RESEARCH ARTICLE Open Access
Sex-speci ic popula ion s uc u e, na u al
selec ion, and linkage disequilib ium in a wild
bi d popula ion as e ealed by genome-wide
mic osa elli e analyses
Meng-Hua Li
*
, Juha Me ilä
Abs ac
Backg ound: Sexual dimo phism in ecologically impo an ai s is widesp ead, ye he di e ences in he genomic
a chi ec u e be ween he wo sexes a e la gely unexplo ed. We employed a genome-wide mul ilocus app oach o
examine he sexual di e ences in popula ion subdi ision, na u al selec ion and linkage disequilib ium (LD) in a wild
Sibe ian jay (Pe iso eus in aus us) popula ion, using geno ypes a a o al o 107 au osomal and Z-ch omosomal
mic osa elli es.
Resul s: Mean obse ed he e ozygosi y was signi ican ly highe in emales (H
O
= 0.567) han in males (H
O
= 0.532),
and au osomal ma ke s (H
O
= 0.561) we e mo e a iable han Z-ch omosomal ma ke s (H
O
= 0.512). Gene ic
di e en ia ion (F
ST
= 0.002, P< 0.05) be ween he wo sexes was low bu signi ican and males we e on a e age
signi ican ly mo e gene ically ela ed o each o he han emales. Genomescan analyses e ealed ha 3 ou o 101
(3%) au osomal loci we e unde di ec ional selec ion, while 4 ou o 6 (67%) Z-ch omosomal ma ke s we e
indica ed o be unde balancing selec ion. This sugges s a signi ican ly g ea e bu con as ing selec ion o ce on
he Z-ch omosome in compa ison o au osomes, which is consis en wi h an o e all signi ican ly (P< 0.05) lowe
F
ST
alue o Z-ch omosomal (-0.014, 95% CI: -0.025 - -0.011) han o he au osomal loci (0.003, 95% CI: 0.001 -
0.004). Analysis o syn enic ma ke pai s e ealed high le els o LD in bo h sexes bu signi ican ly (P< 0.05) lowe
le els o LD in he emales bo h on au osomes and Z-ch omosome, p obably due o he highe a e o dispe sal
and he highe ecombina ion a es on au osomes, as well as he pseudoau osomal ma ke s. In bo h sexes LD
decayed apidly wi h gene ic dis ance in a simila ashion on au osomes, while a mo e apid decay o LD in
Z-ch omosome was de ec ed in emales han in males.
Conclusion: We conclude ha he e a e many clea di e ences in genomic a chi ec u e be ween he sexes
s udied he e which can be a leas pa ly unde s ood in he ligh o highe dispe sal a e o emales as compa ed
o males and he unusual s uc u e o he Z-ch omosome o he species.
Backg ound
La ge di e ences in he o ces o e olu ion - mu a ion,
ecombina ion, selec ion, gene low, and gene ic d i -
a e known o occu be ween males and emales (e.g. [1]).
The e o e, unde s anding he ela i e impo ance o
e olu iona y o ces ha shape pa e ns o sex-speci ic
genomic dimo phism is essen ial o ou unde s anding o
he gene ic basis o sexual dimo phism and sex-speci ic
gene exp ession (e.g. [2,3]). Recen ly, g owing e o s
ha e been in es ed on elucida ing he ine scale gene ic
a chi ec u e o sexual dimo phism in complex pheno-
ypes (e.g. [3,4]). Howe e , li le is known abou he
gene ic a chi ec u e unde lying sex-biased e olu iona y
p ocesses a he genomic le el (bu see [1,5]), especially
in he wild.
In mos bi ds sexual di e ences in beha iou wi h
espec o ma ing and dispe sal p ac ices a e a
ule a he han excep ion (see he e iews in [6-8]).
* Co espondence: [email p o ec ed]
Ecological Gene ics Resea ch Uni , Depa men o Biosciences, PO Box 65, FI-
00014 Uni e si y o Helsinki, Finland
Li and Me ilä BMC E olu iona y Biology 2010, 10:66
h p://www.biomedcen al.com/1471-2148/10/66
© 2010 Li and Me ilä; 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.
Likewise, sex-speci ic selec ion o example in he o m
o compe i i e exclusion o indi iduals o one sex du ing
ec ui men o es ablishmen phases is a commonplace
occu ence in bi ds (e.g. [9]), as is also sexual dimo ph-
isminsize(seee.g.[10])andplumage ai s[11].Asa
consequence, also sexual dimo phism in genomic a chi-
ec u e - including he ex en o linkage disequilib ium
(LD) - is expec ed o occu bo h o selec i e and demo-
g aphic easons. While de ailed unde s anding o he
ex en and pa e ns o LD in he wo sexes is in e es ing
in i sel , his knowledge will also acili a e he choice o
app op ia e me hodology o sex-speci ic QTL mapping
in he wild. In addi ion, inb eeding a oidance, sexual
con lic , and sex-biased in es men a he popula ion
le el can cause non- andom ma ing esul ing in di e -
ences in geno ypic dis ibu ion be ween he sexes
(e.g. [12,13]). To his end, a mul ilocus app oach as used
e.g. in D osophila, maize and humans (e.g. [14,15]) is a
powe ul way o disen angle he e ec s o sex- ela ed
e olu iona y o ces on genomic a ia ion.
The Sibe ian jay (Pe iso eus in aus us)isapasse ine
bi d which has been subjec o conside able ecological
and e olu iona y esea ch du ing he pas decades. I
exhibi s sexual dimo phism in mo phological measu e-
men s [16], nepo is ic beha iou [17], gene ic s uc u ing
[18], li e ime ep oduc i e success and dispe sal pa e ns
([19], Gienapp and Me ilä, unpublished esul s). Mo e-
o e , sex-speci ic genome-wide he e ogenei y in ecombi-
na ion a es and linkage pa e ns o bo h au osomes and
he Z-ch omosome ha e been ecen ly de ec ed [20].
He e, we adop ed a genome-wide mul ilocus app oach
o examine he sex-speci ic genomic di e ences in
Sibe ian jays aking ad an age o a o al o 107 mic osa-
elli e ma ke s geno yped in 172 males and 177 emales.
In pa icula , we analysed and compa ed Z ch omoso-
mal and au osomal mic osa elli es o iden i y possible
sex-speci ic p ocesses ha may ha e shaped genomic
pa e ns o a iabili y. We also in es iga ed he possible
sexual dimo phism in magni ude and ex en o LD. In
pa icula , using sex-speci ic linkage maps, we in es i-
ga ed he decay o LD wi h gene ic dis ance in he
sexes. The main objec i es we e o: 1) de e mine he
pa e n o a iabili y and gene ic di e en ia ion be ween
sexes in bo h au osomal and Z-ch omosomal loci; 2)
in es iga e he po en ial signa u es o na u al selec ion
ela ed o sex on au osomal and Z-ch omosomal ma -
ke s; and 3) in es iga e he di e ences in ex en and
pa e n o LD be ween he sexes.
Resul s
Gene ic di e si y, di e en ia ion and ela edness
Locus speci ic he e ozygosi ies and deg ee o gene ic
di e en ia ion a e shown in Figu e 1 and Table 1. A sig-
ni ican ly lowe a e age obse ed he e ozygosi y was
ound in males (H
O
= 0.532 ± 0.004) han in emales
(0.567 ± 0.004; pai ed - es :
212
=1.16,P<0.01;
Table 1). The mean unbiased es ima ed he e ozygosi y
o au osomal ma ke s (H
E
=0.566±0.003)wassigni i-
can ly (unpai ed - es :
105
= 3.36, P< 0.05) highe han
ha o Z-ch omosomal ma ke s (H
E
= 0.512 ± 0.001;
Figu e 1 Dis ibu ion o obse ed F
ST
alues o each loci as a unc ion o hei Nei’s unbiased he e ozygosi y (H
E
).Thesimula ed
median line and 95% con idence limi s a e ep esen ed by dashed and solid lines, espec i ely, o he Fdis 2 me hod. G ay shading indica es
a ea on he g aph wi hin he 95% con idence limi s.
Li and Me ilä BMC E olu iona y Biology 2010, 10:66
h p://www.biomedcen al.com/1471-2148/10/66
Page 2 o 11
Figu e 1). F
ST
alues o indi idual loci among sexes a -
ied om -0.023 (SD = 0.011) a SJ046 on Z-ch omosome
o 0.069 (SD = 0.037) a SJ036 on au osome. O e all
gene ic di e en ia ion ac oss all he loci was F
ST
=0.002
(95% CI 0.001 - 0.003, P< 0.05) be ween males
and emales, wi h F
ST
= 0.003 (95% CI 0.001 - 0.004,
P< 0.05) o au sosomal ma ke s and F
ST
= -0.014
(95% CI -0.025 - -0.011, P< 0.05) o Z-ch omosomal
ma ke s. Ac oss all he loci, he mean F
IS
was 0.035
(± 0.014) o e all he samples and de ia ed signi ican ly
(P< 0.05) om ze o, while being 0.053 (± 0.027,
P< 0.05) o males and -0.018 (± 0.013, P<0.05) o
emales. The a e age pai wise ela edness was 0.036 (SE:
0.006, 95% CI: 0.032 - 0.039) o males, 0.019 (SE: 0.003,
95% CI: 0.017 - 0.022) o emales and 0.041 (SE: 0.008,
95% CI: 0.038 - 0.045) o males and emales combined
(Table 1).
E idence o selec ion be ween males and emales
The FDIST2 analyses iden i ied eigh loci as ou lie s show-
ing oo p in s o na u al selec ion be ween males and
emales a he 0.5% signi icance le el (Figu e 1; Table 2).
O he eigh signi ican loci, ou we e au osomal (SJ014,
SJ022, SJ036 and CKL5) wi h highe han expec ed F
ST
a-
lues indica ing di ec ional selec ion, while ou Z-ch omo-
somal loci (SJ009, SJ046, SJ048 and SJ108) appea ing in
he lowe ail o he F
ST
dis ibu ion showed signa u es
ypical o balancing selec ion (Figu e 1).
Bayesian F
ST
- es based on a hie a chical eg ession
model indica ed h ee au osomal loci (SJ022, SJ036 and
CKL5) as di ec ionally selec ed and ou Z-ch omosomal
loci (SJ009, SJ046, SJ048 and SJ108) o be unde balan-
cing selec ion (Table 2). Thus, se en o he loci (SJ009,
SJ022, SJ036, CKL5, SJ046, SJ048 and SJ108) we e picked
up by bo h me hods gi ing suppo o hei s a us as ou -
lie s due o selec ion. In addi ion, locus SJ014 can be
seen as candida e a ec ed by di ec ional selec ion, bu
only one o he wo es s suppo ed his s a is ically.
Ou o he 107 mic osa elli e analysed, he ou Z-
ch omosomal loci (SJ046, SJ048, SJ009, and SJ108)
unde balancing selec ion showed he lowes dis ibu ion
o F
ST
alues in bo h es s being as nega i e ou lie s in
he BAYESFST esul s; whe eas he o he ou au oso-
mal loci unde di ec ing selec ion de ec ed by one o
bo h o he es s show he highes di e en ia ion alues
in bo h analyses (Figu es 1 and 2).
Table 1 Summa y s a is ics o obse ed he e ozygosi y (H
O
), Nei’s unbiased he e ozygosi y (H
E
), F-s a is ics (F
IS
and
F
ST
), and ela edness (R) a e aged o males, emales and he o e all sample a au osomal and Z-ch omosomal
mic osa elli es in he Sibe ian jays.
Sex H
O
(± SD)
H
E
(± SD)
F
IS
(± SD)
R
(± SE, 95% C.I.)
F
ST
(95% CI)
Au osomal Z-ch omosomal
Male 0.532
(0.004)
0.559
(0.019)
0.053*
(0.027)
0.036
(0.006, 0.032 - 0.039)
--
Female 0.567
(0.004)
0.562
(0.019)
-0.018*
(0.013)
0.019
(0.003, 0.017 - 0.022)
--
O e all 0.550
(0.003)
0.561
(0.019)
0.035*
(0.014)
0.041
(0.008, 0.038 - 0.045)
0.003*
(0.001 - 0.004)
-0.014*
(-0.025 - -0.011)
0.002 (0.001 - 0.003)*
*P< 0.05
Table 2 Summa y s a is ics o ou lie s de ec ed be ween male s. emale Sibe ian jays, using FDIST2 and BAYESFST
me hods.
Locus Loca ion
a
FDIST2 BAYESFST
F
ST
H
E
PSelec ion F
ST
PSelec ion
SJ014 A-LG4 0.036 0.865 0.0032* di ec ional 0.019 0.08 -
SJ022* A-LG1 0.051 0.766 0.0022* di ec ional 0.024 0.03* di ec ional
SJ036* A-LG2 0.069 0.404 0.0034* di ec ional 0.038 0.01* di ec ional
CKL5* A-LG2 0.043 0.819 0.0015* di ec ional 0.023 0.03* di ec ional
SJ009* Z -0.023 0.467 0.0027* balancing 0.006 0.01* balancing
SJ046* Z -0.023 0.544 0.0041* balancing 0.005 0.03* balancing
SJ048* Z -0.021 0.677 0.0033* balancing 0.005 0.04* balancing
SJ108* Z -0.022 0.651 0.0029* balancing 0.005 0.04* balancing
a
A, au osome; Z, Z-ch omosome; LG, linkage g oup
* Signi ican alue o 5% le el o he BAYESFST me hod and 0.5% le el o he FDIST2 me hod, and locus labels ma ked wi h as e isks i bo h me hods indica e a
signi ican esul
Li and Me ilä BMC E olu iona y Biology 2010, 10:66
h p://www.biomedcen al.com/1471-2148/10/66
Page 3 o 11
Pa e ns o LD
Figu e 3 and Table 3 show he dis ibu ion o D’ alues
o he au osomal syn enic ma ke pai s as a unc ion o
gene ic dis ance (in cM) in bo h emales and males. In
bo h cases, D’decayed wi h dis ance a ying be ween
0.02 - 0.947 in males and be ween 0.018 - 0.928 in
emales (see Figu e 3). In o al, signi ican ly lowe mean
D’- alue o he au osomal ma ke s was e ealed in
emales han in males (Males: 0.378 ± 0.145; Females:
0.299 ± 0.126; unpai ed - es :
1874
=9.13,P< 0.001).
Mean D’- alues we e 0.574 (± 0.162) in males and 0.532
(± 0.154) in emales o ma ke s sepa a ed by 10 cM o
less, while a he dis ance in e als o > 10 cM be ween
ma ke pai s a e all <0.5 in bo h sexes (Table 3). All he
mean D’- alues a e lowe han ha in he sample o
pooled sexes (see [21]). Thus, he ‘hal -leng h’o LD
(measu ed as he dis ance a which mean D’ alls o 0.5)
is ca. 10 cM, hal o ha measu ed in he pooled sample
o males and emales (ca. 20 cM; see [21]). Fo he au o-
somal ma ke pai s, he mean D’- alue dec eased wi h
inc easing gene ic dis ance and was sys ema ically lowe
in he emales han in he males a compa able dis ances
(Figu e 3). Bo h o he nega i e co ela ions be ween LD
and gene ic dis ance we e s a is ically signi ican (F- es ,
P< 0.01). The coe icien o de e mina ion (R
2
)was
0.219 o males and somewha highe (R
2
= 0.367) o
he emales when exponen ial end lines we e i ed o
he da a. As u he indica ed in Figu e 3, D’decays
ela i ely apidly in he i s ens o cen imo gans, while
o ma ke pai s spaced by > ca.100cMslowe
dec eases a e de ec ed. Howe e , signi ican (P< 0.05)
and s ong associa ions (D’> 0.5) we e obse ed also o
a ew compa isons among loci sepa a ed by > 100 cM in
bo h sexes (Figu e 3).
We also compa ed he le el o pai wise LD measu ed
o Z-linked ma ke s as a unc ion o gene ic dis ance in
males and emales. The es ima es o D’- alues a e signi -
ican ly (unpai ed S uden ’s - es :
28
=1.25,P< 0.001)
lowe in emales (mean D’= 0.469) han in males (mean
D’= 0.517; Table 3). In bo h sexes, he D’decayed as a
unc ion o gene ic dis ance. Howe e , a he gene ic
dis ances ≤10.6 cM ( he leng h o emale-speci ic
Figu e 2 Resul s o he Bayesian F
ST
- es . The solid line indica es
he c i ical cu -o poin o he P- alue a he 5% le el.
Figu e 3 Linkage disequilib ium as measu ed by D’as a unc ion o gene ic dis ances (cM) be ween au osomal syn enic ma ke s. The
da k and g ey lines gi e exponen ial end line i s o males and emales, espec i ely.
Li and Me ilä BMC E olu iona y Biology 2010, 10:66
h p://www.biomedcen al.com/1471-2148/10/66
Page 4 o 11
Z-ch omosome linkage g oup) a mo e apid decay o D’
was de ec ed in emales han in males (Figu e 4). When
loga i hmic end lines we e i ed o he da a o emales
and males, he coe icien s o de e mina ion (R
2
)we e
0.319 and 0.025, espec i ely. We also de ec ed signi i-
can (P<0.01)LDwi hD’> 0.5 a all locus pai s
among h ee adjacen loci (SJ046, SJ048, and SJ009), i.e.
LD block, in bo h males and emales (Figu e 5).
Game ic LD was also de e mined o he 4718 non-
syn enic ma ke pai s in bo h emales and males sepa-
a ely. Fo he nonsyn enic ma ke pai s D’ a ies
be ween 0.005 - 0.431 in males and 0.005 - 0.373 in
emales (da a no shown). The mean D’- alues we e
0.153 (± 0.063) in males and 0.104 (± 0.034) in emales
(Table 3), signi ican ly lowe han ha o he syn enic
ma ke pai s (Males: 0.378 ± 0.145,
5669
=5.11,P<
0.0001; Females: 0.299 ± 0.126,
5669
= 6.24, P< 0.0001).
Discussion
Le els and pa e ns o gene ic di e si y, sexual
di e en ia ion and ela edness
Bo h sexes o Sibe ian jays exhibi ed subs an ial le els o
gene ic a ia ion. Mo e in e es ingly, males we e
signi ican ly less he e ozygous han emales. This gen-
ome-wide pa e n appea s compa ible wi h he emale-
biased dispe sal in his species (see below). Also di e -
en ial na u al selec ion on he wo sexes could con i-
bu e o his obse a ion. Fo ins ance, mo e in ensi e
wi hin-sex compe i ion could educe gene ic a iabili y
in males mo e han in emales (see [8]). Ou indings o
lowe le els o di e si y in males and on he Z-ch omo-
some ela i e o he au osomes a e conco dan wi h
esul s o p e ious s udies based on nucleo ide
sequences o mic osa elli e a ia ion in a wide ange o
species including he colla ed lyca che [22], chicken
[23], human [24,25], house mice [26] and D osophila
[27,28]. Indeed, many e olu iona y models, such as
ecen popula ion bo leneck and ecu en selec i e
sweeps a e expec ed o educe he le els o Z-linked (o
X-linked) gene ic di e si y ela i e o au osomal a ia-
bili y (see [1]).
We ound e idence o a signi ican sex-speci ic pa -
e n o popula ion gene ic s uc u e a au osomal mic o-
sa elli es. The obse ed gene ic di e en ia ion be ween
males and emales was unexpec ed gi en ha males and
emales a e om he same coho s and sha e
Table 3 LD summa y s a is ics o mean alues o he Lewon in’s no malised D’(Lewon in 1964) in all linkage g oups
(au osome and Z-ch omosome) sepa a ely o syn enic and nonsyn enic ma ke s in males and emales.
Syn enic (cM)
Sex Au osome Z-Ch . Nonsyn enic
0-10 10-20 20-40 40-60 60-100 100-171.1 171.1-229.2 0-48.9
Males 0.574
(0.162)*
0.495
(0.151)
0.389
(0.142)
0.354
(0.131)
0.299
(0.124)
0.234
(0.110)
- 0.517
(0.176)
0.153
(0.063)
Females 0.532
(0.154)
0.443
(0.151)
0.345
(0.129)
0.261
(0.122)
0.232
(0.117)
0.196
(0.104)
0.061
(0.034)
0.469
(0.169)
0.104
(0.034)
* Value in pa en hesis a e s anda d e o s o D’.
Figu e 4 Linkage disequilib ium as measu ed by D’as a unc ion o gene ic dis ances (cM) be ween Z-ch omosomal syn enic ma ke s.
The da k and g ey lines gi e exponen ial end line i s in males and emales, espec i ely.
Li and Me ilä BMC E olu iona y Biology 2010, 10:66
h p://www.biomedcen al.com/1471-2148/10/66
Page 5 o 11

genealogies wi hin he pedig ees: hei au osomal ma -
ke s a e mixed a each gene a ion. The e o e, long- e m
e olu iona y o ces such as gene ic d i o mu a ion
could no gene a e au osomal gene ic di e ence be ween
sexes [29]. Mo eo e , selec i e e ec s we e a ely (4/101,
3.97%) de ec ed in he au osomal mic osa elli es. Thus,
he sho - e m (one gene a ion) e ec o mig a ion, o
example he di e en ial pa e ns o dispe sal be ween
sexes, appea o p o ide he mos likely explana ion.
Comple e da a on li e- ime ep oduc i e success and
dispe sal his o y o a la ge numbe o males and
emales collec ed o o e 30 yea s ha e e ealed ha
na al dispe sal dis ance was ela ed o sex and dispe sal-
iming: emales and ea ly dispe se s a elled on a e age
a he han males and delayed dispe se s (Gienapp and
Me ilä, unpublished esul s). Ou gene ic da a also co -
obo a e hese indings, wi h a endency o nega i e F
IS
and lowe gene ic ela edness among emales. The sig-
ni ican posi i e F
IS
obse ed in males is consis en wi h
his a gumen as well. Al e na i e explana ions o he
sex-speci ic gene ic di e en ia ion would be di e en ial
iabili y, li espan o ha ching a e be ween sexes, which
can lead o he allele equency di e ence be ween
emales and males de ec ed by F
ST
(see [8,29]). Un o u-
na ely, gi en he a ailable da a in he Sibe ian jays,
asce aining he assump ions and quan i ying he ele-
an pa ame e s unde lying hese non-mu ually exclu-
si e hypo heses is di icul .
E idence o selec i e sweeps on he Z-ch omosome
We iden i ied e idence o selec ion on he Z-ch omoso-
mal (nega i e ou lie s) and au osomal (posi i e ou lie s)
ma ke s wi h wo di e en me hods. In gene al, he
ela i e equency o ou lie s was highe in he Z-ch o-
mosome (67%, i.e. 4 ou o 6) han in he au osomes
(3%, 3/101 om he FDIST2 me hod and 4%, 4/101
om he BAYESFST me hod). Se e al s udies o a ian
genomes ha e indeed e ealed di e ences be ween he
Z-ch omosome and he au osomes in he a es o gene
di e gence (e.g. [30]), pa e ns o gene exp ession (e.g.
[31]) and a es o gene mo emen be ween ch omo-
somes (e.g. [20,22,32]), which may explain he di e en
pa e ns o selec ion obse ed he e. This obse a ion is
also compa ible wi h heo e ical expec a ion (see
[33,34]) and genomic e idence (e.g. [1]) o di e en ial
selec ion on au osomal and Z-ch omosomal (o X-ch o-
mosomal) loci: he Z-ch omosome is p edic ed o
expe ience selec i e sweeps mo e o en han au osomal
ch omosomes (see [1]). Gi en he high le els and ex en
o LD obse ed in he popula ion in gene al [21], he
Figu e 5 De ailed iew o he ex en and signi icance o LD in males and emales using Haplo iew 4.0 p og am. The LD blocks de ined
a e as desc ibed in he Ma e ials and Me hods sec ion. Numbe s in he blocks indica e he pe cen age o he LD me ic D’ alues. Shadings
indica e Fishe ’s exac es signi icance le els: whi e, P> 0.05; ligh shading, P< 0.05; da k shading, P< 0.01. The leng hs o male- and emale-
speci ic linkage g oups in he igu es a e no in p opo ion.
Li and Me ilä BMC E olu iona y Biology 2010, 10:66
h p://www.biomedcen al.com/1471-2148/10/66
Page 6 o 11
obse ed pa e n o a ew posi i e au osomal ou lie s ou
o mo e han 100 au osomal loci seems mo e consis en
wi h locus-speci ic e ec s due o local adap a ion, a he
han selec i e sweeps. I is also possible ha alse posi-
i es ( ype I e o ) esul ing om some s a is ical s o-
chas ici y o bias a e esponsible o he obse ed
au osomal ou lie s (see e.g. [35-38]). Howe e , he
g ea e p opo ion o Z-ch omosomal nega i e ou lie s
may be a ec ed by ch omosome-speci ic e ec s due o
sex- ela ed e olu iona y o ces, which would ha e
a ec ed he genome in a mo e ex ensi e manne . O he
loci de ec ed o be subjec o selec ion, h ee ha e been
mapped o he chicken ch omosomes (SJ009 on GgaZ,
SJ022 on Gga1 and SJ036 on Gga 2; see [20]). The e o e,
he loci p o ide good candida es o u he in es iga-
ions aiming o iden i y genes unde sex-speci ic
selec ion.
The ea eanumbe o sex-biased e olu iona y o ces
(e.g. sex-biased demog aphic p ocesses and sexual di e -
ence in ep oduc i e success, li espan and mo ali y
a es) ac ing wi hin wild bi ds ha a e known o ha e
di e en ial e ec s on loci wi h di e en modes o inhe i-
ance (see [34]). In bi ds, emales a e he he e ogame ic
sex (ZW) whe eas males a e homogame ic (ZZ). Because
dele e ious mu a ions will be exposed o selec ion on
he Z-ch omosome when being hemizygous in emales,
selec ion, in pa icula selec i e sweeps, may he e o e
a ou educed a ia ion on he Z-ch omosome leading
o a balanced polymo phism [30]. In he chicken, se e al
lines o e idence sugges ha selec i e sweeps could be
a po en o ce in shaping Z-ch omosome a iabili y (e.g.
[23]). Ano he po en ial explana ion o he balancing
selec ion be ween sexes on he Z-ch omosomal loci is a
o m o equency-dependence selec ion. The selec ion
on he sex a io main ains males and emales in he
Sibe ian jay popula ion wi h a Z/W sex ch omosome
sys em,whichbeha eslikeasinglegene,asla gepa s
o he ch omosomes, including he egions con aining
he sex-de e mining egion, do no unde go gene ic
c ossing-o e [39]. Since long- e m balancing selec ion
is unusual e en in he classic case o D osophila poly-
mo phism [40] and he e olu iona y o ces a e belie ed
o be sho - e m he e, he balancing selec ion on Z-
ch omosomal loci de ec ed in his s udy should ep e-
sen ecen e en s. Balancing selec ion may hus o en
occu , al hough he di icul ies exis in de ec ing he
‘ eal’signa u e due o a e ac s in he s a is ical me hods
(see e.g. [36]), and could be he basis o much quan i a-
i e a iabili y, including a ia ion in i ness [39].
De ia ions om neu ali y a he ou Z-ch omosomal
loci wi h bo h he neu ali y es s migh ha e been in lu-
enced by gene ic hi chhiking due o selec ion ac ing a
dis inc bu closely linked loci. Signi ican e idence has
been ound o he balancing selec ion on majo
his ocompa ibili y complex (MHC) genes in e eb a es
including bi ds (see [41,42]). Howe e , he linkage o
hese candida e loci o any unc ional loci is cu en ly
unknown. The e o e, al hough signa u es o selec ion
can be iden i ied wi h he aid o mic osa elli es, mic osa-
elli es in gene al do no lend hemsel es well o s udy-
ing he e ec s o na u al selec ion. The e a e now many
mo e es s o selec ion o coding and noncoding DNA
sequences which need o be i s iden i ied (see [39,43]).
Thus, analyses o DNA sequences ha e he p omise o
ad ance unde s anding o he di e en o ms o balan-
cing selec ion [37].
Mapping loca ions o he ou Z-ch omosomal loci
(SJ009, SJ046, SJ048 and SJ108) indica ed o be unde
balancing selec ion a e si ua ed in he middle o he
ch omosome, a he han he elome ic egions [20].
This aligns wi h ea lie indings: selec ion is ypically
de ec ed o ac on ma ke s in he cen e o Z-ch omo-
somes in bi ds, as opposed o he elome ic egions (see
[44]). This could be due o he gene ally highe ecom-
bina ion (and mu a ion) a e be ween loci in he cen o-
me ic hanin elome ic egionsalong heZ-
ch omosome (see [20,22,45]). Al e na i ely, i could be
also due o he ac ha he pseudoau osomal egion
(PAR), which is hypo hesized o sha e he p ope ies o
au osomes, is si ua ed a one ip o he Z-ch omosome
in spi e o i s a ying sizes among species. Thus, he
wo candida e loci (SJ009 and SJ048) in he PAR we e
expec ed o be mo e simila in di e si y and dynamics
o au osomal loci han o Z-speci ic loci. We specula e
ha he selec ion pa e n o candida e loci in he PAR
de ec ed he e may be in luenced by linkage o Z-speci ic
loci (e.g. he LD blocks in bo h he sexes, see Figu e 5),
o o some eason expe iences selec ion in insic o
genes su ounding hem in he egion. Simila e idence
o a depa u e om a s anda d neu al model in PAR
loci was also de ec ed on he Z-ch omosome o he
Emu D omainus no aehollandiae [46].
Sex-speci ic pa e ns o LD and p ospec s o LD mapping
wi hin sex
Conside able amoun s o LD be ween loci we e de ec ed
wi hin each sex. Al hough many o ces can lead o he
high le els and ex en o LD, his inding as well as he
long-dis ance LD can be po en ially a ec ed by he pedi-
g ee samples whe e many a e ull-sibs o hal -sibs e en
wi hin he same sex. Howe e , since high le els o LD
we e s ill obse ed in he ounde s o he pedig ee sam-
ples [21], indica ing he in e enceislikely oapply o
he whole popula ion oo. A small e ec i e popula ion
size and closely ela ed indi iduals may lie behind he
high LD obse ed in his s udy, which, ne e heless, do
no necessa ily a ec much on he compa ison o LD
be ween he sexes. We obse ed signi ican ly lowe LD
Li and Me ilä BMC E olu iona y Biology 2010, 10:66
h p://www.biomedcen al.com/1471-2148/10/66
Page 7 o 11
in emales han males in e ms o mean D’ alues o all
he syne nic ma ke pai s and o ma ke pai s sepa a ed
< 10 cM. Female-biased dispe sal, which may ha e
esul ed in he lowe a e age pai wise ela edness and
he lowe inb eeding a e in emales (see abo e), can
p o ide an explana ion o his. Mo eo e , he signi i-
can ly highe gene ic s uc u ing o male popula ion can
also explain he highe LD in males.
We obse ed a highe le el o LD o ma ke s on he
Z-ch omosome han au osomes in bo h males and
emales, indica ing highly a iable pa e n and ex en o
LD ac oss he genome. Simila inding o highe LD o
ma ke s on he X-ch omosome has been ypically ound
in he X-Y sex de e mina ion sys em e.g. in humans (e.
g. [47]) and ca le (e.g. [48]). In addi ion o dis inc
selec i e o mu a ional o ces, he highe le el o LD on
he Z-ch omosome is p ima ily hough o esul om
highe gene ic d i because o i s smalle e ec i e size
(3/4 o au osomes) as compa ed o au osomes. The ac
ha he inc eased le els o LD we e also obse ed o
ma ke s loca ed > 10 cM apa sugges s ha he ac o s
a e s ill ope a ing o ha e been ope a ing un il ecen ly
(see [48]). Since ecombina ion occu ing be ween wo
si es will usually educe he LD be ween hem, he
ecombina ion a e is likely o be nega i ely co ela ed
wi h he LD be ween he pai o si es. Con a y o his
expec a ion, in gene al lowe ecombina ion a e bu
lowe LD was obse ed in emales han in males in his
s udy. The signi ican ly lowe le els o LD on he Z-
ch omosome in emales is mainly asc ibed o he h ee
sex-speci ic ma ke s, which show ela i ely low LD bu
no ecombina ion de ec ed be ween he ma ke pai s
(see Figu e 5).
The decay o LD in bo h sexes o he Sibe ian jay
popula ion is e y use ul o high- esolu ion mapping in
sex-speci ic associa ion s udies, p o ided ha app op i-
a e candida e genes a e chosen. Howe e , as obse ed in
he samples including bo h sexes [21], conside able,
al hough ela i ely lowe le els o LD we e s ill obse ed
be ween dis an ly linked ma ke s as well as be ween
many nonsyn enic ma ke s wi hin he sexes. The com-
mon occu ence o nonsyn enic LD e okes se ious con-
ce ns abou he gene a ion o ype I e o s when using
LD mapping as he only means o loca e genes unde ly-
ing sexually dimo phic ai s in he popula ion [49]. Fo
his pu pose, a new s a egy o join linkage and LD
mapping in na u al popula ions [50] by use o an e en
highe densi y ma ke sc eening would be needed o
a oid alse posi i e esul s when mapping sex-speci ic o
sex-biased genes in he popula ion. P e ious s udies
epo ed ha he Z-ch omosome ha bou s many QTLs
and genes a ec ing ai s o ecological and e olu iona y
impo ance in bi ds. Coupled wi h he LD blocks ound,
he highe le el and decay o LD obse ed on he Z-
ch omosome in each sex sugges s ha LD mapping may
be possible on he Z-ch omosome in he bi d popula-
ion using low-densi y ma ke maps. In pa icula , he
sex-speci ic LD mapping has he po en ial o be mo e
e ec i e in emales due o he much mo e apid decay
o LD wi h dis ance.
Conclusions
To ou knowledge his is he i s s udy a emp ing o
disen angle he sexual di e ence in genome-wide gene ic
a chi ec u e in a wild bi d popula ion. We obse ed a di -
e ence in gene ic a iabili y, na u al selec ion and LD
be ween males and emales in a Sibe ian jay popula ion.
Se e al di e en e olu iona y o ces and demog aphic
p ocesses including di e en ial dispe sal a e, ecombina-
ion he e ogenei y and Z-ch omosomal genomic s uc-
u e may unde lie he obse ed gene ic di e ences
among sexes. Ou esul s sugges ha he sex-speci ic LD
mapping could be p omising in his popula ion and i
would be ad an ageous o include Z-ch omosome ma -
ke s. Fu u e heo e ical wo k (e.g. examining he join
e ec s o mul iple e olu iona y p ocesses on genomic di -
e ence be ween sexes) and expe imen al esea ch [e.g.
aimed a sex-speci ic mapping Z-ch omosomal QTLs o
genes and in es iga ing he e olu iona y pa e n o di e -
ences in gene exp ession (see he e iew by [31]) be ween
he sexes in a wild bi d popula ion] would inc ease ou
unde s anding o he gene ic basis o sexually dimo phic
ai s.
Me hods
S udy species, s udy popula ion and pedig ee
The Sibe ian jay (Pe iso eus in aus us) is a medium sized
(body mass ca. 85-90 g) and ela i ely long-li ed (a e -
age gene a ion ime ca. 4 yea s) oscine passe ine bi d
om he Co idae amily. I has a s able socially mono-
gamous b eeding sys em in which li e-long pai -bonds
a e o med in pe manen e i o ies es ablished in he
coni e ous o es o he no he n Eu asian aiga (see e.g.
[16]). The species has been desc ibed in mo e de ail in
[19,20,51].
A long- e m ield s udy o he Sibe ian jay (Pe iso eus
in aus us)popula ioninSuupohja(ca. 66°18’N, 29°29’E)
in Wes e n Finland has been conduc ed since 1974. The
s udy popula ion, ield me hods and he da a s uc u e
a e desc ibed in de ail elsewhe e [20,52]. The pedig ee
used in his s udy was buil based on ield obse a ions
[52] and molecula ools as desc ibed in [20]. The pedi-
g ee consis s o 349 animals (172 males and 177
emales) sampled in 1975-2006.
Mic osa elli e ma ke s and geno ypes
The pedig ee has been geno yped in a o al o 107 mic o-
sa elli es comp ising 101 au osomal and 6 Z-ch omosomal
Li and Me ilä BMC E olu iona y Biology 2010, 10:66
h p://www.biomedcen al.com/1471-2148/10/66
Page 8 o 11
ma ke s sepa a ed by a dis ance o 10.6 o 229.2 cM in
emales and 7.7 o 168.3 cM in males [20]. A gene ic map
has been cons uc ed by ollowing he co-seg ega ion
wi hin he pedig ee and de e mining ma ke o de and
dis ances wi h he so wa e CRIMAP [53], desc ibed in
de ail in [20]. The mic osa elli es we e dis ibu ed in 9
au osomal linkage g oups (LG1 - LG9) and one Z-ch o-
mosome linkage g oup (LGZ) co e ing in o al 999 cM in
emales and 822.9 cM in males. A o al o 938 au osomal
and 15 Z-ch omosomal syn enic ma ke pai s we e
ob ained. The ma ke o de and all pai wise in e -ma ke
gene icdis ancesa ebasedon hesex-speci iclinkage
maps (see [20]).
Mic osa elli e a ia ion, gene ic di e en ia ion and
ela edness analyses
Mic osa elli e (au osomal and Z-ch omosomal) a ia ion
was assessed by he numbe o alleles and he obse ed
he e ozygosi y in he males and emales using he Excel
Mic osa elli e Toolki e sion 3.1.1 [54]. Gene ic di e -
en ia ion be ween sexes was quan i ied by he θes ima-
o o F
ST
[55]. We also calcula ed he alue o ,which
co esponds o W igh ’s [56] wi hin-popula ion inb eed-
ing coe icien F
IS
, ac oss he loci o e he sexes as well
as o he o al sample. These calcula ions we e pe -
o med wi h FSTAT e sion 2.9.3.2 [57] and he signi i-
cance and he 95% con idence in e als o θand we e
de e mined by 10 000 pe mu a ions. In addi ion, an
index o ela edness (R) based on all he 107 selec ed
loci was compu ed wi hin sexes using he p og am
RELATEDNESS e sion 5.08 [58]. A e age ela edness
was calcula ed ac oss he pai wise alues o males,
emales and all indi iduals, espec i ely. The s anda d
de ia ion and 95% con idence in e als o ela edness
alues we e es ima ed by jackkni ing.
Linkage disequilib ium es s
We measu ed he s eng h o LD using he Lewon in’s
no malised D’[59] modi ied o mul iple alleles [60].
This s a is ic has been used ea lie o measu e he
ex en o LD be ween ma ke s spanning one single
ch omosome and/o he whole genome in humans (e.g.
[61]), li es ock (e.g. [49]) and wild e eb a es (e.g.
[21,32,62]). Apa om acili a ing he compa ison o
he esul s wi h hose o he s udies, D’is a con enien
measu e o LD as i allows use o highly polymo phic
ma ke s and is less sensi i e o a ia ion in ma ke allele
equencies han o he measu es o LD [58]. The calcu-
la ions o D’a e de ailed in [34].
MIDAS (Mul i-allelic In e allelic Disequilib ium Ana-
lysis So wa e) so wa e [63] p o ided es ima es o D’
ij
,
p
i
,andq
i
o allelic combina ions; D’ o each ma ke
pai we e hen calcula ed using he equa ions (1) in
[21]. The LD was compu ed be ween all syn enic
ma ke s among he linkage g oups in emales and
males, sepa a ely. The s a is ical signi icance (P- alue) o
he obse ed associa ion be ween locus pai s was ca ied
ou using a Mon e-Ca lo app oxima ion o Fishe ’s
exac es as implemen ed in he so wa e ARLEQUIN
[64]. Fo each locus pai wi hin he sex, he unbiased
es ima e o he P- alue is calcula ed as he sum o he
p obabili ies o all ables (wi h he same ma ginal alues
as he obse ed one) wi h a lowe o equal p obabili y
han he obse ed able. G aphic summa y o he ex en
and signi icance o LD de e mina ions was displayed by
using he p og am HaploView e sion 4.0 [65]. Plo s o
pai wise compa isons among associa ions ela i e o
gene ic dis ances we e gene a ed in au osomal and Z-
ch omosomal ma ke s using Mic oso Excel.
Tes s o de ec loci unde selec ion be ween sexes
We examined possible depa u es om he s anda d
neu al model o molecula e olu ion - po en ially
e ealing demog aphic e en s o he exis ence o selec-
i e e ec s a ce ain loci - be ween sexes by he Beau-
mon and Nichols’s modi ied equen is me hod [66], as
well as a mo e obus Bayesian es [67].
We used he equen is me hod p oposed by Beau-
mon and Nichols [66], u he de eloped by Beaumon
and Balding [67], and implemen ed in he p og am
FDIST2, a cu en dis ibu ed e sion o he o iginal
p og am FDIST as desc ibed in [67]. FDIST2 calcula es
θ, Wei & Cocke ham’s [55] es ima o o di e si y o
each locus in he sample. Coalescen simula ions a e
hen pe o med o gene a e da a se s wi h a dis ibu ion
o θcen ed on he empi ical es ima es. We hen de e -
mined he quan iles o he simula ed F
ST
wi hin which
he obse ed F
ST
’s ell, and signi icance le el (P- alues)
o each locus. Ini ially, we used an island model o
popula ion di e en ia ion and epea ed he p ocedu e
50,000 imes o gene a e 95% con idence in e als o
neu al di e en ia ion. Simula ion pa ame e s we e
unde an in ini e allele mu a ion model o 100 demes,
sample sizes 100, and a mean weigh ed F
ST
o 0.0014
calcula ed om he 107 mic osa elli e loci. This me hod
p o ides e idence o selec ion by looking o ou lie s
wi h highe /lowe obse ed F
ST
- alues, con olling o
he e ozygosi y [67]. This app oach is ai ly obus o
a ia ion in mu a ion a e be ween loci, o sample size,
and whe he popula ions a e equilib ium o non-equili-
b ium [67].
Beaumon and Balding’s [67] hie a chical-Bayesian
me hod was pe o med using he p og am BAYESFST
package [68], which gene a es 2000 Ma ko chain
Mon e Ca lo (MCMC) simula ed loci on he basis o he
dis ibu ion o F
ST
gi en he da a. The me hod com-
bines in o ma ion o e loci and popula ion in o de o
simul aneously es ima e F
ST
ac oss he locus/popula ion
Li and Me ilä BMC E olu iona y Biology 2010, 10:66
h p://www.biomedcen al.com/1471-2148/10/66
Page 9 o 11