated by Acaciu
Vleck. 1998).
sp. and annual grasses. Annual biomass production is
Animal models used to estimate materna1 effects
closely related to the amount of rainfall the area receives
typically include direct and materna1 genetic effects and
each year.
rnaternal permanent environmental effects and consider the
Native pasture is the main source of feed. The quan-
covariance between genetic effects. Meyer ( 1992) pointed
tity and quality of the pasture vary considerably during the
year. During the rainy season, pastures are of good quality
and abundant. During the dry season. the nutritive value
of the forage decreases and supplemental feeding with
ground nutcake or cottonseed is provided. especially for
suckling cows and weaned calves.
Breeding females were randomly assigned to sires
(30-50 cows/sire) for a breeding season from December
to Mat-ch. Cows that were not pregnant 3 months after the
breeding season started were reassigned to a different bull.
.
364
Diop <‘/ tri.
Over the years, thc si/e of’ III~ hcrd Iras Iluctua~cd around
300 CO~S. Males and fèrnales wcrc sclec~ed bas~cl ou
weight at 6 months (wraning). Malca werc again sclccted
based on their wcight a( 1 X months aftcr which the IO bcsr
males underwcnl ii growth pcrt’ormancc
(est hct’orc~ final
selection for the two or thrce bulls to bc uscd as rcplacc-
ment bulls. Replacemen~ I’cIllalcs werc sclcctcd a( 24
months based on Ihcir wcighl. About 5’4 of thc malch and
80% of the femalcs selectcd aficr wcaning \\vcrc LIWI as
replacements.
Culling of thc c’ows was bnscd on polar rc-
productive performance (ion,0 calving iiiler\\,al OI’ failurc
to calve after two breeding aensons~ or pour pwth pcr-
formance of offspring.
The data for the present anaIysc\\ consistcd of the
records of animais h-n I’ronI 1963 IO 19X9. C’cmG~ttwcy
rcspccti\\.cIy. a1111 N ix thc numbcr of records. A is Ihc nu-
checks werc performcd upon tho idcntifica~ion I)I‘ ihc ani-
mcrator rclationshi p rnali is anlong aninlals in the pedi-
mals and their pedigrees. Records of thc progcny OI sires
frcc Iï Ic. ancl thc 1 malriccs arc identily matrices of speci-
with less than five pro~eny wcrc delctcd from I~IC ;maly-
Iïcd ordcr. Modclx I und 3 rcquirc cach iminltll with il record
sis. The PROC MIXEIl) option of SAS (SAS. 1093) was
lo havc a trandtl;~ln. Thcrclore. “du1~111y” granddam identi-
used to lest the significancc of thc fixed cfli’c~~ 01’ ~ncmth
Iïc~alic>ll \\\\‘a$ ;(S\\i cncd I~)I- records \\\\rith unkno\\\\sn granddam
of birth, year of birth. scx and parity in a modcl with Grc
in@;inp thc assomption lh;lI ~hcsc ~randdams are unrelated
considered to be a random cffcct.
;lnll~ll~ ~hcIll~t!~~~~IS il1111 1lnlL’bX~ t0 thosc \\Vith know’n idell-
For estimation of thc gcnctic parmnctc~~s. thrcc ani-
lilïcalion Tilljlt’ 1 sui~ini;iri/cs thC d3til LlSCd in thC analyses.
mal models wcrc fitted for the four traits consitlcrc4 :
i’omponcnts of \\ ;Iri;\\llcc were cstimated b> re-
strictcd ~naximunl likelihood tREM1.) usin? a n average
inli)rnialicm alyjrirhnl (.lohns~~n XI~ Thompson. 19%). The
Mode] 2: y = Xp + Z, a + ZI 111 + Cc’, ç + \\?‘? y +- c
alyorithin is a Newton ~nc~hoci w h i c h LI~~S first and WC-
Mode] 3: y = Xp + Z, a + Zz m + x3 g + \\Y, c + \\Y? q + e
ond rlci i\\.alivc\\ of thc iogilrilhm 01’ the likclihood given
thc tlat;l t0 iïnd cstirllntc~~ or \\ iliiilWC Cl~~llpO~lC~ltS
ht
where y is an N x I vector of ohsl‘rvafions,
p is lhc \\ cc101
lllilXiIlli/r~ lhc likitli hood l’unclion.
of fixed effects of ycar of birth. month of birth. WL and
A pclinlillilry
ilKll?\\iS OI‘ \\~ilriallcc ShoWed that the
parity of dam. a, m and g iII’C \\:eclors of hrcc‘tliirg \\ alues
fixcd cl‘li.*~*l\\ of iiioiilh and !c;lr of birth. scx. illld parity
for direct. malcrnal and LLrilnC11ii;ltcl.Il;II gcncilic ~~~I~~CI~, ç
wcrc 4gnil‘ic.ant. (‘c)tlxc’titlt’tlllv. ~hc\\e eff’cc~s wcre included
and q are vectors of matcrnal pcrmancnt
cn\\,ironiircnlal
in ail thrcc nll.ltlcls.
and grandmaternal
permanenl
en\\‘ironrncnt~~~ cl’fticlb and
f:xlim;itc\\ of rcIali\\‘c’ \\ arianccs attributaI.+ to ad-
e is an N x 1 vector of rcsidual random cffccis. ;IIKI X, %,,
dilivc ci~rect #~II:). nlaternal th,:.) and _«randmaterna1 (11;)
Z,, Z,, W, and W, are known incidence matriccx rclating
(~cnctic Cl’lCcts \\‘1crc cillc~ll;ltctl a4 riltiC)s of the esiimates
observations to their respective fixed and randonl cll&..~s.
Of adltil i\\re diret.t CC?). nlatcrnill gcnctic (of,,) and grand-
Matrices Z,, Z2 and Z, were augmenled for animais with-
matcrnal genctic. c$) varianccs. rcspectively. to the phe-
out records that were included in Ihe relalionship matrix.
notypic varimlcc CC$>. Thc cliI.cct-mnternal (ri,,,)). direct-
For the most completc mode1 (model 3). E (y) =
grandma~erna1 (r,?) and matcln;1l-=randmaternal
(r,,,<) ge-
Xp and
nctic correlalionx were estimated as ratios of the direct-
nlaternirl (<r,,,,,), I,lirccr-eranclnlrlternal (o,~J and maternal-
a
grandrnaternal (cF,,,,) genetic covariances t6 the square roots
of ~hc products of CJ~~ and of,, 0: and 05, and oi,. and 0:.
m I
respcctivcly. ‘Thc estimalcs of‘ relative variantes attribut-
ahle to nlatcrnal pcrnlancnt environmental effects (c’) and
g
<rr;iiidiil;ilcrii~it
pcrnianenr en~~ii’oiimenta1 effects (q’) were
V
=
ZillCllliltCd ;lb th! ratios of thC CstillliltêS of maternal (0:)
C
illld ~r~indiiialcrii31
((ri) permanent cnvironmental variancés
to thc phcnotypic variancc ($1.
RESULTS
e
The \\2riilncc components iltld estimates of genetic
where Nd and Ng are numbers 01‘ dams :III~ grand(ia~ns.
parainclcrs
(IdiOS aiitl correlations)
for birth weight are
prehcnted in Table II. The direct (hi) and maternai (hf,, )
Thc cstimates for yearling ueight (Table III)
heritabilities were 0.08 -t 0.03 and 0.03 k 0.02. respec-
sl~owctl that the inclusion of grandmatcrnal genetic effects
tively. for the three modela. The estimates were similar for
in model 3 slightly incrcased thc additive direct genetic
a11 effects included in a11 models. The cstimates of vuri-
v:ukmce wirh an estimate of grandmuternal genetic heri-
ances attributable to grandmatcrnal Fenctic and/or grimd-
tahility (hi‘) of 0.03 + 0.03. The estimated values of the
materna1 permanent environmentu1 effkcts werc close to
<Fcnctic correlations bctwern the direci 2nd grandmater-
2
zero in models 2 and 3. The estimate of the corrclation
nal genetic effects (r,_) and between the materna1 and
between direct and materna1 genetic effects L;IS -0. I7 +
grandmaterna cl‘fects i r,,,?) were small to moderate with
0.41 . The estimate of& was 0.0-i -t 0.02 for ail modelh. As
large srandard errors (O.% rt 0.38 and -0.33 tr 0.67, re-
expected from the estimates ashociated with grandtmtter-
spcctivcly). Thc eatimates of variante components attrib-
na1 effects. the three models resulted in c\\scntiidly the siune
utahlc to direct and marna1 effects and their correlation
value for the log likelihood.
uc‘re similar for the three rnodels. The estimates of hi
The pattern of the estimates for weaning wright
and hf,, were 0.36 k 0.07 2nd 0.16 + 0.07, respectively.
(Table II) was similar to tllat pattcnl of’ estimates l(n
Thc cstimate of r.,,,, was -05.5 IL 0.37 for mode1 3. The
birth weipht. i.e., the estimates of thc variance and co-
cstimates of c’ ~+ercf 0.09 k 0.03 for mode1 1 and 0.07 t-
variante components for includcd effects were 01‘ the
0.0-l for modcl 3.
same magnitude in the different ~~~odels. Variancc\\ at-
Thr estimatc of bariance for final weight attribut-
tributable to grandmaternal cffects [vert: small. with
ahlc to ~landtnatcrnnl el’fccts wx nrar zero (Tilble III).
values near zero. The estimatcs of direct nnd maternai
The estimates of direct (hi) and maternai (ht),) heritabilities
heritabilities were 0.20 $r 0.05 and 0.2 1 rt 0.05. rcspec-
wcrc O.I3 31 0.06 and 0. I6 + 0.06, respectively. The esti-
tively. The direct-materna1 genetic corrclation ~~1s -0.58
mate of c’ W;IS 0.04 $r 0.05. Thct correlation between the
IL 0.33 and c? wus 0.15 4 0.0-I.
direct and maternai genztic effects was -0.34 * 0.37.
4.
.
366
Table 111 - Estimates of çompon~~ts of variancc and gcnetic paramctcr\\
for )ç;~rlin~
and final weights for Gohru cxttle.--_
-
-
Yearling weight
Final wcirhl
-
-
-
-
ParameteW
Mode1 1
Model 2
Mode1 3
Modcl I
Moclcl 2
hlodel
3
s
169.7
169.4
171.3
I20.3
129.1
120.1
cl:,
10X.7
106.0
107.X
144.9
110.5 IN.5
7:
2 0 . 8
0.0
CT .ll/>
-70.4
-70.0
-76.6
-46.7
45.3
-15.3
cT.,g
16.X
0.0
on1p
-15.4
0.0
ls,f
5 7 . 1
53.2
4x.5
42.9
3.5.
I
35.2
s
5 . 2
0 . 0
10.3 10.7
0;
39X.4
398.5
394. I
6 3 Z . 0
632.3
633.1
0,:
6635
663.2
663.2
w3.3
~~01.0 902.4
11;
0.26 k 0.07
0.26 k 0.07
0.26 f 0.07
0.13 !: 0.06
0.14 7t O.Oh 0.11 i 0 . 0 6
hi,
0.16 k 0.05
0.16 + 0.0
0.16 f 0.07
0.16 -r 0.06
0.16 t 0.06 0.16 t O.Oh
h;
0.03 Ek 0.03
0.00 k 0.00
r.,,.
-0.52 k 0.34
-0.52 + 0.35
-o.ss * 0.37
-0.3-I i 0.27
- 0 . 7 4 -f 0 . i7 -0.34 k 0..37
r.lis
0.28 k 0.48
0.(10 i 0.00
hg
-0.33 k 0.67
0.00 r 0.00
$
0.09 f 0.03
0.08 f 0.04
0.07 Ii 0.04
0.05 + 0.04
0.03 i 0.0s 0.04 d. 0.05
Y?
0.01 -I 0.02
0.00 i 0.00
O.OI zk 0.03 O.OI f 0.03
-2 log L”
2033s.2
20.135. I
20334.2
IOJ4.5.6
I h.J-li.4
I h-115.-I
-_--.-
%si. Direct genctic variante:
o:,, materna1 genetic variante:
0:. gnuntlmnlcrnnl
selletic \\ ;ui;mcc: (T,,?. rlirec~l and malcr-
nal genetic covariance: CT,,,,, direct and grdmatunül genctlrc covariancc:
<T,,,L. m;~tcrnal and ui-;tndiiiatcr1lal
gcnctk
r
covariance:
of, maternul permanent environmentid
variancc: a. ~r;intlln;ltcrn;II
pcrm;mcnt
cm ilrulmcnt;ll
\\ ariaicx:
0:. temporary
environmental variante:
$, phcnotypic
variancc:
ha. clirccl hcritahility:
II,:.. nlalt’mal
hcrilahilit!:
h:
grandmaternal heritability; r,,,,,, direct-malcrnal
gcnetic corrclntion:
r,,_. tlirCCt-~r;llltliIl;ltCrIl:Il
pcnclic ckWrcl;ition:
r
/on
malemal-frandinaternal
genctic corr&tion:
c’= a’io;: cl?= c$k$,. “-2 lop 1. = Minus 1% icc tlic Io? ~1 lihclih(~~d $\\cn thc tlat;l.
DISCUSSION
MacNeil. M.D.. and Van VIe&. I,.D., unpublished results).
For yearling wcight. thcre was a small. positive
The inclusion ofgrandmaternal effects in the mode1
correlation bctwccn direct and ~randmaternal genetic ef-
did not change the estimates of the direct and materna1
fects (r,,r) and a modcrate. negative correlation betwYen
heritabilities or of the genetic correlation between the di-
the mat&nal ancl grandmaternal gcnetic effects (r,,iP). The
rect and materna1 genetic effects for the four traits ana-
standard en-ors associated Lvith thcse estimates uerk large
lyzed. However, for yearling weight. the inclusion of grand-
which means they arc not significantly different from zero.
materna1 genetic effects in mode1 3 slightly increased the
Dodenhoff et ol. (1998) rcported small to large positive
additive direct genetic variante. The estimates of variante
estimates of rZ,E at weaning for differcnt lines of Hereford
attributable to grandmaternal effects were very small or
cattle. The estimates for r,,,, were negative and large for
zero, suggesting that these effects did not play an impor-
weaning wcight in these Ii&. The negative correlation
tant role in these traits in this population of cattle which
between direct and materna1 and between materna1 and
exhibits large differences in materna1 genetic effects. Simi-
grandmaternal genctic effccts may bc explained as the
lar results were found with native Korean cattle (Lee et
negative influence of dams on thc materna1 ability of theil
al., 1998) which also exhibit large differences in 4 month
female offspring through ovcrfeeding ( Koch. 1972 ). Tawah
and 6 month weights as a result of materna] genetic ef-
et al. ( 1993) suggested that these negati\\,e correlations may
fects. Dodenhoff et a/. (1998) reported significant varia-
retlect adaptation 01‘ the animais to a dry tropical environ-
tion in weaning weight attributable to grandmaternal ef-
ment where l’ood rcsourccs are scarcc. In such an environ-
fects in a population of Hereford cattle with moderate ex-
ment, small cows tend to meet their nutritional reyuire-
pression of materna] effects and a large negative correla-
ments for maintenance and growth of their calves more
tion between direct and materna1 genetic effects. These
easily than larger cows. The latter would produce smaller
authors noted that when estimates of grandmaternal heri-
calves, especially at weanin,.
(y than would smaller cows of
tability were not zero. the estimates of materna] herita-
similar age.
bility increased, whereas direct heritability was not af-
Materna1 heritability was important in all cases.
fected. A similar pattern was also observed in another
except birth weight, with estimates of the same magnitude
population of Hereford cattle that had undergone long-
as for direct heritability. The estimates for weaning weight
term selection in Montana (USA) (Ferreira, G.B.,
agreed with values reported for Hereford cattle (Meyer.
1 YY2 ), Nclorr: cattle tElu ut CI/.. 1995). Makwa ;mtl GuJali
0.26 L 0,07 e 0.1 h k 0.07 piIra ]XSO COI11 1 iUl0 tZ O,I4 + 0,06 e
cattlc (Tawah or (11.. 1993). and Mashona cattlc (Khombc
0. Ih -t 0.06 para o peso final. As estimativas da correlaç5o
pt ll/.. 19%). For yearling and final weights. the estimatcs
gcn6tica cntrc os clc‘itos direto e materna1 para os pesos a0
nn5cimcnto. ao desmame. com 1 an0 e final Ioram -0.17 f 0.40.
of direct. but not matrrnal. heritability were consistent with
-0.5X f 0.32. -0,52 + 0.34 e -0.34 rt 0.37. respectivamente. Para
most published estimates. Materna1 effccts would he ex-
o IW\\O com I ano com herdabilidade relativa à av6 estimada
pected t0 diminish ilt thcse agcs bccWsc the animais Il0
coma scndo apcnus 0.03. o modela 3 deu estimativas da
longer clcpend on thcir mothers. Relatively high estimalcs
corrclnc;50 gen6tica entre os cleitos direto e relative à av6 e entre
of hi, were also reported hy Eler cif rrl. ( I 995) and
05 cltiitos matemal c relative :I avd de 0,28 + 0.48 e -0,33 IL 0X17,
Mockinnon rl nl. ( 199 1 ) af yearling and later ages. reslxc-
qxxctivamente. As estimatives de herdabilidade direta e mater-
tivcly. Thcsc authorx both also suggested that for animais
nnl Cio se altcrkma~~~ quündo os efeitos relativos à av6 ii5o foram
ritised 011 pasture with Little or no feed suppletnent. the letgth
incluidos no modela.
of time between wcaning and yearling ages may not be
etlot~gh to buffer tn:ttertliJ effccts present a~ we:ming.
REFERENCES
The cstimate of thc ratio of matemal permanent
environnlental variance to the phcnotypic v:tri>mcc. c‘. wus
Bakw. R.I.. ( IWO). The rolc of tuu~crnal cllècti, on thc efficicncy OF seleçtion
larger for weaning weight than for the other traits. as also
in hcef’catllc>. A rcvicw. I’lr~c,.
N. Z. Snc.. A~IN~. Prod. 40: 2X5-303.
Diop. %l. antl
reported hy Eler ef tri. ( 199.5) and Meycr ( I993). Perrna-
Vdll \\‘leîk, Id. ( 1
WX). Ehtirnates of genetic parama% for
gn~lh traits of Gobra cattle. A~ini. Sci. 66: 349-35.5.
nent cnvironmental effects rcsult from incidents that af-
Dodenhoff. J., Van Vleck, L.D., Kachman, S.D. and Koch, R.M. (199X).
fect nll progeny OI’ the same CIW. The cllccts before wcan-
f’ammclcr
estiiiiatcz
fi,r direct. materna1
and jrantlmaiernal
genetic
ing niay be caused by sequels of diseases or accidents to
cl’fcct\\ for hirlh wcight and seaninp
Wright in Hereford cattle. J.
,Irrirx S<~i. 7fl: 321-2527.
the ul&r. which Will affect thc rnilk production ofthc dam.
Klcr. J.l?, Van Vlrçk, L.D., Ferraz, J.R.S. and Lôbo. R.R. (19%). Esti-
whcrcas the estimate of c’ at later ages may rcllect il carry-
mation OF \\al-iancch
dur to direct and materna1
cffects for growth
over et’fect on weaning
weight.
Iralts 01 Nclorc catlle. .1. ,Irrirrr. SU. , . __
‘?. 375L32SX.
Falconrr. D.S. ( 1965 I. Matcrnal
effects antl sclcction
response. In: Gcrlef-
il2 ‘Todri~. f’~r~rwc/irrg.s o/ thc XI Inietmrfior~trl
Coqwss of’ Gmrf-
CONCLUSIONS
ic 1. ‘l’hc 1 Iague. The Nelhcrlands.
Vol 3. p. 763.
Johnson, D.I.. antl Thompson.
R. ( 1995). Restricted
maximum likelihood
Grandniaternal
ellècts do net play an important rolc
estimation
of \\ariance
components
for univariate
animal modçls us-
in thc weight of Gobra cattle at birth, weaning (6 months).
ing \\parsc matrix tcchniclues
und avt:rage information. .I. ZIclir;v :,ci.
7s: 149~4.56.
yGl~l~ll~ ( 13 months) and 18 months of age. In contrast.
Khoml)e, C.T., Hales, J.F., Cue, R.I. ml WI&, K.M. ( 1995~. Estimation
niuternal etlècts are important, even at 18 nionths of ilgt!.
r>f direct atlditivc and maternal
additive genetic effects for weaning
~Modcls l’or the genetic evaluation of Gobra cattlc could
vzight in MAona cattle of Zimbahwr usine an individual animal
sakly ignore grandmaternal
efkcts but should include di-
modcl. i\\rlirU. I’lTl<I. 60: -l I--IX.
Koch, R.M. t 1972). The rolc of matrrnal
effecth in animal hreeding.
VI.
rect ml materna1 penetic effects with the corresponding
M;~lcrnal cffcct\\ in hccf car~le. J. A/rGn. .W. .<S: 1.3 lb- 1.323.
dlrcct- ninternal genetic covariance.
Korh. R.M. UKI Clark, R.T. ( IYSS). Genctic and cnvironmental rclation-
\\hip\\ ;IIIIOII~ cconomic
charactcr\\
in hcef caltle. III. Evaluating
TXI-
RESUMO
iernal cnvironmcnt.
./. Anirn. Sci. 14: 979.996.
IN, J.W., Chai, S.B. and Van Vleck, L.D. ( lY98). Lack of grandmaternal
genetic effect\\ for birth wcight, weaning
weight, and six month weight
Forain obtidas as estimativas dos pesos uo nascimento
in Korcan
native cattle. P,r~ccwlin,q
of’ th 8th Wdtl Congres.~ on
(N = 3909). ao desmame (N = 342.5). com 1 ano de idade (N =
A~ir& Pro~lrrctior~,
Seoul National University, Seoul, Korea. June
2764) e final (N = 2114) a partir dos registres de gado C;obra
2X-July 4. 199X. Vol. 2. pp. 630-631,
coletados no Centro de Pesyuisns em Zootecnia de Dahra,
Mackinnnn, M.J.. Meyer, K. and Hetzel, J.D.S. (199 1). Genetic variation
Senegal. Três modelos animais fora-n adaptados para ohtel
and covariation
for growth. parasite resistance
and hent tolerance
in
tropical cattle. /.ii~I. Prml. sci. 27: 1 os- 122.
estimativns por REML usando uma abordagem dc informa@
Mryw, K. ( lYY3). Variance components
due to direct and materna1 effects
média (AI). 0 modela 1 con4itlerou os cfeitos amhientais
for srowth traits of Australian
heef cattle. Lil~,sr. Pnjr/. Scj. 3I: 179-
alcatdrios dircto, gcntitico maternai e matemal permanente. No
2 0 4 .
modclo 2. uni efeito geixl relative %s üvds foi adiçionado ;IOS
Robison, 0.W. ( 19X I ). Thc influence of materna1 effects on the efficiency
efeitoh aleatcirios considerados no modela 1. e no modelo 3 0
of selection:
a rcview. Lii~r. Pr&. S?i. 8: 121.137.
cl’eito gcrul relative às av6s foi dividido em eteitos amhientais
SAS Institute Il 9Y2). .SAS/sïA7: K&tr.sr 6.05’. SAS Institute Inc., Cary, NC.
genético c permanente. Todos os modelos admitiram covariâncias
T’zwah, C.I.., Mbah, D.A., Rege, J.E.O. and Oumate, H. (1993). Genetic
evalua~ion of hirth and weaning
wcights of Gudali and two-breed
entre 0% eI‘eitos genéticos. A inclus%) dos efeitos relatives 5s
\\ynthclic
Wakwa hecfcattle
populations under selection
in Cameroon:
av6s nos modelos 2 t’ 3 n5o alterou as estimativas dos parâmetroh
pcnetic ;wd phenotypic
parametera.
Atrim. Prori. 57: 73-79.
gcnéticos comparados com o modela 1. As vnrifincins atribuiveis
‘fotusek, R., Stephens, O.F.. Krnpp, J.R., Holloway, S.W., Knosi, J.W.
aos elèitos relatives às av6s tornaram-se negativai e i!xam
and Whiteman, J.V. (1 Y7 1). Milk production of range CO~S. 0X-la.
posicionadas pr6ximas a zero, exceto para 0 peso coiil 1 ano,
Aqrk. E\\t. Srtr. Mi.SC~. Pllh. x5.
para 0 qtial ii herdabilidade relativn à avo foi 0,O.J 5~ 0.01. As
Willham, R.L. ( 1963 1. The covariance between relatives for characters com-
estimativas para as herdahilidades
direta e materna1 foram.
po\\ed of componcnts
contrihuted
hy related individuals.
Riomrtri<~.s
/Y: 1X-27.
re\\peclivamente.
0.08 f 0,03 c 0,03 tr 0.02 para peso ao
nnscimento, 0.20 * 0.05 e 0.2 1 f 0.05 para peso ii0 desmame.