Digging into the Genomic Past of Swiss Honey Bees by Whole-Genome Sequencing Museum Specimens

Digging in o he Genomic Pas o Swiss Honey Bees by
Whole-Genome Sequencing Museum Specimens
Melanie Pa ejo
1,2,
*, Da id W agg
3
, Do a Hen iques
4
, Jean-Daniel Cha ie` e
1
, and Andone Es onba
2
1
Ag oscope, Swiss Bee Resea ch Cen e , Be n, Swi ze land
2
Lab. Gene ics, Depa men o Gene ics, Facul y o Science and Technology, Uni e si y o he Basque Coun y (UPV/EHU), Leioa, Spain
3
The Roslin Ins i u e, Uni e si y o Edinbu gh, Edinbu gh, Uni ed Kingdom
4
Ins i u o Poli 
ecnico de B aganc¸a, Cen o de In es igac¸~
ao de Mon anha (CIMO), B aganc¸a, Po ugal
*Co esponding au ho : E-mails: melanie.pa [email protected], melanie.pa [email p o ec ed].
Accep ed: 24 Augus 2020
Da a deposi ion: This p ojec has been deposi ed a he Eu opean Nucleo ide A chi e (h p://www.ebi.ac.uk/ena, las accessed Sep embe 5,
2020) unde he accession PRJEB34590.
Abs ac
His o ical specimens in museum collec ions p o ide oppo uni ies o gain insigh s in o he genomic pas . Fo he Wes e n honey bee,
Apis melli e a L., his is pa icula ly impo an because i s popula ions a e cu en ly unde h ea wo ldwide and ha e expe ienced
many changes in managemen and en i onmen o e he las cen u y. Using Swiss Apis melli e a melli e a as a case s udy, ou
esea ch p o ides impo an insigh s in o he gene ic di e si y o na i e honey bees p io o he indus ial-scale in oduc ions and
ade o non-na i e s ocks du ing he 20 h cen u y— he onse o in ensi e comme cial b eeding and he decline o wild honey bees
ollowing he a i al o Va oa des uc o . We sequenced whole-genomes o 22 honey bees om he Na u al His o y Museum in Be n
collec ed in Swi ze land, including he oldes A. melli e a sample e e sequenced. We iden i y bo h, a his o ic and a ecen mig an ,
na u al o human-media ed, which co obo a es wi h he popula ion his o y o honey bees in Swi ze land. Con a y o wha we
expec ed, we ind no e idence o a signi ican gene ic bo leneck in Swiss honey bees, and ind ha gene ic di e si y is no only
main ained, bu e en sligh ly inc eased, mos p obably due o mode n apicul u al p ac ices. Finally, we iden i y signals o selec ion
be ween his o ic and mode n honey bee popula ions associa ed wi h genes en iched in unc ions linked o xenobio ics, sugges ing a
possible selec i e p essu e om he inc easing use and di e si y o chemicals used in ag icul u e and apicul u e o e he las cen u y.
Key wo ds: Apis melli e a melli e a, museum genomics, gene ic di e si y, selec ion signa u es, haplo ype phasing,
biodi e si y.
Significance
Li le is known abou na i e honey bees’ gene ic di e si y and s uc u e p eag icul u al and apicul u al e olu ions du ing he 20 h
cen u y— he beginning o comme cial bee b eeding and decline o wild honey bees ollowing he a i al an in asi e ec opa asi e. We
ind no educ ion in gene ic di e si y o a his o ic honey bee popula ion compa ed wi h i s con empo a y conspeci ics. We u he iden i y
genes en iched in unc ions linked o immuni y, and he de oxi ica ion o possible ag ochemicals. The esul s do no only e eal no el
insigh s in o he honey bee genomic pas , bu also p o ide aluable baseline genomic da a o na i e popula ions aiming a making
imp o ed conse a ion managemen decisions. In addi ion, ou app oach o sequence honey bee museum samples se es as a case s udy
o he sequencing o o he p ecious museum specimens.
ßThe Au ho (s) 2020. Published by Ox o d Uni e si y P ess on behal o he Socie y o Molecula Biology and E olu ion.
This is an Open Access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion Non-Comme cial License (h p://c ea i ecommons.o g/licenses/by-nc/4.0/), which pe mi s
non-comme cial e-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. Fo comme cial e-use, please con ac jou nals.pe [email protected]
Genome Biol. E ol. 12(12):2535–2551. doi:10.1093/gbe/e aa188 Ad ance Access publica ion 2 Sep embe 2020 2535
GBE
Downloaded om h ps://academic.oup.com/gbe/a icle/12/12/2535/5900668 by UNIVERSIDAD DEL PAIS VASCO use on 02 Ma ch 2021
In oduc ion
Fo he mos impo an pollina o o wild and cul u ed plan s,
he Wes e n honey bee (Apis melli e a)(Klein e al. 2007;
Gallai e al. 2009;IPBES 2016), much has changed in he
las decades. Today, i is unde p essu e globally om new
in asi e pa asi es and eme gen pa hogens, inc eased use o
pes icides, habi a loss, and clima e change (Neumann and
Ca eck 2010;Po s e al. 2010;Vanengelsdo p and
Meixne 2010), culmina ing in majo losses o managed
honey bee colonies wo ldwide (Liu e al. 2016;Maggi e al.
2016;G ay e al. 2019;Mo awe z e al. 2019).
One o he p ima y ac o s d i ing colony losses is he ec-
opa asi ic mi e, Va oa des uc o , and i s associa ed i uses
(Guzm
an-No oa e al. 2010;Daina e al. 2012). In he la e
1970s, his pa asi e, na i e o Asia, sp ead h oughou
Wes e n Eu ope and No h Ame ica decima ing wild
A. melli e a colonies. Such is he scale o he h ea ha oday
he majo i y o honey bees canno su i e wi hou human
in e en ion (Rosenk anz e al. 2010). I is hus gene ally ac-
cep ed ha e al honey bee colonies nea ly became ex inc
a e he a i al o he mi e (Mo i z e al. 2007;De la R
ua e al.
2009), al hough he e a e epo s o wild honey bee popula-
ions pe sis ing in la ge woodlands (Seeley 2017;Kohl and
Ru schmann 2018).
The widesp ead colony losses and associa ed popula ion
size decline migh ha e po en ially esul ed in a gene ic bo -
leneck o he emaining wild Eu opean honey bees. Such
popula ion collapses can lead o loss o gene ic di e si y and
he eby h ea en long- e m adap i e po en ial o u u e en i-
onmen al changes (F ankham e al. 2002;Allendo e al.
2013). Honey bees, due o hei haplodiploid ma ing sys em
and social o ganiza ion, a e pa icula ly sensi i e o inb eeding
dep ession (Zayed 2009). S udies ha e demons a ed ha
high in acolony di e si y dec eases pa hogen load (Desai
and Cu ie 2015) and enhances p oduc i i y (Old oyd e al.
1992;Ma ila and Seeley 2007), su i o ship (Ta py e al.
2013), he mo egula ion (Jones e al. 2004), and homeos asis
(Old oyd and Fewell 2007). Mo eo e , i has also been shown
ha locally adap ed honey bees ha e highe su i al (Bu¨chle
e al. 2014;Bu nham e al. 2019) and lowe pa hogen le els
(F ancis e al. 2014), om which ollows ha he e is a need o
conse e he unde lying geno ypic a ia ion (F ankham e al.
2002).
In he second hal o he 20 h cen u y, Eu ope unde wen
la ge-scale ag icul u al in ensi ica ion associa ed wi h d as ic
land-use changes, igge ing a signi ican decline in insec bio-
di e si y (Robinson and Su he land 2002; an Lexmond e al.
2015). One o he majo d i e s o his decline is he inc easing
use o pes icides (Le F
eon e al. 2010;Goulson e al. 2015).
The la e 20 h cen u y also wi nessed an in ensi ica ion o api-
cul u e wi h beekeepe s beginning o apply chemicals inside
he colony o con ol pes s and pa hogens (Johnson 2015).
Chemicals applied wi hin he hi e, such as mi icides and
an ibio ics, as well as ag ochemicals acqui ed ex e nally can
pe sis o many yea s in beeswax and a ec honey bee col-
onies in he long- e m (Mullin e al. 2010).In hesame ime
ame, apicul u e has u he expe ienced apid p o essional-
iza ion including mig a o y beekeeping, inc eased b eeding
e o s, and impo a ions o non-na i e subspecies and se-
lec ed s ock. Fo ins ance, h oughou la ge pa s o i s dis i-
bu ional ange he na i e Eu opean M-lineage honey bee
subspecies, Apis melli e a melli e a, has been eplaced by C-
lineage bees, mainly Apis melli e a ca nica,Apis melli e a lig-
us ica and Buck as bees p e e ed by beekeepe s (Pin o e al.
2014;Pa ejo e al. 2016).
Pas and con empo a y popula ions di e by na u al and
human-media ed ac o s and ci cums ances, such as bee-
keeping p ac ices, p e ailing pa hogens, o he pes icide e-
gime on c ops. Mode n honey bee popula ions ely hea ily on
human managemen , and hei gene ic composi ion is he e-
o e in luenced by comme cial ade (Vanengelsdo p and
Meixne 2010) and a i icial selec ion (W agg e al. 2016;
Pa ejo e al. 2017). Mo eo e , hey a e much mo e exposed
o he d as ic land-use changes and p e ailing ag icul u al
p ac ices o ecen imes. Th oughou much o Eu ope un il
he 1950s, honey bees we e much less in ensi ely managed,
mo e closely e lec ing na u al condi ions wi h li le o no hu-
man selec ion, and mos ly kep by swa m beekeeping, and
he eby in cons an gene low wi h he wild popula ion.
Gaining a g ea e unde s anding o he gene ic di e si y in
he pas can in o m ou unde s anding o he impac o he
ag icul u al and apicul u al e olu ions on honey bee
popula ions.
One powe ul way o in es iga e he changes be ween
pas and mode n popula ions is by analyzing samples ha
p eda e he d as ic en i onmen al and human-induced ans-
o ma ions. Museum specimens, he e o e, o e an excellen
oppo uni y by p o iding a window in o he pas (Lis e 2011).
Compa ing his o ic and con empo a y allelic equencies is
he mos di ec and powe ul way o de ec mic oe olu iona y
change (Mikheye e al. 2015). The main ca ea o museum
samples, howe e , is he di icul y o ob aining high-quali y
DNA o molecula gene ic analysis (S aa s e al. 2013), al-
hough imp o emen s in DNA ex ac ion p o ocols con inue
o be de eloped (Tin e al. 2014;Sp oul and Maddison 2017).
Un il ecen ly he majo i y o s udies using museum specimens
ha e been based on PCR-ampli ica ion o speci ic genes o
mi ochond ial DNA (Habel e al. 2009;La Haye e al. 2012).
Howe e , due o DNA deg ada ion, agmen s which a e
sho e han he PCR a ge egion canno be ampli ied (Tin
e al. 2014). The ecen ad ances in high- h oughpu se-
quencing enable us now o o e come he challenges o
ex ac ing genomic in o ma ion om museum specimens,
as mos me hods a e designed o sho agmen ed DNA
(S aa s e al. 2013;Bu ell e al. 2015). In he ield o human
e olu ion, p o ocols o high- h oughpu sequencing applied
o ancien DNA om a chaeological si es a e well es ablished
Pa ejo e al. GBE
2536 Genome Biol. E ol. 12(12):2535–2551 doi:10.1093/gbe/e aa188 Ad ance Access publica ion 2 Sep embe 2020
Downloaded om h ps://academic.oup.com/gbe/a icle/12/12/2535/5900668 by UNIVERSIDAD DEL PAIS VASCO use on 02 Ma ch 2021
( e iewed by Sla kin and Racimo 2016). Howe e , ewe
e o s ha e been made in he applica ion o his o ical mu-
seum collec ions om animals and plan s (e.g., sequencing o
he mi ogenome as in Mille e al. 2008;Hung e al. 2013;
Besna d e al. 2016), and only mos ecen ly high-densi y ge-
nome-wide analyses o museum specimens ha e been
epo ed (Mikheye e al. 2015;Linck e al. 2017;S
anchez
Ba ei o e al. 2017;C idland e al. 2018).
To da e, o honey bees, he e ha e been wo s udies
published using whole-genome sequence da a o museum
specimen, bo h o which conce ned he in oduced ange
o A. melli e a in No h Ame ica. Mikheye e al.(2015) in es-
iga ed he genomic changes o e a 33-yea pe iod o a wild
honey bee popula ion in I haca ollowing he in oduc ion o
V. des uc o . The au ho s ound e idence o a mi ochond ial
bo leneck bu wi h li le loss o nuclea gene ic di e si y and
popula ion size. C idland e al. (2018) documen ed he em-
po al gene ic changes in Cali o nian popula ions wi h no h-
e n popula ions expe iencing a shi in gene ic ances y om
M- o C-lineage since he 1960s and sou he n popula ions
unde going A icaniza ion.
He e, we hypo hesize ha he d as ic changes ega ding
popula ion decline, ag icul u al and apicul u al in ensi ica ion
du ing he las decades ha e had p o ound e ec s on he
gene ic di e si y and ances y o na i e honey bee popula ions
and le signa u es o selec ion on hei genomes. Using he
Swiss da k honey bee popula ion as a case s udy, we se-
quenced museum specimens da ed be ween 1879 and
1959 ha we e p o ided by he Na u al His o y Museum in
Be n, Swi ze land, o in es iga e he genomic pas o he na-
i e A. m. melli e a bees. Na i e Swiss honey bees ha e su -
e ed om a se e e popula ion size decline in ecen decades
due, in pa , o Va oa mi es, as well as in oduc ions and
eplacemen s o non-na i e s ocks. Toge he wi h whole-
genome sequence da a om a p e ious s udy o con empo-
a y Swiss bees (Pa ejo e al. 2016), we in es iga ed gene ic
di e si y, mi ochond ial DNA haplo ypes, admix u e, and se-
lec ion signa u es. To ou knowledge, his is he i s s udy o
whole-genome sequence his o ic honey bee specimen om
hei na i e ange.
Ma e ials and Me hods
Samples
Museum Samples
Twen y- wo Swiss A. m. melli e a museum specimens da ed
be ween 1879 and 1959 (61–141 yea s old) we e ob ained
om he Na u al His o y Museum in Be n, Swi ze land ( ig. 1).
Specimen consis ed o d ied and pinned wo ke bees (diploid)
s o ed by he museum bu o igina ing om se e al p i a e
collec ions o Swiss en omologis s. Samples ha e been
assigned wi h a QR-code deposi ed in he Na u al His o y
Museum in Be n (supplemen a y able S1,Supplemen a y
Ma e ial online).
Mode n Samples
Whole-genome sequence da a we e a ailable om a p e ious
s udy (Pa ejo e al. 2016) o which we selec ed 40 pu e Swiss
A. m. melli e a d ones (haploid). These samples co e a sligh ly
la ge geog aphical ange han he a ailable museum bees,
bu due o b eeding ac i i ies hey a e e y simila o each
o he ep esen ing he cu en popula ion. In addi ion, o in-
es iga e o e all gene ic s uc u e and admix u e p opo ions,
36 honey bees om a di e en e olu iona y lineage widely
employed by beekeepe s in Swi ze land we e included in
some analyses. These included 24 A. m. ca nica and 12 A.
m. ligus ica d ones (supplemen a y able S1,Supplemen a y
Ma e ial online) om Pa ejo e al. (2016) and Hen iques e al.
(2018).
DNA Ex ac ion and Sequencing
Genomic DNA was ex ac ed om he hind legs o museum
specimen ( ig. 1) ca e ully insed wi h Ringe solu ion, using a
phenol–chlo o o m–isoamyl alcohol (25:24:1) me hod
(Ausubel 1988). Pai -end (2 125 bp) lib a ies (ki ) we e p e-
pa ed ollowing manu ac u e s p o ocol using he NEBNex
Ul a II ki (New England Biolabs, Inc) and sequenced on he
Illumina HiSeq3000 pla o m wi h 20 samples pe lane.
Mapping, Va ian Calling, and Single Nucleo ide
Polymo phism Se s
Mapping
Raw sequence da a om mode n and his o ic samples we e
p ocessed using Cu adap 1.8 (Ma in 2011) o emo e
Illumina uni e sal adap o s and keep only eads wi h minimum
leng hs o 20 bp and a minimum base quali y sco e o 20.
T immed eads we e hen mapped agains he da k honey
bee e e ence genome INRA_AMelMel_1.0 (www.ncbi.
nlm.nih.go /assembly/GCA_003314205.1, las accessed
Sep embe 5, 2020) using bwa mem 0.7.10 (Li and Du bin
2009). PCR duplica es we e ma ked using PICARD 2.18.23
(h p://b oadins i u e.gi hub.io/pica d/, las accessed
Sep embe 5, 2020). Mapping s a is ics including dep h o
co e age and pe cen age o mapped eads we e calcula ed
using sam ools 1.7 (Li e al. 2009) and GATK 4.1.0.0
(Mckenna e al. 2010;Van De Auwe a e al. 2013).
DamageP o ile (h ps://damagep o ile . ead hedocs.io/en/
la es /index.h ml, las accessed Sep embe 5, 2020) was
used on museum samples o gene a e damage p o iles o
he mapped DNA eads caused by deamina ion o cy osine
o e ime which leads o misinco po a ions o G!Aa he5
0
and C!Ta he3
0ends (B iggs 2010;Sawye e al. 2012).
Whole-Genome Sequencing Museum Specimens GBE
Genome Biol. E ol. 12(12):2535–2551 doi:10.1093/gbe/e aa188 Ad ance Access publica ion 2 Sep embe 2020 2537
Downloaded om h ps://academic.oup.com/gbe/a icle/12/12/2535/5900668 by UNIVERSIDAD DEL PAIS VASCO use on 02 Ma ch 2021
Va ian Calling
To inc ease a ian con idence, single nucleo ide polymo -
phism (SNP) calling o he A. m. melli e a (N
MUSEUM
¼22
wo ke s, N
MODERN
¼40 d ones) was pe o med using wo di -
e en so wa e ools (GATK’s Haplo ypecalle and SAM ools
mpileup). Fi s , ollowing GATK’s bes p ac ices, indi idual
GVCFs we e p oduced using Haplo ypecalle wi h pa ame-
e s: minimum mapping quali y ¼20, max al e na e
alleles ¼2, minimum quali y sco e ¼20, and sample-
ploidy ¼2 o museum (diploid wo ke s) and sample-
ploidy ¼1 o mode n samples (haploids d ones).
Subsequen ly, GVCFs we e combined and geno yped o p o-
duce a VCF- ile con aining aw a ian s o all A. m. melli e a
samples. Va ian s we e il e ed acco ding GATK’s ha d-
il e ing ecommenda ion (MQ <40.0, FS >60.0, QUAL <
30.0, MQRankSum <12.5, ReadPosRankSum <8.0) ex-
cep o quali y by dep h (QD), whe e a s ic e il e was ap-
plied (QD >5) a e QD dis ibu ion was in es iga ed be o e
and a e il e ing (supplemen a y igs. S1 and S2,
Supplemen a y Ma e ial online). Second, mul isample SNP
calling was also pe o med using SAM ools/BCF ools mpileup
1.7 (Li e al. 2009) wi h pa ame e s mapQuali y (q>30),
baseQuali y (Q>20), and il e ing low-quali y a ian s
(QUAL <30). Bo h call se s (GATK and SAM ools) we e il-
e ed on dep h (minimum 5, maximum 3*a e age DP) and
o include only biallelic SNPs on ch omosomes 1–16. Finally,
he a ian s om bo h call se s we e me ged using BCF ools
isec o keep only SNPs iden i ied in bo h se s. Va ian calling
s a is ics we e calcula ed wi h BCF ools s a s.
Anno a ion
No published anno a ion is a ailable o he da k honey bee
e e ence genome (A. m. melli e a; INRA_AMelMel_1.0).
Thus, anno a ion iles (g , anno a ion elease 104) om he
la es honey bee genome Amel_HA 3.1 (Wallbe g e al.
2019) we e emapped on o he INRA_AMelMel_1.0 genome
using NCBI’s emapping se ice (www.ncbi.nlm.nih.go /ge-
nome/ ools/ emap, las accessed Sep embe 5, 2020).
Finally, a cus om da abase o SnpE 4.3 (Cingolani e al.
2012) as pe so wa e ins uc ions was gene a ed o anno a e
he a ian s and p edic hei po en ial e ec s excluding in e -
genic, up- and downs eam anno a ions.
Haplo ype Phasing
Phasing geno ypes in o haplo ypes is a undamen al equi e-
men o some analyses, such as ha o ex ended haplo ype
homozygosi y (see below), which seek o exploi linkage dis-
equilib ium (LD) be ween ma ke s. S a is ical phasing can be
pe o med wi h o wi hou a e e ence panel, and gene ally
he use o an ex e nal e e ence panel has been shown o
inc ease phasing accu acy (Delaneau e al. 2013). Howe e ,
using haplo ypes om he mode n d one (haploid) da a se o
phase he his o ic wo ke (diploid) da a se isks po en ially
Fig. 1.—Sampling si es o he 22 A. m. melli e a museum specimens. Mos samples o igina e om he egion a ound Be n da ing be ween 1941 and
1959, bu some a e om moun ain a eas. The oldes sample is om Luze n (1879), Cen al Swi ze land. The second oldes sample (1884) is om Ze ma ,
Valais, in he Sou he n Alps. Map c ea ed wi h Da aw appe (www.da aw appe .de, accessed Feb ua y 2020).
Pa ejo e al. GBE
2538 Genome Biol. E ol. 12(12):2535–2551 doi:10.1093/gbe/e aa188 Ad ance Access publica ion 2 Sep embe 2020
Downloaded om h ps://academic.oup.com/gbe/a icle/12/12/2535/5900668 by UNIVERSIDAD DEL PAIS VASCO use on 02 Ma ch 2021
c ea ing a i ac ual museum haplo ypes ha a e mode n
ecombinan s o ances al a ia ion. We he e o e es ed he
impac o phasing bo h wi h and wi hou he use o he mod-
e n bees as a e e ence panel on he museum bees.
Fi s , SNPs o all samples we e il e ed on call a e >0.95
using VCF ools (Danecek e al. 2011) lea ing 2,651,904 SNPs
and missing SNPs we e impu ed wi hin SHAPEIT4 (Delaneau
e al. 2019). P elimina y analyses masking 5% o he SNPs
wi h 100% call a e e ealed ha impu a ion using his ap-
p oach is highly accu a e (>90% accu acy, supplemen a y
ig. S6,Supplemen a y Ma e ial online) while keeping a la ge
numbe o SNPs. Subsequen ly, he da a we e phased wi h-
ou a e e ence panel in museum and mode n da a se s inde-
penden ly, and using he sel -impu ed mode n bees as
e e ences we phased he unphased museum da a once
mo e. Phasing was pe o med wi h SHAPEIT4 (Delaneau
e al. 2019) wi h he sequencing lag, a minimum window
size o 0.1 Mb ( he minimum pe mi ed by SHAPEIT4) and an
e ec i e popula ion (Ne) size o 150,000 which app oxima es
he Ne calcula ion o Wallbe g e al. (2014) o he No he n
A. m. melli e a samples hey s udied. Gene ic maps o phas-
ing we e gene a ed by li ing o e he Amel4.5 e e ence ge-
nome physical posi ions in he c osso e da a gene a ed by Liu
e al. (2015) o hose o INRA_AMelMel_1.0, and used he
c osso e in o ma ion o es ima e gene ic posi ions o each
SNP. The combined phased da a se s including ei he e e -
ence o sel -phased museum wo ke bees we e il e ed on
mino allele equency (MAF >0.05), each lea ing
1,828,439 SNPs o signa u es o selec ion analyses and link-
age decay es ima ion.
Mi ochond ial DNA
Va ian s in he mi ochond ial genome o he A. m. melli e a
bees we e called using SAM ools mpileup 1.7 wi h he—
ploidy 1 op ion and keeping only SNP a ian s wi h high qual-
i y (QUAL >30). The esul ing 254 SNPs we e geno yped in
he C-lineage samples using GATK’s Haplo ypecalle wi h
mode geno ype-gi en-alleles. C-lineage and A. m. melli e a
samples we e combined and SNPs wi h >20% missing calls
we e emo ed. This le 205 SNPs o pe o m Median-Joining
ne wo k analysis (Bandel e al. 1999) in PopART (Leigh and
B yan 2015).
Popula ion S uc u e Analyses
Combining Haploid D ones o Diploid Indi iduals
Popula ion s uc u e analyses a e sensi i e when haploid and
diploid da a se s a e analyzed oge he (Du esne e al. 2014;
W agg e al. 2016). To ge he mos in o ma i e and unbiased
esul s o popula ion s uc u e analyses, we he e o e an-
domly combined he haploid geno ypes o wo mode n A.
m. melli e a d ones using a cus om sc ip o gene a e in silico
diploids ollowing he p ocedu e in W agg e al. (2016).
Fo popula ion s uc u e analyses, he da a se wi h he
diploid museum samples and “diploidized” mode n samples
o A. m. melli e a was il e ed and p uned on LD using
PLINK1.9 (Chang e al. 2015). We applied –indep-pai wise
50 10 0.1 o il e ou a ian s wi h a co ela ion o >0.1 in
each window o 50 a ian s as ecommended by Alexande
e al. (2009) and kep only SNPs wi h a 100% call a e. This
le 59 K independen SNPs, which we e hen geno yped in
he C-lineage d ones using GATK’s Haplo ypecalle wi h
mode geno ype-gi en-alleles. Finally, simila o he mode n
A. m. melli e a samples, he haploid C-lineage d ones we e
andomly combined in o diploid indi iduals using a cus om
sc ip .
The combined da a se o he popula ion s uc u e analy-
ses comp ised 59 K SNPs geno yped in 60 samples (22 mu-
seum A. m. melli e a, 20 “diploidized” mode n A. m.
melli e a, 12 “diploidized” A. m. ca nica, and 6 “diploidized”
A. m. ligus ica). This da a se was used o es ima e he a e age
genome-wide di e gence, model-based ances y, and in p in-
cipal componen analysis (PCA).
Ances y, PCA, and Popula ion Di e en ia ion
To in e he gene ic ances y o each indi idual, we pe o med
model-based clus e ing as implemen ed in ADMIXTURE
(Alexande e al. 2009). We an he analysis unsupe ised
wi h 10,000 i e a ions o 1–5 hypo he ical ances al (K) clus-
e s. C oss- alida ion e o was es ima ed o each clus e and
used o de e mine he op imal numbe o K clus e s. We also
pe o med PCA o assess he popula ion s uc u e in he ab-
sence o a model (P ice e al. 2006). PCA was applied o he
pai wise gene ic ela ionships be ween all indi iduals (N¼60)
acco ding o hei iden i y-by-s a e alues compu ed in PLINK
1.9 (Chang e al. 2015). Admix u e and PCA esul s we e
p ocessed and plo ed in R (R De elopmen Co e Team
2013). Based on hese esul s, a single admixed museum
bee was iden i ied, which was subsequen ly excluded om
downs eam analyses (popula ion di e en ia ion, gene ic di-
e si y, linkage, and selec ion signa u es).
Popula ion di e en ia ion was es ima ed as mean pai wise
F
ST
(Wei and Cocke ham 1984) pe si e as implemen ed in
VCF ools (Danecek e al. 2011). The mean and con idence
in e als we e calcula ed om 10 andomly selec ed boo -
s ap samples o 10 mode n A. m. melli e a,10his o icA.
m. melli e a, and 10 C-lineage bees.
Gene ic Di e si y
To in e he adap i e po en ial wi hin he mode n and his o ic
A. m. melli e a popula ions wo gene ic di e si y measu es
we e employed: 1) Expec ed he e ozygosi y (H
Exp
) o each
indi idual was calcula ed using he da a se o 59 K unlinked
SNPs; and 2) nucleo ide di e si y (p;Nei 1982), which was
calcula ed om he whole-genome da a o each popula ion
Whole-Genome Sequencing Museum Specimens GBE
Genome Biol. E ol. 12(12):2535–2551 doi:10.1093/gbe/e aa188 Ad ance Access publica ion 2 Sep embe 2020 2539
Downloaded om h ps://academic.oup.com/gbe/a icle/12/12/2535/5900668 by UNIVERSIDAD DEL PAIS VASCO use on 02 Ma ch 2021

in window sizes o 10 kb wi h 5 kb o e lap using VCF ools
(Danecek e al. 2011). The mean and con idence in e als o
H
Exp
and pwe e calcula ed wi h 10 andomly selec ed sam-
ples o eacho 10mode nand10his o icA. m. melli e a.
LD Es ima ion
LD be ween each pai o SNPs in he mode n and his o ic
popula ions was calcula ed using he sel - and e e ence-
phased museum bees (N¼21 samples !N¼42 haplo ypes)
and he haploid mode n d ones (N¼40) by es ima ing he
Pea son’s squa ed co ela ion coe icien (
2
) in Plink 1.09
(Chang e al. 2015). Pai wise LD (as
2
alues) we e calcula ed
be ween all he pai s o SNPs wi hin each ch omosome based
on he exac solu ion o he Hill equa ion (Hill 1974;Gaun
e al. 2007) and applying a MAF il e o 0.05. The ex en o
LD decay was es ima ed based on physical dis ance be ween
each SNP pai . LD decay cu es we e calcula ed as he a e age
2
wi hin bins o 200 base pai s, up o a dis ance o 10 kbp
and plo ed in R.
Selec ion Signa u es Analyses
Genome Scan
Se e al measu es a e a ailable o in e signa u es o selec ion,
employing a ange o es s a is ics each wi h i s own limi s
and me i s ( e iewed in Vi i e al. 2013). He e, we employed
c oss-popula ion ex ended haplo ype homozygosi y (XP-EHH)
desc ibed by Sabe i e al. (2007). The XP-EHH es s a is ic is
basedonex endedhaplo ypeleng hwhichhasbeenshown
o be mos sui able o in e ecen selec ion and is especially
use ul o iden i ying ha d and so selec i e sweeps wi h
causa i e alleles ha ha e no eached ixa ion (Sabe i e al.
2007;Vi i e al. 2013). The XP-EHH analysis is based on hap-
lo ype leng h and is, hus, sensi i e o accu a e phasing.
Hence, we pe o med he analysis wice—once compa ing
he sel -phased museum o mode n bees, and secondly com-
pa ing museum bees phased wi h he mode n bees as a e -
e ence panel o he mode n bees. The analyses we e
pe o med using selscan 1.2.1 (Szpiech and He nandez
2014) wi h de aul MAF and EHH unca ion alues o 0.05.
The ob ained XP-EHH sco es o each SNP we e no malized
(Z- ans o med) by sub ac ing genome-wide mean XP-EHH
and di iding by he s anda d de ia ion (supplemen a y ig. S4,
Supplemen a y Ma e ial online). SNPs wi h absolu e Z- ans-
o med XP-EHH alues in he 99 h pe cen ile we e conside ed
signi ican . A simila app oach has been employed by o he
s udies in es iga ing selec ion signa u es in he honey bee
genome (Wallbe g e al. 2016;Mon e o-Mendie a e al.
2019). Finally, genes associa ed wi h SNPs anno a ed wi h
SNPe as in onic, exonic, 30UTR, 50UTR, and splice a ian s,
ha we e iden i ied in bo h genome scans (sel -phased and
e e ence-phased wo ke bees) a e conside ed as pu a i e
candida es unde selec ion.
Gene On ology En ichmen Analysis
Gene On ology (GO) anno a ions p o ide a con enien means
o g ouping genes by hei known unc ions and p edic ed
biological oles, enabling en ichmen analyses o be con-
duc ed. The online esou ce DAVID .6.8 (Da abase o
Anno a ion, Visualiza ion, and In eg a ed Disco e y; Huang
e al. 2007) was accessed in June 2020 o es i he candida e
genes iden i ied demons a ed en ichmen o any pa icula
unc ion (Huang e al. 2009). We used as a backg ound gene
se all honey bee genes associa ed wi h a leas one SNP in ou
analyses. Func ionally ela ed genes we e clus e ed using he
gene unc ional classi ica ion ool se o highes s ingency.
En ichmen o GO ca ego y e ms was pe o med wi h he
unc ional anno a ion analysis ool using he GO ca ego ies o
Biological P ocess, Cellula Componen , Molecula Func ion,
and KEGG pa hway. The unc ional anno a ion clus e ing ool
was subsequen ly used o clus e simila GO e ms.
Resul s
Mapping and SNP Calling
In o al, 686,376,114 sequencing eads we e gene a ed om
he 22 museum samples. A summa y o alignmen s a is ics o
hese sequence eads in addi ion o hose o he 76 mode n
samplesisp o idedinsupplemen a y able S1, and igu es S1
and S2, Supplemen a y Ma e ial online. The a e age mapping
a e ac oss he museum samples was 93.7%, wi h a mean
dep h o co e age o 13.9, anging om 5.55 o 27.39,
in compa ison wi h he mode n A. m. melli e a d ones wi h
mean 10.3and ange 7.3–21.2. Only 34% o eads o
sample Ki BE_1941 mapped o he e e ence genome, possi-
bly indica ing con amina ion, ne e heless, he dep h o co -
e age was 7.8and he sample e ained in downs eam
analyses. On a e age o museum and mode n samples, e-
spec i ely, 85% and 92.2% o he genome pe sample was
callable, ha ing a dep h o co e age >4, whe eas he low-
es b ead h o co e age was obse ed in Be BE_1947-2
(57.7%). The oldes sample, LuzLU_1879-2, e u ned a dep h
and b ead h o co e age o 9.02and 69.6%, espec i ely.
Analysis o DNA deg ada ion by DamageP o ile (supple-
men a y able S2 and ig. S5,Supplemen a y Ma e ial online)
indica ed only mino 50(2.1% 60.6SD)and3
0(2.4% 60.5
SD) misinco po a ions compa ed wi h s udies o ancien DNA
(e.g., 8%, Pel ze e al.2018). The misinco po a ions o wo
he oldes museum samples da ing om 1884 o 1879 we e
es ima ed a 3% and 4.5% o G!Aa he5
0end, whe eas
o he C!Ta he3
0end i was 2.5% and 4.3%, espec-
i ely. Howe e , because misma ches a he ends o he se-
quence eads a e so -clipped by bwa mem du ing alignmen ,
no addi ional ead il e ing o clipping was pe o med.
O e all, wi h excep o he wo oldes samples and
Be BE_1947-2 wi h he lowes b ea h o co e age, he se-
quence quali y o he museum samples can be conside ed
Pa ejo e al. GBE
2540 Genome Biol. E ol. 12(12):2535–2551 doi:10.1093/gbe/e aa188 Ad ance Access publica ion 2 Sep embe 2020
Downloaded om h ps://academic.oup.com/gbe/a icle/12/12/2535/5900668 by UNIVERSIDAD DEL PAIS VASCO use on 02 Ma ch 2021
compa able wi h hose o he mode n samples (supplemen-
a y ables S1 and S2,Supplemen a y Ma e ial online).
Va ian calling was pe o med using GATK’s
Haplo ypecalle and SAM ools mpileup, esul ing in
4,611,541 and 4,145,606 SNPs, espec i ely (supplemen a y
able S3,Supplemen a y Ma e ial online). Applying il e s o
emo e low-quali y a ian s and in e sec ing bo h a ian se s,
le 3,252,197 genome-wide SNPs. The a e age geno yping
call a e was 0.94 in he museum samples and 0.97 in he
mode n bees, wi h all samples excep o h ee exceeding
90% call a e (supplemen a y able S3A, igs. S6 and S7,
Supplemen a y Ma e ial online). The museum sample wi h
he lowes geno yping a e o 0.78 was Be BE_1947-2, which
was also he sample wi h he lowes mapping a e po en ially
due o con amina ion.
Di e en addi ional il e s we e applied o subsequen anal-
yses (supplemen a y able S3,Supplemen a y Ma e ial online):
Fo es ima ing nucleo ide di e si y and o phasing he da a se
was addi ionally il e ed on call a e 0.95 lea ing 2,651,904
SNPs, o linkage and XP-EHH a MAF il e o >0.05 was applied
lea ing 1,828,439 SNPs, and o PCA, model-based admix u e,
F
ST
,andH
Exp
he da a se was il e ed o unlinked SNPs wi h a
call a e o 100% lea ing 59,320 SNPs.
Mi ochond ial Haplo ype Ne wo ks
Mi ochond ial ne wo k analysis e ealed e idence o mi o-
nuclea disco dance. The na i e da k honey bee o
Swi ze land belongs o he M-lineage, bu we ound ha
one indi idual sampled in Liebe eld in 1959 (LieBE_1959-1)
and ano he sampled in he Swiss Alps in 1958 (LoeVS_1958)
possess mi ochond ial haplo ypes ha clus e wi h C-lineage
bees ( ig. 2). A phylogene ic analysis o SNPs iden i ied on he
comple e mi ochond ial genome showed ha all bu wo
mode n and his o ic A. m. melli e a samples clus e in wo
M-lineage clades (supplemen a y ig. S8,Supplemen a y
Ma e ial online). The wo his o ic samples ha a e placed in
a dis an b anch in he phylogene ic analysis a e he same
samples ha clus e wi hin he C-linage o he haplo ype
ne wo k.
Popula ion S uc u e
Popula ion s uc u e in e ed by he model-based ADMIXTURE
and PCA each e ealed ou (sub-) popula ions ep esen ing
his o ic A. m. melli e a, mode n A. m. melli e a,A. m. ca nica
and A. m. ligus ica bees ( ig. 3). The op imal numbe o clus e s
iden i ied om he lowes c oss- alida ion e o is K¼2(sup-
plemen a y ig. S9,Supplemen a y Ma e ial online), which co -
espond o he wo majo e olu iona y lineages: M- (A. m.
melli e a) and C-lineage (A. m. ca nica and A. m. ligus ica).
One o he museum bees, LieBE_1959-1 shows e idence o
gene ic admix u e wi h he C-lineage (60.5% M-lineage and
39.5% C-lineage ances y), as indica ed by he dual-colo ge-
ne ic backg ound in he ADMIXTURE plo ( ig. 3A)andi s
in e media e placemen along PC1 in he PCA plo ( ig. 3B).
LieBE_1959-1 is one o wo bees iden i ied om he mi ochon-
d ial analyses o possess C-lineage m DNA ( ig. 2).
When es ima ing popula ion di e en ia ion (F
ST
)using he
59 K SNP da a se , he museum bee iden i ied as being
admixed was excluded om he analysis. The F
ST
analysis
e ealed a high di e gence be ween C-lineage bees and he
his o ic and mode n A. m. melli e a popula ions (F
ST
>0.4),
and, whe eas he di e gence be ween he mode n and mu-
seum samples was expec edly e y low, i was signi ican ly
di e en om 0 based on andom subsampling (F
ST
¼
0.007, 95% CI 0.004–0.009).
Gene ic Di e si y
To in es iga e di e ences in gene ic di e si y and popula ion
his o ies be ween mode n and museum A. m. melli e a,we
calcula ed expec ed he e ozygosi y (H
Exp
) and nucleo ide di-
e si y (p). He e ozygosi y di e ed signi ican ly be ween bo h
popula ions (H
E(Museum)
¼0.243, 95% CI 0.241–0.245,
H
E(Mode n)
¼0.258, 95% CI 0. 257–0.259). The o mula o
H
Exp
is based on allele equencies and is he e o e no in lu-
enced by ploidy, no he gene a ion o in silico diploids (as we
es ed in p elimina y analyses). Mo eo e , we also es ima ed
nucleo ide di e si y which is calcula ed on haploid sequences
(phased geno ypes) and, hus, insensi i e o ploidy. Simila ly
o H
Exp
also nucleo ide di e si y was sligh ly highe in he
mode n A. m. melli e a popula ion (p¼0.00241, 95% CI
0.00240–0.00243) compa ed wi h he his o ic (p¼0.00227,
95% CI 0.00224–0.00230). Bo h measu es o gene ic di e -
si y being la ge in he mode n popula ion sugges enhanced
adap i e po en ial. Obse ed he e ozygosi y in he iden i ied
admixed museum bee (LieBE_1959-1) was conside ably
highe (H
Obs
¼0.42) han he expec ed he e ozygosi y
(H
Exp
¼0.26), and also highe han o all o he A. m. melli e a
samples, sugges ing he admix u e o be ecen .
LD Decay
LD decay be ween SNPs in he mode n and his o ic A. m.
melli e a popula ions as measu ed using
2
o e inc easing
dis ances be ween pai wise SNPs is shown in igu e 4.The
maximum a e age LD o SNPs less han 200 bps apa was
1.5 imes as high in he mode n popula ion (
2
¼0.43) com-
pa ed wi h he his o ic popula ion (
2
0.28). LD decays
quickly o bo h popula ions, bu long- ange LD was ound
o be conside ably lowe o he his o ic (
2
0.02) han he
mode n popula ion (
2
0.12) po en ially e lec ing he ecen
popula ion his o y o a small, inb ed o admixed popula ion.
The e is a sligh , bu insigni ican endency o LD o be highe
in museum bees phased wi h he d ones as e e ences ( ig. 4).
Whole-Genome Sequencing Museum Specimens GBE
Genome Biol. E ol. 12(12):2535–2551 doi:10.1093/gbe/e aa188 Ad ance Access publica ion 2 Sep embe 2020 2541
Downloaded om h ps://academic.oup.com/gbe/a icle/12/12/2535/5900668 by UNIVERSIDAD DEL PAIS VASCO use on 02 Ma ch 2021
Selec ion Signa u es be ween His o ic and Mode n A. m.
melli e a Popula ions
We in es iga ed he p esence o signa u es o selec ion be-
ween he his o ic and mode n A. m. melli e a popula ions
using he XP-EHH me hod (Sabe i e al. 2007) and wi h wo
di e en da a se s including ei he he sel -phased o
e e ence-phased museum wo ke haplo ypes. The e was
an 84% o e lap o he SNPs (15,354 SNPs) alling wi hin
he 99 h pe cen ile in bo h scans (supplemen a y ig. S10,
Fig. 2.—Median-joining ne wo k in e ed om 205 m DNA SNPs and 60 samples (N¼22 museum A. m. melli e a,N¼20 mode n A. m. melli e a,
N¼12 A. m. ca nica,andN¼6A. m. ligus ica). Hypo he ical (unsampled o ex inc ) haplo ypes a e deno ed as illed black ci cles. The alues in b acke s
indica e base pai di e ences be ween haplo ypes. M-lineage samples including mode n and his o ic A. m. melli e a a e g ouped in o wo clades, wi h he
excep ion o wo museum samples (LieBE_1959-1 and LoeVS_1958) which clus e wi h C-lineage bees (deno ed by he a ows).
Fig. 3.—Popula ion s uc u e in e ed om he LD-p uned 59 K SNPs and 60 samples (N¼22 museum A. m. melli e a,N¼20 “diploidized” mode n A.
m. melli e a,N¼12 “diploidized” A. m. ca nica,andN¼6 “diploidized” A. m. ligus ica). (A) Gene ic ances y as calcula ed wi h ADMIXTURE o K¼2 o4
hypo he ical ances al popula ions. Each colo ep esen s one o Kclus e s. Each indi idual is ep esen ed by a ho izon al ba and colo ed acco ding o he
p opo ion o he genome ha was de i ed om each clus e . The op imal numbe o clus e s iden i ied by c oss- alida ion is K¼2. (B) PCA o gene ic
dis ance be ween indi iduals. The i s p incipal componen (PC1) explains 97% o he a ia ion indica ing s ong di e gence be ween M- and C-lineage
honey bees, whe eas PC2 accoun s only o 0.2% o he a iance.
Pa ejo e al. GBE
2542 Genome Biol. E ol. 12(12):2535–2551 doi:10.1093/gbe/e aa188 Ad ance Access publica ion 2 Sep embe 2020
Downloaded om h ps://academic.oup.com/gbe/a icle/12/12/2535/5900668 by UNIVERSIDAD DEL PAIS VASCO use on 02 Ma ch 2021
Supplemen a y Ma e ial online) which we e conside ed as be-
ing associa ed wi h pu a i e signa u es o selec ion ( ig. 5 and
supplemen a y able S4,Supplemen a y Ma e ial online). In a
i s sc eening, we iden i ied an ex eme peak on ch omo-
some 11 (supplemen a y ig. S11,Supplemen a y Ma e ial
online), which on u he in es iga ion was iden i ied o
mos likely be a duplica ion no cap u ed in he e e ence
genome due o 1) he a e age dep h o co e age in he e-
gion being 1.6 imes he ch omosome a e age, and 2) he
p esence o eads wi h bo h alleles ( e e ence and al e na e) in
he haploid d ones. Fu he mo e, his peak lies wi hin a low-
complexi y cen ome e and no gene is anno a ed wi hin
615 kb. We ha e hus excluded i om u he analyses.
The emaining SNPs a e associa ed wi h 644 candida e genes
(supplemen a y able S5,Supplemen a y Ma e ial online).
Many o he genes iden i ied a e uncha ac e ized loci
(29.8%) as he anno a ion o he honey bee genome is a
om comple e. O he 15,354 SNPs, 275 and 13 SNPs we e
p edic ed by SNPe o ha e MODERATE and HIGH impac ,
espec i ely (supplemen a y able S6,Supplemen a y Ma e ial
online). These included mos ly nonsynonymous base pai
changes, and also splice egion a ian s as well as s op los /
gained a ian s.
GO analysis is a s a egy o iden i y he mos impo an
biological p ocesses o candida e egions iden i ied in
whole-genome selec ion scans. Howe e , he powe o hese
analyses depends on he numbe and quali y o anno a ed
genes a ailable o he ocal o ganism (Yon Rhee e al. 2008).
We conduc ed a GO analysis using he online esou ce
Fig. 4.—LD be ween SNPs as measu ed by
2
(yaxis) o inc easing
dis ance be ween SNPs (xaxis) o A. m. melli e a mode n d ones (N¼40)
and A. m. melli e a museum bee haplo ypes (N¼42 haplo ypes) sel -
phased and phased using he d ones as a e e ence panel.
Fig. 5.—Signa u es o selec ion be ween his o ic and mode n A. m. melli e a om Swi ze land. XP-EHH was pe o med using 42 haplo ypes de i ed
om 21 museum samples (diploid) (A) sel -phased and (B) e e ence-phased, and 40 haplo ypes de i ed om mode n d ones (haploid). XP-EHH sco es a e
plo ed along he 16 honey bee ch omosomes wi h nega i e alues indica ing selec ion in he mode n popula ion. The dashed lines deno e SNPs in he 99 h
pe cen ile o he absolu e XP-EHH sco es. This igu e excludes he alse posi i e peak on ch omosome 11 (4945317-4945798), which can be seen in igu e
S11,Supplemen a y Ma e ial online. The i e highes peaks o each analysis a e labeled wi h hei pu a i e genes unde selec ion.
Whole-Genome Sequencing Museum Specimens GBE
Genome Biol. E ol. 12(12):2535–2551 doi:10.1093/gbe/e aa188 Ad ance Access publica ion 2 Sep embe 2020 2543
Downloaded om h ps://academic.oup.com/gbe/a icle/12/12/2535/5900668 by UNIVERSIDAD DEL PAIS VASCO use on 02 Ma ch 2021
Bonn, Ge many: Sec e a ia o he In e go e nmen al Science-Policy
Pla o m on Biodi e si y and Ecosys em Se ices.
Johnson RM. 2015. Honey bee oxicology. Annu Re En omol.
60(1):415–434.
Jones JC, Mye scough MR, G aham S, Old oyd BP. 2004. Honey bee nes
he mo egula ion: di e si y p omo es s abili y. Science
305(5682):402–404.
Klein A-M, e al. 2007. Impo ance o pollina o s in changing landscapes
o wo ld c ops. P oc R Soc B 274(1608):303–313. Vol. p
Kohl PL, Ru schmann B. 2018. The neglec ed bee ees: Eu opean beech
o es s as a home o e al honey bee colonies. Pee J 6:e4602.
La Haye MJJ, Neumann K, Koelewijn HP. 2012. S ong decline o gene
di e si y in local popula ions o he highly endange ed Common ham-
s e (C ice us c ice us) in he wes e n pa o i s Eu opean ange.
Conse Gene . 13(2):311–322.
Le F
eon V, e al. 2010. In ensi ica ion o ag icul u e, landscape composi-
ion and wild bee communi ies: a la ge scale s udy in ou Eu opean
coun ies. Ag ic Ecosys En i on. 137(1–2):143–150
Leigh JW, B yan D. 2015. PopART: ull- ea u e so wa e o haplo ype
ne wo k cons uc ion. Me hods Ecol E ol. 6(9):1110–1116.
Lemmon MA, Schlessinge J. 2010. Cell signaling by ecep o y osine
kinases. Cell 141(7):1117–1134.
Li H, e al. 2009. The sequence alignmen /map o ma and SAM ools.
Bioin o ma ics 25(16):2078–2079.
Li H, Du bin R. 2009. Fas and accu a e sho ead alignmen wi h
Bu ows–Wheele ans o m. Bioin o ma ics 25(14):1754–1760.
Linck EB, Hanna ZR, Sellas A, Dumbache JP. 2017. E alua ing hyb idiza-
ion cap u e wi h RAD p obes as a ool o museum genomics wi h
his o ical bi d specimens. Ecol E ol. 7(13):4755–4767.
Lis e AM. 2011. Na u al his o y collec ions as sou ces o long- e m da a-
se s. T ends Ecol E ol. 26(4):153–154.
Liu H, e al. 2015. Causes and consequences o c ossing-o e e idenced
ia a high- esolu ion ecombina ional landscape o he honey bee.
Genome Biol. 16(1):15.
Liu Z, e al. 2016. Su ey esul s o honey bee (Apis melli e a) colony losses
in China (2010–2013). J Apic Res. 55(1):29–37.
Maggi M, e al. 2016. Honeybee heal h in Sou h Ame ica. Apidologie
47(6):835–854.
Ma in M. 2011. Cu adap emo es adap e sequences om high-
h oughpu sequencing eads. EMBne J. 17(1):10–12.
Ma ila HR, Seeley TD. 2007. Gene ic di e si y in honey bee colonies
enhances p oduc i i y and i ness. Science 317(5836):362–364.
Mckenna A, e al. 2010. The genome analysis oolki : a MapReduce ame-
wo k o analyzing nex -gene a ion DNA sequencing da a. Genome
Res. 20(9):1297–1303.
Miguel I, e al. 2011. Bo h geome ic mo phome ic and mic osa elli e
da a consis en ly suppo he di e en ia ion o he Apis melli e a M
e olu iona y b anch. Apidologie 42(2):150–161.
Mikheye AS, Tin MM, A o a J, Seeley TD. 2015. Museum samples e eal
apid e olu ion by wild honey bees exposed o a no el pa asi e. Na
Commun. 6:7991.
Mille W, e al. 2008. The mi ochond ial genome sequence o he
Tasmanian ige (Thylacinus cynocephalus). Genome Res.
19(2):213–220.
Mon e o-Mendie a S, e al. 2019. The genomic basis o adap a ion o
high-al i ude habi a s in he eas e n honey bee (Apis ce ana). Mol
Ecol. 28(4):746–760.
Mo awe z L, e al. 2019. Heal h s a us o honey bee colonies (Apis melli-
e a) and disease- ela ed isk ac o s o colony losses in Aus ia. PLoS
One 14(7):e0219293.
Mo eau-Fau a que C, Taillebou g E, P 
ea T, Du a J-M. 2002. Mu a ion o
lino e causes beha io al de ec s independen ly o pigeon in
D osophila. Neu oRepo 13(17):2309–2312.
Mo i z RF, K aus FB, K yge P, C ewe RM. 2007. The size o wild honeybee
popula ions (Apis melli e a) and i s implica ions o he conse a ion o
honeybees. J Insec Conse . 11(4):391–397.
Mullin CA, e al. 2010. High le els o mi icides and ag ochemicals in No h
Ame ican apia ies: implica ions o honey bee heal h. PLoS One
5(3):e9754.
Nei M. 1982. E olu ion o human aces a he gene le el. Hum Gene A
167–181.
Neumann P, Ca eck NL. 2010. Honey bee colony losses. J Apic Res.
49(1):1–6.
Ng TH, Ku z J. 2020. Dscam in immuni y: a ques ion o di e si y in insec s
and c us aceans. De Comp Immunol. 105:103539.
Old oyd BP, Fewell J. 2007. Gene ic di e si y p omo es homeos asis in
insec colonies. T ends Ecol E ol. 22(8):408–413.
Old oyd BP, Rinde e T, Ha bo J, Buco S. 1992. E ec s o in acolonial
gene ic di e si y on honey bee (Hymenop e a: Apidae) colony pe o -
mance. Ann En omol Soc Am. 85(3):335–343.
Oxley PR, Spi ak M, Old oyd BP. 2010. Six quan i a i e ai loci in luence
ask h esholds o hygienic beha iou in honeybees (Apis melli e a).
Mol Ecol. 19(7):1452–1461.
Pa ejo M, e al. 2016. Using whole-genome sequence in o ma ion o
os e conse a ion e o s o he Eu opean Da k Honey Bee, Apis
melli e a melli e a. F on Ecol E ol. 4:140. doi:10.3389/
e o.2016.00140.
Pa ejo M, W agg D, Hen iques D, Vignal A, Neudi schko M. 2017.
Genome-wide scans be ween wo honeybee popula ions e eal pu-
a i e signa u es o human-media ed selec ion. Anim Gene .
48(6):704–707.
Pel ze A, e al. 2018. In e ing gene ic o igins and pheno ypic ai s o
Geo ge B€
ah , he a chi ec o he D esden F auenki che. Sci Rep. 8(1).
doi:10.1038/s41598-018-20180-z
Pin o MA, e al. 2014. Gene ic in eg i y o he Da k Eu opean honey bee
(Apis melli e a melli e a) om p o ec ed popula ions: a genome-wide
assessmen using SNPs and m DNA sequence da a. J Apic Res.
53(2):269–278.
Po s SG, e al. 2010. Global pollina o declines: ends, impac s and d i -
e s. T ends Ecol E ol. 25(6):345–353.
P ice AL, e al. 2006. P incipal componen s analysis co ec s o
s a i ica ion in genome-wide associa ion s udies. Na Gene .
38(8):904–909.
P ies ley CM, Williamson EM, Wa o d KA, Sa elle DB. 2003. Thymol, a
cons i uen o hyme essen ial oil, is a posi i e allos e ic modula o o
human GABAA ecep o s and a homo-oligome ic GABA ecep o
om D osophila melanogas e . B J Pha macol. 140(8):1363–1372.
P owell DP, McMichael M, Sil ain J-F. 2004. Mul ilocus gene ic analysis o
hos use, in og ession, and specia ion in hos s ains o all a mywo m
(Lepidop e a: Noc uidae). Ann En omol Soc Am. 97(5):1034–1044.
R De elopmen Co e Team. 2013. R: a language and en i onmen o
s a is ical compu ing. Vienna, Aus ia: R Founda ion o S a is ical
Compu ing.
Robinson RA, Su he land WJ. 2002. Pos -wa changes in a able a ming
and biodi e si y in G ea B i ain. J Appl Ecol. 39(1):157–176.
Roe schi A, Be houd H, Kuhn R, Imdo A. 2008. In ec ion a e based on
quan i a i e eal- ime PCR o Melissococcus plu onius, hecausal
agen o Eu opean oulb ood, in honeybee colonies be o e and a e
apia y sani a ion. Apidologie 39(3):362–371.
Rosenk anz P, Aumeie P, Ziegelmann B. 2010. Biology and con ol o
Va oa des uc o . J In e eb Pa hol. 103:S96–119.
Sabe i PC, e al. 2007. Genome-wide de ec ion and cha ac e iza ion o
posi i e selec ion in human popula ions. Na u e 449(7164):913–918.
S
anchez Ba ei o F, e al. 2017. Cha ac e izing es ic ion enzyme-
associa ed loci in his o ic agweed (Amb osia a emisii olia) ouche
specimens using cus om-designed RNA p obes. Mol Ecol Resou .
17(2):209–220.
Pa ejo e al. GBE
2550 Genome Biol. E ol. 12(12):2535–2551 doi:10.1093/gbe/e aa188 Ad ance Access publica ion 2 Sep embe 2020
Downloaded om h ps://academic.oup.com/gbe/a icle/12/12/2535/5900668 by UNIVERSIDAD DEL PAIS VASCO use on 02 Ma ch 2021

Sawye S, K ause J, Guschanski K, Sa olainen V, P€
a€
abo S. 2012. Tempo al
pa e ns o nucleo ide misinco po a ions and DNA agmen a ion in
ancien DNA. PLoS One 7(3):e34131.
Seeley TD. 2017. Li e-his o y ai s o wild honey bee colonies li ing in
o es s a ound I haca, NY, USA. Apidologie 48(6):743–754.
Sgolas a F, e al. 2020. Bees and pes icide egula ion: lessons om he
neonico inoid expe ience. Biol Conse . 241:108356.
Shi JF, Zhu TT, Guo WC, Li GQ. 2016. Recep o y osine kinase genes
espond ansc ip ionally o suble hal doses o i e insec icides by a
mode-o -ac ion independen way in Lep ino a sa decemlinea a (Say). J
Asia-Pac En omol. 19(4):1103–1110.
Simon-Delso N, e al. 2015. Sys emic insec icides (neonico inoids and
ip onil): ends, uses, mode o ac ion and me aboli es. En i on Sci
Pollu Res. 22(1):5–34.
Sla kin M. 2008. Linkage disequilib ium—unde s anding he e olu iona y
pas and mapping he medical u u e. Na Re Gene . 9(6):477–485.
Sla kin M, Racimo F. 2016. Ancien DNA and human his o y. P oc Na l
Acad Sci USA. 113(23):6380–6387.
Smi h NM, e al. 2019. S ikingly high le els o he e ozygosi y despi e 20
yea s o inb eeding in a clonal honey bee. J E ol Biol. 32(2):144–152.
Sp oul JS, Maddison DR. 2017. Sequencing his o ical specimens: success ul
p epa a ion o small specimens wi h low amoun s o deg aded DNA.
Mol Ecol Resou . 17(6):1183–1201.
S aa s M, e al. 2013. Genomic easu e o es: comple e genome se-
quencing o he ba ium and insec museum specimens. PLoS One
8(7):e69189.
Szpiech ZA, He nandez RD. 2014. selscan: an e icien mul i h eaded p o-
g am o pe o m EHH-based scans o posi i e selec ion. Mol Biol E ol.
31(10):2824–2827.
Ta py DR, anEngelsdo p D, Pe is JS. 2013. Gene ic di e si y a ec s colony
su i o ship in comme cial honey bee colonies. Na u wissenscha en
100(8):723–728.
Tewhey R, Bansal V, To kamani A, Topol EJ, Scho k NJ. 2011. The impo -
ance o phase in o ma ion o human genomics. Na Re Gene .
12(3):215–223.
Thompson HM. 2003. Beha iou al e ec s o pes icides in bees— hei po-
en ial o use in isk assessmen . Eco oxicology 12(1/4):317–330.
Tin MM-Y, Economo EP, Mikheye AS. 2014. Sequencing deg aded DNA
om non-des uc i ely sampled museum specimens o RAD- agging
and low-co e age sho gun phylogene ics. PLoS One 9(5):e96793.
Toews DPL, B els o d A. 2012. The biogeog aphy o mi ochond ial and
nuclea disco dance in animals. Mol Ecol. 21(16):3907–3930.
Tong F, Coa s JR. 2010. E ec s o mono e penoid insec icides on [3H]-
TBOB binding in house ly GABA ecep o and 36Clup ake in
Ame ican cock oach en al ne e co d. Pes ic Biochem Physiol.
98(3):317–324.
Tsu uda JM, Ha is JW, Bou geois L, Danka RG, Hun GJ. 2012.
High- esolu ion linkage analyses o iden i y genes ha in luence
Va oa sensi i e hygiene beha io in honey bees. PLoS One
7(11):e48276.
Vanengelsdo p D, Meixne MD. 2010. A his o ical e iew o managed
honey bee popula ions in Eu ope and he Uni ed S a es and he ac o s
ha may a ec hem. J In e eb Pa hol. 103:80–95.
an Lexmond MB, Bonma in J-M, Goulson D, Noome DA. 2015.
Wo ldwide in eg a ed assessmen on sys emic pes icides. En i on Sci
Pollu Res. 22(1):1–4.
Vi i JJ, G ossman SR, Sabe i PC. 2013. De ec ing na u al selec ion in ge-
nomic da a. Annu Re Gene . 47(1):97–120.
Vogel C, Teichmann SA, Cho hia C. 2003. The immunoglobulin supe -
amily in D osophila melanogas e and Caeno habdi is elegans and he
e olu ion o complexi y. De elopmen 130(25):6317–6328.
Waliwi iya R, Bel on P, Nicholson RA, Lowenbe ge CA. 2010. E ec s o
he essen ial oil cons i uen hymol and o he neu oac i e chemicals
on ligh mo o ac i i y and wing bea equency in he blow ly
Phaenicia se ica a. Pes Manag Sci. 66(3):277–289.
Wallbe g A, e al. 2014. A wo ldwide su ey o genome sequence a i-
a ion p o ides insigh in o he e olu iona y his o y o he honeybee
Apis melli e a. Na Gene . 46(10):1081–1088.
Wallbe g A, e al. 2019. A hyb id de no o genome assembly o he hon-
eybee, Apis melli e a, wi h ch omosome-leng h sca olds. BMC
Genomics 20(1):275.
Wallbe g A, Pi k CW, Allsopp MH, Webs e MT. 2016. Iden i ica ion o
mul iple loci associa ed wi h social pa asi ism in honeybees. PLoS
Gene . 12(6):e1006097.
Wal on C, e al. 2008. Gene ic popula ion s uc u e and in og ession in
Anopheles di us mosqui oes in Sou h-eas Asia. Mol Ecol. 10(3):569–580.
Wa son FL. 2005. Ex ensi e di e si y o Ig-supe amily p o eins in he im-
mune sys em o insec s. Science 309(5742):1874–1878.
Wei BS, Cocke ham CC. 1984. Es ima ing F-s a is ics o he analysis o
popula ion s uc u e. E olu ion 38(6):1358–1370.
Wilson EB. 1905. The ch omosomes in ela ion o he de e mina ion o sex
in insec s. Science 22(564):500–502.
Wins on ML, D opkin JA, Taylo OR. 1981. Demog aphy and li e his o y
cha ac e is ics o wo honey bee aces (Apis melli e a). Oecologia
48(3):407–413.
W agg D, e al. 2016. Whole-genome esequencing o honeybee d ones
o de ec genomic selec ion in a popula ion managed o oyal jelly. Sci
Rep. 6:1–13.
W agg D, e al. 2018. Au osomal and mi ochond ial adap a ion ollowing
admix u e: a case s udy on he honeybees o Reunion Island. Genome
Biol E ol. 10(1):220–238.
Yon Rhee S, Wood V, Dolinski K, D aghici S. 2008. Use and misuse o he
gene on ology anno a ions. Na Re Gene . 9(7):509–515.
Zayed A. 2009. Bee gene ics and conse a ion. Apidologie 40(3):237–262.
Associa e edi o : And ea Be ancou
Whole-Genome Sequencing Museum Specimens GBE
Genome Biol. E ol. 12(12):2535–2551 doi:10.1093/gbe/e aa188 Ad ance Access publica ion 2 Sep embe 2020 2551
Downloaded om h ps://academic.oup.com/gbe/a icle/12/12/2535/5900668 by UNIVERSIDAD DEL PAIS VASCO use on 02 Ma ch 2021