RESEARCH ARTICLE Open Access
Gene ic suscep ibili y o in ec ious disease in Eas
A ican Sho ho n Zebu: a genome-wide analysis
o he e ec o he e ozygosi y and exo ic
in og ession
Gemma GR Mu ay
1,2*
, Ma k EJ Woolhouse
1
, Miika Tapio
3
, Ma y N Mbole-Ka iuki
4
, Tad S Sons ega d
5
,
Samuel M Thumbi
1,9
, Amy E Jennings
6
, Ilana Con adie an Wyk
7
, Ma go Chase-Topping
1
, Hen y Kia a
8
, Phil Toye
8
,
Koos Coe ze
7
, Ba end M deC B ons oo
6
and Oli ie Hano e
4
Abs ac
Backg ound: Posi i e mul i-locus he e ozygosi y- i ness co ela ions ha e been obse ed in a numbe o na u al
popula ions. They ha e been explained by he co ela ion be ween he e ozygosi y and inb eeding, and he
nega i e e ec o inb eeding on i ness (inb eeding dep ession). Exo ic in og ession in a locally adap ed popula ion
has also been ound o educe i ness (ou b eeding dep ession) h ough he b eaking-up o co-adap ed genes, o
he in oduc ion o non-locally adap ed gene a ian s.
In his s udy we examined he in e - ela ionships be ween genome-wide he e ozygosi y, in og ession, and dea h
o illness as a esul o in ec ious disease in a sample o cal es om an indigenous popula ion o Eas A ican
Sho ho n Zebu (c ossb ed Bos au us xBos indicus) in wes e n Kenya. These cal es we e obse ed om bi h o
one yea o age as pa o he In ec ious Disease in Eas A ican Li es ock (IDEAL) p ojec . Some o he cal es we e
ound o be gene ic hyb ids, esul ing om he ecen in og ession o Eu opean ca le b eed(s) in o he indigenous
popula ion. Eu opean ca le a e known o be less well adap ed o he in ec ious diseases p esen in Eas A ica. I
dea h and illness as a esul o in ec ious disease ha e a gene ic basis wi hin he popula ion, we would expec bo h
a nega i e associa ion o hese ou comes wi h in og ession and a posi i e associa ion wi h he e ozygosi y.
Resul s: In his indigenous li es ock popula ion we obse ed nega i e associa ions be ween he e ozygosi y and
bo h dea h and illness as a esul o in ec ious disease and a posi i e associa ion be ween Eu opean au ine
in og ession and episodes o clinical illness.
Conclusion: We obse e he e ec s o bo h inb eeding and ou b eeding dep ession in he Eas A ican Sho ho n
Zebu, and he e o e ind e idence o a gene ic componen o ulne abili y o in ec ious disease. These esul s
indica e ha he signi ican bu den o in ec ious disease in his popula ion could, in p inciple, be educed by
al e ed b eeding p ac ices.
Keywo ds: Inb eeding dep ession, He e ozygosi y, Fi ness, In ec ious disease, Exo ic in og ession, Local adap a ion,
Ca le
* Co espondence: [email p o ec ed]
1
Ins i u e o E olu iona y Biology, and Cen e o Immuni y, In ec ion and E olu ion,
Uni e si y o Edinbu gh, Edinbu gh EH9 3JT, UK
2
Depa men o Gene ics, Uni e si y o Camb idge, Camb idge, Downing S ee ,
Camb idge CB2 3EH, UK
Full lis o au ho in o ma ion is a ailable a he end o he a icle
© 2013 Mu ay e al.; licensee BioMed Cen al L d. This is an open access a icle dis ibu ed unde he e ms o he C ea i e
Commons A ibu ion License (h p://c ea i ecommons.o g/licenses/by/2.0), which pe mi s un es ic ed 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.
Mu ay e al. BMC E olu iona y Biology 2013, 13:246
h p://www.biomedcen al.com/1471-2148/13/246
Backg ound
Posi i e mul i-locus he e ozygosi y- i ness co ela ions
ha e been obse ed in a numbe o na u al popula ions
[1,2]. Genome-wide he e ozygosi y is co ela ed wi h
inb eeding wi hin a popula ion [3] and he e ozygosi y-
i ness co ela ions ha e he e o e been associa ed wi h
inb eeding dep ession [4]. Inb eeding may be de i-
men al o i ness i : (i) i ness ela ed ai s a e o e -
dominan , such ha he he e ozygous geno ype esul s
in g ea e i ness han ei he homozygo e, o (ii) he e
a e ecessi e alleles p esen in he popula ion ha e-
sul in lowe i ness [5]. The ex en o which he he -
e ozygosi y o gene ic ma ke s e lec s genome-wide
he e ozygosi y and inb eeding is a con en ious issue,
and he e ozygosi y- i ness co ela ions may esul om
a ew loci o majo e ec in linkage disequilib ium wi h
he gene ic ma ke s a he han he e ec o inb eeding
ac oss he genome [2,3,6,7]. Howe e , when la ge numbe s
o ma ke s a e used, es ima es o genome-wide he e ozy-
gosi y a e g ea ly imp o ed, a e mo e likely o co ela e
wi h inb eeding, and a e less likely o co ela e wi h he
i ness e ec s o a ew loci [2,3].
Inb eeding has been obse ed o ha e a nega i e e ec
on i ness- ela ed ai s in many popula ions [8], e.g.
[9,10], including some Eu opean ca le popula ions
[11,12]. In a numbe o na u al popula ions suscep ibil-
i y o in ec ious disease has been ound o be associa ed
wi h he e ozygosi y e.g. [13-15] and inb eeding e.g. [16].
In ec ious disease suscep ibili y in humans has been
ound o ha e a s ong gene ic componen in nume ous
s udies e iewed in [17], wi h many genes and gene ic
pa hways associa ed wi h suscep ibili y o di e en dis-
eases. The majo his ocompa ibili y complex (MHC) is
conside ed o play a signi ican ole in de e mining sus-
cep ibili y o a ious pa hogens [18], and MHC- a ge ed
b eeding p og ammes ha e been sugges ed o cap i e
popula ions [19,20]. In one na u al popula ion i was
ound ha he e ec o a single locus linked o he
MHC exceeded he e ec o genome-wide he e ozygosi y
on i ness [21].
Exo ic in og ession in a locally adap ed popula ion may
also educe i ness. A numbe o s udies ha e obse ed
a nega i e e ec o in og ession on i ness (ou b eeding
dep ession) [22,23]. These may esul om ei he (i) he
b eak-up o co-adap ed epis a ic in e ac ion be ween
genes, o (ii) he in oduc ion o non-locally adap ed
alleles [24].
In his s udy, we es ed o associa ions o he e ozygosi y
and exo ic Eu opean au ine b eed in og ession wi h
dea h and illness due o in ec ious disease du ing he i s
yea o li e, in a popula ion o Eas A ican sho ho n zebu
(EASZ) ca le (c ossb ed Bos au us xBos indicus) om
wes e n Kenya. Mo e han 500 cal es we e ollowed o
he i s yea o hei li e as pa o he In ec ious Disease
in Eas A ican Li es ock (IDEAL) p ojec [25]. The cal es
we e closely moni o ed h oughou ; dea hs and episodes
o clinical illness we e epo ed and ollowed up by ained
e e ina y s a . Du ing he s udy, mo e han 15% o he
cal es died, wi h he majo i y o dea hs due o in ec ious
diseases, mos commonly Eas Coas Fe e , helmin h in-
ec ions, hea wa e and ypanosomiasis [26]. Gi en he
hea y in ec ious disease bu den on hese ca le coupled
wi h minimal disease con ol o ea men , an abili y o
su i e in ec ion is likely o be signi ican ly associa ed wi h
i ness. Fu he mo e, mo e han hal he su i ing cal es
we e epo ed o ha e a leas one episode o clinical ill-
ness du ing he s udy pe iod, and almos all hese episodes
we e a ibu ed o in ec ious disease (see Me hods). We
p opose ha clinical illness is also indica i e o a ulne -
abili y o in ec ious disease ha could a ec i ness in his
en i onmen .
All cal es in he IDEAL popula ion we e geno yped
using he 50K Illumina® Bo ineSNP50 beadchip . 1
(55,777 SNPs be o e quali y con ol). These SNPs a e
ela i ely e enly dis ibu ed ac oss he genome, wi h an
a e age o 1,895 on each au osome ( anging om 1,009
o 3,553) and 1,362 on he X ch omosome (102 emain
unassigned).
The s udy popula ion was ound o be pa ly in o-
g essed by one o possibly mo e Eu opean au ine b eeds
[27,28] (de e mined om analysis o genome-wide SNP
da a using STRUCTURE [29]). This e lec s he e ec s o
a subs an ial b eed imp o emen p og amme, in ol ing
he c ossing o local ca le wi h Eu opean b eeds, ha ook
place in he mid-1990s wi h in e mi en c oss-b eeding
he ea e . As a esul , app oxima ely 20% o he cal es
show le els o in og ession consis en wi h c ossing wi h
Eu opean b eeds ≤5 gene a ions ago ( ac ion Eu opean
au ine > 2
-6
) [27].
We would also expec some inb eeding o occu in
his popula ion, since ma ing is la gely unmanaged, he e
is a ela i ely low numbe o b eeding bulls, and ca le
a e gene ally only anspo ed o e sho dis ances (only
a small ac ion o ca le a e aded and hese o e dis-
ances ypically jus 20-30 km). Also, his o ically, his e-
gion has expe ienced majo inde pes epidemics, which
ha e educed he e ec i e popula ion size o he EASZ
[28]. In his s udy we calcula e he genome-wide he e o-
zygosi y o he cal es in he IDEAL popula ion using
he same SNP da a se as was used o calcula e exo ic
in og ession [28].
In his s udy we aim o de e mine whe he inb eeding
(as es ima ed by SNP he e ozygosi y) and ou b eeding
(as es ima ed by he es ima ed p opo ion o gene ic
ma ke s ha descend om a Eu opean au ine b eed)
a ec he likelihood o dea h o illness due o in ec ious
disease and he e o e whe he he e is e idence o a gen-
e ic basis o ulne abili y o in ec ious disease in EASZ.
Mu ay e al. BMC E olu iona y Biology 2013, 13:246 Page 2 o 8
h p://www.biomedcen al.com/1471-2148/13/246
We use esul s om ea lie s udies ha ha e de e mined
he deg ee o Eu opean au ine in og ession [28] and
desc ibe and analyse he causes o ins ances o dea h and
illness o all he cal es in he IDEAL p ojec [25,26].
Th ough analysis o hese da a we ind associa ions o
bo h he e ozygosi y and Eu opean au ine in og ession
wi h ulne abili y o in ec ious disease du ing he i s yea
o li e in his popula ion o EASZ.
Resul s
The sample size a ailable o analysis was 518 cal es (see
Me hods o inclusion/exclusion c i e ia), o which 68
(13%) died o in ec ion- ela ed disease wi hin one yea o
bi h (see Me hods o mo e de ails). O he su i o s,
episodes o clinical illness (almos always ela ed o in-
ec ions) we e epo ed in 243 (54%) (Table 1, Addi ional
ile 1: Table S1).
The dis ibu ions o obse ed he e ozygosi y alues o
non-in og essed (de ined as < 2
-6
Eu opean au ine) and
in og essed cal es (> 2
-6
Eu opean au ine) a e shown in
Figu e 1. The mean he e ozygosi y o non-in og essed
cal es is 0.28 (wi h a s anda d de ia ion (SD) o 0.01). This
is no an unusual le el o he e ozygosi y o a ca le popu-
la ion. Hols ein-F iesians (Eu opean au ine) ha e been
ound o ha e a e age he e ozygosi y o 0.33 (SD = 0.01),
Je sey (Eu opean au ine) ha e an a e age he e ozygosi y
o 0.25 (SD = 0.03), N’dama (A ican au ine) ha e an a e -
age he e ozygosi y o 0.17 (SD = 0.03) and E hiopian Sheko
(admixed A ican au ine and Asian zebu) ha e an a e age
he e ozygosi y o 0.26 (SD = 0.0003) [27].
Ino de o es whe he o no ou es ima eso
he e ozygosi y a e likely o e lec ue genome-wide
he e ozygosi y,we andomlysampledou SNPswi h-
ou eplacemen 100 imes and de e mined he deg ee
o co ela ion in he he e ozygosi y es ima es o cal es
based on one hal o he ma ke s and he es ima es
based on he o he hal [2,6]. We ound a consis en ly
s ong co ela ion be ween he es ima es (mean Pea son’s
= 0.98, SD = 0.0012, ange = 0.977-0.982). Since some
SNPs may be in linkage disequilib ium wi h one ano he ,
we epea ed he analysis wi h wo smalle subse s (each
ep esen ing a di e en andom sample, wi hou e-
placemen , o one 10
h
o he o al SNPs). E en wi h his
much smalle se o ma ke s, a s ong co ela ion
be ween he es ima es was ound (mean Pea son’s =0.91,
SD = 0.0053, ange = 0.89-0.92).
The non-in og essed cal es show a skewed dis ibu ion
o he e ozygosi y alues, wi h he majo i y (93%) con-
ained wi hin a nea -symme ical dis ibu ion ± 0.010
om a mode o 0.284 (Figu e 1). Howe e , he e is a long
ail o 28 cal es wi h lowe alues. We ca ego ise cal es
wi h alues < 0.274 as “low”he e ozygosi y, assumed o e-
lec a highe deg ee o inb eeding han he popula ion
no m. In og essed cal es show a di e en dis ibu ion o
he e ozygosi y alues ha a e gene ally highe han hose
o non-in og essed cal es (Figu e 1) (Mann Whi ney U
es , P<0.001).
The ela ionship be ween he e ozygosi y and in og es-
sion is shown in Figu e 2. As expec ed, he e is a clea in-
c ease in he e ozygosi y wi h inc easing in og ession.
Howe e , in addi ion o he 28 non-in og essed cal es
wi h low he e ozygosi y, he e a e se e al in og essed
cal es wi h he e ozygosi y alues ha appea low ela i e
o he amoun o in og ession. In o de o in es iga e he
e ec o a ia ion in he e ozygosi y on he i ness o in o-
g essed cal es independen ly o he e ec o in og ession
we de ine a g oup o cal es wi h low he e ozygosi y ela-
i e o hei deg ee o in og ession. We de ine his g oup
by inding he bes i linea eg ession and i e a i ely ex-
cluding cal es wi h he e ozygosi y alues > 0.01 below
hei expec ed alue. This p ocedu e esul s in he i ed
eg ession line shown in Figu e 2 (wi h he -0.01 h eshold
also indica ed). Using his c i e ion, 9 o he 113 in o-
g essed cal es a e ca ego ised as low he e ozygosi y. Two
o hese 9 died and 6 o he s expe ienced a leas one clin-
ical episode (see Table 1).
We compa ed bo h cal es ha died and hose ha
su i ed bu expe ienced any episode o clinical illness
o a e e ence ca ego y o heal hy cal es ( hose ha su -
i ed and had no epo ed episode o clinical illness).
We conside ed wo ca ego ical p edic o s o hese ou -
comes, low he e ozygosi y and being in og essed, using
a mul inomial logis ic eg ession (desc ibed in Me hods;
esul s in Table 2). We used hese ca ego ical p edic-
o s, a he han analysing he es ima es di ec ly, due o
he non-no mali y o he da a ( he dis ibu ion o alues
o bo h he e ozygosi y and in og ession a e highly
skewed) and o allow us o obus ly accoun o he
Table 1 Numbe s o cal es ha all in o he di e en ca ego ies used o he s a is ical analysis o he e ec s o
he e ozygosi y and in og ession on likelihood o dea h and clinical illness
Low he e ozygosi y High he e ozygosi y
In og essed Non-in og essed In og essed Non-in og essed
Died 2 111144
Su i ed wi h clinical episode 6 13 64 160
Su i ed wi hou clinical episode 1 4 29 173
Cu -o s o high/low he e ozygosi y and in og essed/non-in og essed a e de ined in Me hods.
Mu ay e al. BMC E olu iona y Biology 2013, 13:246 Page 3 o 8
h p://www.biomedcen al.com/1471-2148/13/246
ela ionship be ween he e ozygosi y and in og ession
(Figu es 1 and 2, Addi ional ile 1: Table S1).
We i s conside he associa ion be ween p edic o s and
cal dea h due o in ec ious disease. Fo non-in og essed
cal es ( ac ion Eu opean au ine < 2
-6
) only he e is s ong
e idence o a ela ionship be ween dea h and low he e ozy-
gosi y: odds a io (OR) = 10.8 (95% CIs 3.0-48.1); P< 0.001
(Table 3). Fo all cal es, his ela ionship emains when
bo h low he e ozygosi y and in og ession a e included as
p edic o s: OR = 9.5 (3.5-27.9); P<0.001(Table2).The eis
no e idence o an in e ac ion be ween low he e ozygosi y
and in og ession (P> 0.05) and he e is no e idence o an
associa ion be ween dea h and deg ee o in og ession
(OR = 1.4 [0.7-2.9]; P= 0.39) (Table 2). No signi ican
di e ences we e ound in he na u e o in ec ious disease-
ela ed dea hs among he low he e ozygosi y cal es. These
esul s a e obus o a ia ion in he cu -o o in o-
g essed cal es (see Me hods).
Clinical illness is a less well de ined pheno ype han
dea h, since i includes a wide ange o disease signs and i
canno be unequi ocally de e mined whe he o no an in-
s ance o clinical illness was he esul o in ec ious disease
o no (al hough i is likely ha he o e whelming majo i y
we e). Despi e his, clinical illness and dea h a e clea ly
ela ed ou comes. Mo eo e , because many mo e cal es
expe ienced clinical illness du ing hei i s yea han
died, his ou come may p o ide inc eased powe o de-
ec small e ec s. Fo non-in og essed cal es he e is
Figu e 1 His og am o equency o he e ozygosi y alues o Eu opean au ine in og essed ( ed) and non-in og essed (blue) Eas
A ican Sho ho n Zebu cal es.
Figu e 2 Plo o he e ozygosi y agains p opo ion Eu opean in og ession o all cal es in ou analysis. Cal es ha died a e highligh ed
( ed symbols). The bes i ela ionship be ween he e ozygosi y (H) and in og ession (I) o non-inb ed cal es is shown ( hick line: eg ession
equa ion H = 0.284 (95% con idence in e als ±0.0003) + I*0.312 (±0.006), R
2
= 0.972). The cu -o o de ine ‘low’he e ozygosi y cal es a ies wi h
in og ession, since he expec ed o mean he e ozygosi y a ies wi h in og ession. Cal es ha a e mo e ha 0.01 below he expec ed he e ozygosi y
(blue dashed line) o hei amoun o in og ession a e de ined as ‘low’he e ozygosi y (see Resul s and Me hods). The o ange dashed line
indica es he cu -o o in og essed cal es (2
-6
%).
Mu ay e al. BMC E olu iona y Biology 2013, 13:246 Page 4 o 8
h p://www.biomedcen al.com/1471-2148/13/246
e idence o a ela ionship be ween illness and low he -
e ozygosi y: OR = 3.5 (95% CIs 1.05-15.1); P= 0.026
(Table 3). Fo all cal es, his ela ionship emains when
bo h low he e ozygosi y and in og ession a e included
as p edic o s: OR = 3.4 [1.3-9.3]; P= 0.018 (Table 2).
Howe e , o his pheno ype he e is also s ong e idence
o an associa ion wi h in og ession, wi h in og essed
cal es being mo e likely o expe ience illness (OR = 2.4
[1.5-3.8]; P< 0.001). Again he e is no e idence o any
in e ac ion be ween low he e ozygosi y and in og ession
(P> 0.05). These esul s a e obus o a ia ion in he cu -
o o in og essed cal es (see Me hods).
Discussion
Ou esul s indica e ha he e is a gene ic basis o ulne -
abili y o in ec ious disease in Eas A ican Sho ho n
Zebu. We ind a posi i e associa ion be ween low he e o-
zygosi y and bo h dea h and illness. Since we use a eason-
able p oxy o genome-wide he e ozygosi y (40,457 SNPs
ac oss all au osomes), and he es ima es o genome-wide
he e ozygosi y a e highly co ela ed o di e en subse s
o SNPs, we conside i likely ha ou esul s a e indica-
i eo hep esenceo inb eedingdep essionin he
s udy popula ion, a he han being a consequence o
localised associa ions be ween i ness and loci o la ge
e ec [6,7,17].
Ou esul s indica e he p esence o i ness- ela ed geno-
ypes ha a e ei he o e -dominan o ha e ecessi e dele-
e ious alleles in he popula ion. Al hough he p opo ion
o cal es ha we classi ied as “low”he e ozygosi y was
small (app oxima ely 7%), he e ec on in ec ious disease-
ela ed mo ali y was la ge (OR ~ 9) and he e was also an
e ec on clinical illness (OR ~ 3). Ou esul s gi e popula-
ion a ibu able ac ions ( o he non-in og essed popu-
la ion) due o low he e ozygosi y o 14% o mo ali y and
4% o clinical illness [30]. In p inciple, he e o e, cases o
illness and dea hs could be educed by be e managemen
p ac ices (e.g. bull o a ion).
We ound e idence ha Eu opean au ine in og ession
a ec ed heal h as i was a signi ican p edic o o clinical
illness (Table 2). As he e we e only 13 dea hs among
hose wi h ac ional in og ession g ea e han 2
-6
we had
e y li le powe o de ec any e ec on dea h alone. None-
heless, ou esul s o hese wo pheno ypes a e no in-
consis en , as he con idence in e als o he odds a ios
o e lap (see Table 2). We sugges ha clinical illness is a
ma ke o ulne abili y in his popula ion and conse-
quen ly ha Eu opean au ine in og essed cal es a e
mo e ulne able in his sense.
We could ind no non-gene ic con ounding ac o s
ha could be d i ing he obse ed associa ions be ween
in og ession and ei he dea h o illness. The in o-
g essed cal es (e en hose wi h high in og ession)
could no be dis inguished om non-in og essed cal es
by he ield eam [25]; exo ic in og ession was only dis-
co e ed upon gene ic analysis o he cal es. And, al hough
he in og essed cal es we e no ound uni o mly dis ib-
u ed among he 20 sub-loca ions o he s udy [28], he e is
no e idence o geog aphic a ia ion in he a e o mo al-
i y o o clinical illness.
Since he EASZ ances o s o he s udy popula ion ha e
been p esen in Eas A ica o many gene a ions [28], i
migh be expec ed ha hey may ha e in some way
adap ed o he speci ic challenges posed by ha en i on-
men , in pa icula o he in ec ious diseases p esen in
ha egion. In ec ious disease was he p ima y cause o
i s yea mo ali y and illness in his popula ion, and
hus he abili y o wi hs and in ec ion is likely o be o
signi ican selec i e ad an age. On he o he hand, he
in og essed animals descend, in pa , om Eu opean
ca le, which ha e no had he same his o y o selec i e
p essu es. I has been ound ha Eu opean b eeds a e
mo e suscep ible han local b eeds o some in ec ious
diseases ha a e p e alen in A ica, such as Eas Coas
Fe e [31,32], bo ine ube culosis [33,34], and ypano-
somiasis [35,36]. Thus, he Eu opean b eeds a e, in his
impo an espec , less well adap ed o he Eas A ican
Table 2 Summa y o esul s o a mul inomial logis ic eg ession o he e ec s o genome-wide he e ozygosi y and
in og ession (as wo le el ac o s) on he likelihood o dea h and he likelihood o clinical illness
Pheno ype Dead cal es (n = 68) Ill cal es (n = 243)
Fac o s Es ima e (SE) OR (95% CI) P- alue Es ima e (SE) OR (95% CI) P- alue
Low he e ozygosi y 1.13 (0.27) 9.5 (3.5-27.9) < 0.001 0.61 (0.26) 3.4 (1.2-9.3) 0.018
In og essed 0.16 (0.19) 1.4 (0.7-2.9) 0.39 0.43 (0.12) 2.4 (1.5-3.8) < 0.001
Cu -o s o high/low he e ozygosi y and in og essed/non-in og essed a e de ined in Me hods. Heal hy cal es we e he e e ence ca ego y. CI = con idence
in e al. SE = s anda d e o . OR = odds a io. P- alues in bold ype a e s a is ically signi ican .
Table 3 Summa y o esul s o Fishe ’s exac es s o he
e ec s o low genome-wide he e ozygosi y (as a
wo-le el ac o ) on he likelihood o dea h and he
likelihood o clinical illness o he sub-popula ions o
non-in og essed and in og essed cal es
Pheno ypes Dead s heal hy cal es Ill s heal hy cal es
Subpopula ion OR 95% CI P- alue OR 95% CI P- alue
Non-in og essed 10.8 3.0-48.1 < 0.001 3.5 1.05-15.1 0.026
In og essed 5.3 0.24-321 0.213 2.72 0.31-129 0.671
Cu -o s o high/low he e ozygosi y and in og essed/non-in og essed a e
de ined in Me hods. OR = odds a io. CI = con idence in e al. P- alues in bold
ype a e s a is ically signi ican .
Mu ay e al. BMC E olu iona y Biology 2013, 13:246 Page 5 o 8
h p://www.biomedcen al.com/1471-2148/13/246

en i onmen , and in og essed animals, while mo e
he e ozygous due o ou b eeding, a e also less locally
adap ed. While inc eased he e ozygosi y is likely o be
bene icial i i ness ela ed ai s a e o e -dominan o
he e a e ecessi e dele e ious alleles in he popula ion,
he b eaking up o o loss o locally adap ed geno ypes is
likely o be de imen al o i ness. Bo h hese compe ing
o ces appea o be a wo k in his popula ion o Eas
A ican Sho ho n Zebu cal es in wes e n Kenya.
Conclusions
In conclusion, in his indigenous ca le popula ion om
he opics we ind s ong associa ions be ween genome-
wide he e ozygosi y and bo h dea h and clinical illness due
o in ec ious diseases. The e is e idence ha in og ession
o Eu opean au ine b eeds also inc eases ulne abili y o
in ec ious diseases. We he e o e belie e ha we ha e ob-
se ed he e ec o bo h inb eeding and ou b eeding de-
p ession, and ha e ound e idence o a gene ic componen
o ulne abili y o in ec ious disease and o adap a ion in
he Eas A ican Sho ho ned Zebu popula ion in wes e n
Kenya. This implies ha he signi ican bu den o in ec-
ious disease in his popula ion could, in p inciple, be
educed by be e b eeding p ac ices.
Me hods
Da a se
The IDEAL s udy popula ion consis ed o 548 cal es,
which we e moni o ed closely h oughou hei i s yea
o li e [25]. Th ee o hese cal es we e excluded om
his analysis because hey ecei ed ea men o disease.
The cal es we e sampled om a egion in wes e n
Kenya (o app oxima ely 45 × 90 km) (co e ing some o
he Busia, Teso, Siaya, Bu e e/Mumias and Bungoma
dis ic s). This egion a e sed ou di e en ypes o
ag o-ecological zones. The egion was di ided in o sub-
loca ions o app oxima ely 10 × 10 km, o which 20 we e
andomly selec ed (wi hin each ag o-ecological zone) o
s udy [25]. Small-scale ca le a ming is common in his
egion and he sub-loca ions con ain app oxima ely 80
o 90 households, o which app oxima ely 60% own ca le
(gene ally anging om 1-10 ca le pe household). New
bo n cal es we e ec ui ed in o he s udy om hese 20
sub-loca ions o e a pe iod o 2 yea s (2007-2009). Cal es
ha showed isible signs o exo ic b eed in og ession
we e excluded, and only cal es om small a ms, whe e
he ca le spen a leas pa o hei ime ou side, exposed
o in ec ious disease, we e ec ui ed [25].
Gene ic da a
All cal es in he IDEAL popula ion we e geno yped
using he 50K Illumina® Bo ineSNP50 beadchip . 1 in-
cluding 55,777 SNPs be o e quali y con ol. S anda d
quali y con ol p ocedu es we e applied o he da a p io
o analysis e iewed in [37]. This was implemen ed using
GenABEL. A mino allele equency o SNPs cu -o o
1% was applied and a SNP call- a e cu -o o 90%. A
cu -o o call a e o an animal o 90% was applied.
And an iden i y by s a e (IBS) h eshold o 90% and a
cu -o odds o 1000-1 was applied o decide whe he a
cal should be excluded based on sex/X-linked ma ke
da a inconsis ency (which bo h elimina e e o s due o
sample mis-iden i ica ion). These quali y con ol h esh-
olds le a o al o 40,457 au osomal SNPs o subsequen
analysis. A Ha dy-Weinbe g equilib ium (HWE) cu -o
was no applied, since a di e gence om HWE can esul
om ac o s o he han geno yping e o (in his case
in og ession).
As a esul o quali y con ol, 9 cal es and 13,856 o
he 54,313 ini ial au osomal SNPs had o be excluded
om he 50K SNP chip da a se due o unsuccess ul
geno yping o mis-iden i ica ion o indi iduals. This e-
duced he geno yping da a se o his s udy o 536 cal es
and 40,457 au osomal SNPs.
Fi ness pheno ype –dea h
The bes p oxy o i ness in his s udy is su i al o i s
yea . Eigh y-eigh o he o iginal 548 cal es died du ing
he s udy, and all bu nine o hese we e subjec o a de-
ailed pos -mo em examina ion [25,26]. In mos cases
he cause o dea h was ound o be in ec ious disease.
These esul s a e desc ibed in de ail elsewhe e [25,26],
bu he mos equen causes o dea h we e Eas Coas
Fe e (caused by he ick-bo ne p o ozoan Theile ia
pa a), hea wa e (caused by he ick-bo ne icke sia
Eh lichia uminan ium) and helmin hosis (especially
associa ed wi h he nema ode Haemonchus placei). O
he 88 dea hs, 9 we e ound o ha e causes unlikely o
be associa ed wi h any he i able i ness, such as auma
and s a a ion, and 10 had an unde e mined cause o
dea h. These cal es we e excluded om he analysis, so
as o make he pheno ype mo e likely o ha e a consis -
en (al hough likely complex) gene ic basis. This esul ed
in a da a se , a e quali y con ol, o 68 cal es ha died
due o in ec ious disease, and 450 cal es ha su i ed
hei i s yea (Table 1, Addi ional ile 1: Table S1).
Follow-up s udies indica ed ha he dea h a e o e he
nex 1-2 yea s was much lowe han du ing he i s
yea .
Fi ness pheno ype –clinical illness
All cal es we e moni o ed o signs o clinical illness
h oughou he s udy. Episodes o clinical illness we e e-
po ed by he cal ’s owne , isi ing animal heal h wo ke s
o he p ojec s a and we e subjec o de ailed in es iga-
ion using a s anda dised p o ocol and, whe e necessa y,
ollowed up by he p ojec ’s e e ina y s a [25]. O he
su i ing cal es in he s udy, 243 expe ienced a leas one
Mu ay e al. BMC E olu iona y Biology 2013, 13:246 Page 6 o 8
h p://www.biomedcen al.com/1471-2148/13/246
epo ed episode o clinical illness. Clinical signs anged
om mild o se e e. T ea men was p o ided i he cal
was in dis ess and i s keepe ag eed; ea ed cal es (n = 3)
we e subsequen ly excluded om he s udy. In a ew in-
s ances a se e ely ill cal was eu hanized; his was ega ded
as equi alen o a dea h. Almos all episodes o clinical ill-
ness could be a ibu ed o in ec ions, howe e causal links
a e di icul o de ini i ely es ablish and so episodes canno
be unequi ocally di ided in o hose ha we e caused by
in ec ious disease and hose ha we e no . Fo his eason
all episodes o clinical illness we e used o dis inguish be-
ween heal hy and unheal hy cal es (whe e his di e ence
in heal h can gene ally be a ibu ed o a ia ion in sus-
cep ibili y o in ec ious disease). This di ision esul s in
207 cal es ha nei he died no expe ienced any e-
po ed episodes o clinical illness up o one yea old.
Eu opean au ine in og ession
Using a cu -o o ac ion 2
-6
in og ession we iden i-
ied 113 in og essed cal es, wi h a maximum o 36.1%
Eu opean au ine in og ession. The in og essed cal es
we e no ound uni o mly dis ibu ed among he 20
sub-loca ions o he s udy: hey a e ound mainly in 12
sub-loca ions in he no he n and cen al egions o he
s udya ea[28].Themos likelyo igino hisEu opean
au ine in og ession in he s udy a ea is a numbe o
ecen and con inuing b eed imp o emen p og ammes
ha use exo ic animals and semen, and ca le ma ke s
in he s udy a ea con inue o sell c ossb ed ca le [28].
Themos ac i ep og amme ookplacein hemid
1990s, equi alen o app oxima ely 5 EASZ gene a ions
be o e ou s udy ook place. This is e lec ed in ou cu -
o o ac ion2
-6
in og ession, co esponding o he
expec ed le el o in og ession a e 5 gene a ions o
back-c ossing wi h indigenous ca le since he F1 gene -
a ion. Howe e , o con i m he obus ness o he p ecise
alue o he cu -o , we epea ed all analyses using al e na-
i e cu -o alues bu ob ained e y simila esul s.
He e ozygosi y
Indi idual mean genome-wide he e ozygosi y was com-
pu ed as he p opo ion o all au osomal geno yped loci
ha passed quali y con ol (40,457 SNPs) ha we e he -
e ozygous. This was implemen ed using GenABEL [38].
To es o he whe he his is likely o be a good es ima e
o genome-wide he e ozygosi y, we andomly sampled
SNPs wi hou eplacemen 100 imes and calcula ed he
Spea man co ela ion coe icien o he he e ozygosi y es-
ima es o cal es based on one hal o he ma ke s and
he es ima es based on he o he hal [2,6]. Since SNPs
may in linkage disequilib ium wi h one ano he we e-
pea ed he analysis compa ing smalle subse s.
S a is ical me hods
The compa isons be ween heal hy cal es (n = 217) and
hose ha died as a esul o in ec ious disease (n = 68)
o su i ed bu expe ienced clinical illness (n = 243) was
in es iga ed using a 3-le el mul inomial logis ic eg ession
analysis using P oc Logis ic SAS e sion 9.3 (SAS Ins i u e
Inc., Ca y, NC). Bo h he e ozygosi y and in og ession
we e included as p edic o s and we e ca ego ised o
he pu poses o analysis. Using a cu -o o ac ion 2
-6
in og ession we iden i ied 113 in og essed cal es. We
epea ed all analyses using al e na i e cu -o s o 2
-7
,2
-5
and 2
-4
(6, 4 and 3 gene a ions o c ossing espec i ely).
We ca ego ised non-in og essed cal es wi h he e ozy-
gosi y < 0.274 (mo e han 0.10 below he mean o his
g oup) as low he e ozygosi y, assumed o e lec a highe
deg ee o inb eeding han he popula ion no m (28 cal es).
We de ined low he e ozygosi y cal es in he in og essed
g oup by inding he bes i linea eg ession o he e ozy-
gosi y agains in og ession and i e a i ely excluding cal es
wi h he e ozygosi y alues > 0.01 below hei expec ed
alue ( esul ing in 9 o he 113 in og essed cal es being
ca ego ised as low he e ozygosi y).
We conduc ed con i ma o y uni a ia e analyses using
Fishe ’s exac es (S a Xac e sion 8, Cy el So wa e
Co p, Camb idge, MA, USA) o he non-in og essed
and in og essed ca ego ies, allowing us o in es iga e
he e ec o he e ozygosi y o hese sepa a ely.
A ailabili y o suppo ing da a
Geno yping da a can be ound a D yad (h p://da ad yad.
o g/ esou ce/doi:10.5061/d yad.bc598).
Addi ional ile
Addi ional ile 1: Table S1. Table desc ibing he e ozygosi y (He ),
p opo ion Eu opean au ine (%ET), whe he hey died du ing he s udy
(and whe he his is known o ha e been he esul o an in ec ious
disease) and whe he hey expe ienced one o mo e clinical episodes, o
all o he cal es included in his analysis.
Compe ing in e es s
The au ho s decla ed ha hey ha e no compe ing in e es s.
Au ho s’con ibu ions
MW, OH, HK, PT, KC and MB designed he IDEAL p ojec . GM, MW and OH
d a ed he manusc ip . GM calcula ed es ima es o he e ozygosi y. MM and
MT calcula ed es ima es o in og ession. TS comple ed he SNP geno yping
and SNP ma ke sco ing. MCT, MW and GM did he s a is ical analysis. ST, AJ
and IC did ieldwo k o he IDEAL p ojec , pe o med au opsies and
diagnos ic es s. All au ho s ead and app o ed he inal manusc ip .
Acknowledgemen s
The IDEAL p ojec was unded by he Wellcome T us (g an e e ence
WT079445MF). USDA-ARS bo ine unc ional labo a o y and Geneseek
e e ina y diagnos ics p o ided in aluable echnical assis ance o he
geno yping o he samples.
This p ojec was also pa ially suppo ed by USDA Ag icul u al Resea ch
Se ice p ojec 1265-31000-104-00D. Men ion o ade names o comme cial
p oduc s in his a icle is solely o he pu pose o p o iding speci ic
Mu ay e al. BMC E olu iona y Biology 2013, 13:246 Page 7 o 8
h p://www.biomedcen al.com/1471-2148/13/246
in o ma ion and does no imply ecommenda ion o endo semen by he
U.S. Depa men o Ag icul u e.
Au ho de ails
1
Ins i u e o E olu iona y Biology, and Cen e o Immuni y, In ec ion and
E olu ion, Uni e si y o Edinbu gh, Edinbu gh EH9 3JT, UK.
2
Depa men o
Gene ics, Uni e si y o Camb idge, Camb idge, Downing S ee , Camb idge
CB2 3EH, UK.
3
MTT Ag i ood Resea ch Finland, Bio echnology and Food
Resea ch, Jokioinen FI-31600, Finland.
4
School o Li e Sciences, Uni e si y o
No ingham, Uni e si y Pa k, No ingham NG7 2RD, UK.
5
Uni ed S a es
Depa men o Ag icul u e, Ag icul u al Resea ch Se ice, Bo ine Func ional
Genomics Labo a o y, 10300 Bal imo e A enue, Bel s ille, MD 20705, USA.
6
Roslin Ins i u e, Eas e Bush, Uni e si y o Edinbu gh, Roslin, Midlo hian,
EH25 9RG, Edinbu gh, UK.
7
Depa men o Ve e ina y T opical Diseases,
Facul y o Ve e ina y Science, Uni e si y o P e o ia, P i a e bag X04,
Onde s epoo , Sou h A ica.
8
In e na ional Li es ock Resea ch, P.O. Box
30709, Nai obi 00100, Kenya.
9
Cu en add ess: Paul G Allen School o Global
Animal Heal h, Washing on S a e Uni e si y, Pullman, WA 99164-7079, USA.
Recei ed: 2 May 2013 Accep ed: 4 No embe 2013
Published: 9 No embe 2013
Re e ences
1. Col man DW, Sla e J: Mic osa elli e measu es o inb eeding: a
me a-analysis. E olu ion 2003, 57:971–983.
2. Chapman JR, Nakagawa S, Col man DW, Sla e J, Sheldon BC: A quan i a i e
e iew o he e ozygosi y- i ness co ela ions in animal popula ions.
Mol Ecol 2009, 18:2746–2765.
3. Sla e J, Da id P, Dodds KG, Veen lie BA, Glass BC, B oad TE, McEwan JC:
Unde s anding he ela ionship be ween he inb eeding coe icien and
mul ilocus he e ozygosi y: heo e ical expec a ions and empi ical da a.
He edi y 2004, 93:255–265.
4. Kelle LF, Walle DM: Inb eeding e ec s in wild popula ions. T ends Ecol
E ol 2002, 17:230–241.
5. Cha leswo h D, Cha leswo h B: Inb eeding dep ession and i s
e olu iona y consequences. Annu Re Ecol Sys 1987, 18:237–268.
6. Balloux F, Amos W, Coulson T: Does he e ozygosi y es ima e inb eeding
in eal popula ions? Mol Ecol 2004, 13:3021–3031.
7. Pembe on J: Measu ing inb eeding dep ession in he wild: he old ways
a e he bes . T ends Ecol E ol 2004, 19:613–615.
8. C nok ak P, Ro DA: Inb eeding dep ession in he wild. He edi y 1999,
83:260–270.
9. Walling CA, Nussey DH, Mo is A, Clu on-B ock TH, K uuk LEB, Pembe on
JM: Inb eeding dep ession in ed dee cal es. BMC E ol Biol 2011, 11:318.
10. Col man DW, Pilking on JG, Smi h JA, Pembe on JM: Pa asi e-media ed
selec ion agains inb ed soay sheep in a ee-li ing, island popula ion.
E olu ion 1999, 53:1259–1267.
11. Sewalem A, Kis emake GJ, Miglio F, Van Doo maal BJ: Analysis o
inb eeding and i s ela ionship wi h unc ional longe i y in canadian
dai y ca le. J Dai y Sci 2006, 89:2210–2216.
12. Thompson JR, E e e RW, Hamme schmid NL: E ec s o inb eeding on
p oduc ion and su i al in hols eins. J Dai y Sci 2000, 83:1856–1864.
13. Luika G, Pilg im K, Vis y J, Ezenwa VO, Schwa z MK: Candida e gene
mic osa elli e a ia ion is associa ed wi h pa asi ism in wild bigho n
sheep. Biol Le 2008, 4:228–231.
14. Lyons EJ, Amos W, Be kley JA, Mwangi I, Sha i M, Williams TN, New on CR,
Peshu N, Ma sh K, Sco JAG, Hill AV: Homozygosi y and isk o childhood
dea h due o in asi e bac e ial disease. BMC Med Gene 2009, 10:55.
15. Rijks JM, Ho man JI, Kuiken T, Os e haus ADME, Amos W: He e ozygosi y
and lungwo m bu den in ha bou seals (Phoca i ulina). He edi y 2008,
100:587–593.
16. Bu che AR, Pale ho pe HM, G o e DI: The suscep ibili y o inb ed mice o
in ec ion wi h B achylaima c ibbi (Digenea: B achylaimidae). Pa asi ol In
2002, 51:109–115.
17. Chapman SJ, Hill AVS: Human gene ic suscep ibili y o in ec ious disease.
Na Re Gene 2012, 13:175–188.
18. Ho on R, Wilming L, Rand V, Lo e ing RC, B u o d EA, Khodiya VK, Lush MJ,
Po ey S, Talbo CC, W igh MW, Wain HM, T owsdale J, Ziegle A, Beck S:
Gene map o he ex ended human MHC. Na Re Gene 2004, 5:889–899.
19. Hughes AL: MHC polymo phism and he design o cap i e b eeding
p og ams. Conse Biol 1991, 5:249–251.
20. Wedekind C: Sexual selec ion and li e-his o y decisions: Implica ions o
suppo i e b eeding and he managemen o cap i e popula ions.
Conse Biol 2002, 16:1204–1211.
21. Lenz TL, Muelle B, T illmich F, Wol JBW: Di e gen allele ad an age a
MHC-DRB h ough di ec and ma e nal geno ypic e ec s and i s
consequences o allele pool composi ion and ma ing. P oc R Soc B 2013,
280:176.
22. Fou nie -Le el A, Ko e A, Coope MD, No dbo g M, Schmi J, Wilczek AM:
A Map o Local Adap a ion in A abidopsis haliana.Science 2011,
334:86–89.
23. Olano-Ma in J, Muelle JC, Kempenae s B: He e ozygosi y and su i al in
blue i s (Cyanis es cae uleus): con as ing e ec s o p esumably
unc ional and neu al loci. Mol Ecol 2011, 20:4028–4041.
24. Szulkin M, Da id P: Nega i e he e ozygosi y- i ness co ela ions obse ed
wi h mic osa elli es loca ed in unc ional a eas o he genome. Mol Ecol
2011, 20:3949–3952.
25. de C B ons oo BM, Thumbi SM, Poole EJ, Kia a H, Tosas Augue O, Handel
IG, Jennings A, Con adie I, Mbole-Ka iuki MN, Toye PG, Hano e O, Coe ze J,
Woolhouse ME: Design and desc ip i e epidemiology o he In ec ious
Diseases o Eas A ican Li es ock (IDEAL) p ojec , a longi udinal cal
coho s udy in wes e n Kenya. BMC Ve Res 2013, 9:171.
26. Thumbi S, B ons oo B, Kia a H, Toye P, Poole J, Mbole-Ka iuki MN, an Wyk
Con adie I, Jennings AE, Handel I, Coe ze JAW, S eyl JCA, Hano e O,
Woolhouse MEJ: Pa asi e co-in ec ions and hei impac on su i al o
indigenous ca le. PLoS One. In P ess.
27. Mbole-Ka iuki MN: Genomic di e si y o Eas A ican Sho ho n zebu in
Wes e n Kenya, PhD Thesis. 2012.
28. Mbole-Ka iuki MN, Sons ega d T, O h A, Thumbi S, B ons oo B, Kia a H,
Toye P, Con adie I, Jennings AE, Coe ze K, Woolhouse MEJ, Hano e O,
Tapio M: Genome–wide analysis e eal he ancien and ecen gene ic
admix u e o Eas A ican Sho ho n Zebu, Unde e iew.
29. P i cha d JK, S ephens M, Donnelly P: In e ence o Popula ion S uc u e
Using Mul ilocus Geno ype Da a. Gene ics 2000, 155:945–959.
30. Dohoo I, Ma in W, S yhn H: Ve e ina y Epidemiologic Resea ch. 2nd edi ion.
Canada: AVC; 2009.
31. Jensen K, Pax on E, Wadding on D, Talbo R, Da ghou h MA, Glass EJ:
Di e ences in he ansc ip ional esponses induced by Theile ia
annula a in ec ion in bo ine monocy es de i ed om esis an and
suscep ible ca le b eeds. In J Pa asi ol 2008, 38:313–325.
32. Glass EJ, P es on PM, Sp ingbe A, C aigmile S, Ki a E, Wilkie G, B own
CGD: Bos au us and Bos indicus (Sahiwal) cal es espond di e en ly o
in ec ion wi h Theile ia annula a and p oduce ma kedly di e en le els
o acu e phase p o eins. In J Pa asi ol 2005, 35:337–347.
33. Vo de meie M, Ameni G, Be g S, Bishop R, Robe son BD, Ase a A,
Hewinson RG, Young DB: The in luence o ca le b eed on suscep ibili y
o bo ine ube culosis in E hiopia. Comp Immunol Mic obiol In ec Dis
2012, 35:227–232.
34. Ameni G, Ase a A, Enge s H, Young D, Go don S, Hewinson G, Vo de meie
M: High p e alence and inc eased se e i y o pa hology o bo ine
ube culosis in Hols eins compa ed o zebu b eeds unde ield ca le
husband y in cen al E hiopia. Clin Vaccine Immunol 2007, 14:1356–1361.
35. Naessens J, Teale AJ, Sileghem M: Iden i ica ion o mechanisms o na u al
esis ance o A ican ypanosomiasis in ca le. Ve Immunol
Immunopa hol 2002, 87:187–194.
36. Mu ay M, Mo ison WI, Whi elaw DD: Hos suscep ibili y o A ican
ypanosomiasis: ypano ole ance. Ad Pa asi ol 1982, 21:1–68.
37. Ande son CA, Pe e sson FH, Cla ke GM, Ca don LR, Mo is AP, Zonde an
KT: Da a quali y con ol in gene ic case-con ol associa ion s udies.
Na P o ocols 2010, 5:1564–1573.
38. Aulchenko YS, Ripke S, Isaacs A, an Duijn CM: GenABEL: an R lib a y o
genome-wide associa ion analysis. Bioin o ma ics 2007, 23:1294–1296.
doi:10.1186/1471-2148-13-246
Ci e his a icle as: Mu ay e al.:Gene ic suscep ibili y o in ec ious
disease in Eas A ican Sho ho n Zebu: a genome-wide analysis o he
e ec o he e ozygosi y and exo ic in og ession. BMC E olu iona y
Biology 2013 13:246.
Mu ay e al. BMC E olu iona y Biology 2013, 13:246 Page 8 o 8
h p://www.biomedcen al.com/1471-2148/13/246