469
Su ace swabs ou pe o m mos adi ional honeybee
(Apis melli e a) hi e samples o eco e y o eDNA and eRNA
E in Hill1, Liz Milla2, F ancisco Encinas-Viso2, James O'Dwye 3, Ben Gooden1, Ma iana Hoppe 4,
John Robe s1
1 CSIRO Heal h and Biosecu i y, Clunies Ross S , Ac on, ACT, Aus alia
2 CSIRO Cen e o Aus alian Na ional Biodi e si y Resea ch, Clunies Ross S , Ac on, ACT, Aus alia
3 CSIRO Aus alian Cen e o Disease P epa edness, Po a ling on Rd, Eas Geelong, VIC, Aus alia
4 CSIRO Heal h and Biosecu i y, Unde wood A e, Flo ea , WA, Aus alia
Co esponding au ho : E in Hill ([email p o ec ed])
Copy igh : © E in Hill e al.
This is an open access a icle dis ibu ed unde
e ms o he C ea i e Commons A ibu ion
License (A ibu ion 4.0 In e na ional – CC BY 4.0).
Resea ch A icle
Abs ac
Gene ic sampling om Eu opean honeybee (Apis melli e a) hi es is a p omising eposi-
o y o e es ial en i onmen al nucleic acids (eNA) om a ange o axonomic sou ces.
eNA samples om honeybee hi es can be used o moni o en i onmen s ia ege a ion
assessmen s and plan pa hogen de ec ion, while simul aneously assessing honeybee
hi e heal h h ough he iden i ica ion o honeybee pes s and pa hogens. Howe e , he e
is s ill limi ed compa ison ac oss he di e en ypes o hi e samples commonly used
o biomoni o ing. The e is also a need o less in asi e sampling me hods o acil-
i a e adop ion o hi e eNA su eillance a scale. The aim o his s udy was o com-
pa e he e ec i eness o using non-in asi e swabbing o hi e en ances wi h common
sample ypes collec ed om honeybee hi es (honey, pollen, and bees) o eNA de ec-
ion o a ious axa. Using a DNA me aba coding app oach o bac e ia (16S), animal
(COI), ungi (ITS1), and plan (ITS2) iden i ica ion, and me a ansc ip omic sequencing
o i us de ec ion, swabbing was ound o p o ide compa able o be e de ec ion o
axa compa ed wi h mo e in asi e sampling me hods. Swabs we e mos in o ma i e
o iden i ying bac e ial and ungal communi ies, while plan de ec ions we e simila
be ween swabs and honey samples. De ec ion o RNA i al di e si y was highes in hon-
ey samples, howe e , swabs also showed good eco e y o plan and insec i uses.
A minimum o wo swab eplica es was ecommended o inc ease axa yield o DNA
me aba coding. Ou s udy sugges s ha swabbing o hi e en ances is a highly e ec i e
biomoni o ing me hod o non-in asi ely collec ing hi e eNA and deli e ing comp ehen-
si e su eillance o animal, plan and mic obial communi ies in he en i onmen .
Key wo ds: Biosecu i y, en i onmen al biomoni o ing, honey, non-in asi e sampling,
pa hogens, pollen, su eillance, weeds
In oduc ion
Sampling en i onmen al nucleic acids (eNA), including DNA and RNA, has be-
come a co ne s one o non-in asi e moni o ing o wildli e, om he de ec ion
o a e and elusi e species and he cha ac e isa ion o species assemblages
wi hin a gi en en i onmen , o he su eillance o eme ging biosecu i y h ea s
Academic edi o :
A mando Espinosa P ie o
Recei ed:
22 July 2025
Accep ed:
13 Sep embe 2025
Published:
10 Oc obe 2025
Ci a ion: Hill E, Milla L, Encinas-Viso
F, O'Dwye J, Gooden B, Hoppe M,
Robe s J (2025) Su ace swabs
ou pe o m mos adi ional honeybee
(Apis melli e a) hi e samples
o eco e y o eDNA and eRNA.
Me aba coding and Me agenomics
9: e165436. h ps://doi.o g/10.3897/
mbmg.9.165436
Me aba coding and Me agenomics 9: 469–499 (2025)
DOI: 10.3897/mbmg.9.165436
470
Me aba coding and Me agenomics 9: 469–499 (2025), DOI: 10.3897/mbmg.9.165436
E in Hill e al.: Su ace swab eco e y o eDNA and eRNA
(Valen ini e al. 2016; T ujillo-González e al. 2019; Piggo e al. 2021). Much
o his ocus o da e has e ol ed a ound he use o wa e samples o species
moni o ing, wi h s anda dised sampling echniques and p o ocols o a ange
o aqua ic en i onmen s and de ec ion applica ions (Pawlowski e al. 2020;
Minamo o e al. 2021). While hese me hods a e eliable o species wholly
o pa ially dependen on aqua ic en i onmen s o su i al, al e na i e ech-
niques ha e been p oposed o he moni o ing o e es ial axa, including soil,
ai , sca s, and in e eb a e pollina o s (Bane jee e al. 2022; Van De Heyde e
al. 2022). Many o hese echniques a e s ill in hei in ancy and equi e ho -
ough examina ion o po en ial ad an ages and limi a ions be o e being b oadly
applied o e es ial biodi e si y moni o ing.
Gene ic sampling om Eu opean honeybee (Apis melli e a) hi es ep esen s
a p omising eposi o y o he de ec ion o e es ial axa om a ange o axo-
nomic sou ces. eNA me hods ha e been used ex ensi ely in honeybee hi es in
he pas , howe e , hese applica ions ha e mos ly assessed hi e heal h h ough
he de ec ion o honeybee pes s and pa hogens (Ribiè e e al. 2019; Ribani e al.
2020, 2022; Bio á e al. 2021), o iden i ying lo al composi ion o honey o unde -
s and honeybee o aging equi emen s (Hawkins e al. 2015; Danne e al. 2017),
and egional p o enance o honeys being sold (P osse and Hebe 2017; Milla e
al. 2021; Wi a e al. 2021). Mo e ecen ly, a en ion has shi ed o u ilising hon-
eybee hi e samples o he iden i ica ion o axa un ela ed o honeybee heal h
and p oduc ion ou pu s; hese ha e included assessmen s o plan biodi e si y
and weed de ec ion (Milla e al. 2022; Ba chelo e al. 2023), iden i ica ion o
plan pa hogens (Robe s e al. 2018b; T emblay e al. 2019; Robe s e al. 2023),
and de ec ion o honeydew p oducing insec s (U ze i e al. 2018). These s udies
demons a e he po en ial o honeybee hi es o be join ly used o a ange o
eNA biosecu i y and ecological moni o ing applica ions ha can bo h bene i he
heal h and ou pu s o he gi en hi e, as well as gaining an unde s anding o he
su ounding en i onmen and he incu sion o new pes s, weeds, and pa hogens.
One o he challenges o using honeybee hi es o biosecu i y and biodi e si y
moni o ing is he collec ion o app op ia e samples and how his can be e icien -
ly upscaled o allow ho ough sampling o b oad a eas o in e es . Commonly
u ilised samples such as honey, bees, and pollen equi e he assis ance o spe-
cialis s in he ield o open and handle hi es o hese samples o be collec ed,
which can be a limi ing ac o when unde aking la ge-scale s udies. The collec-
ion o in e nal hi e samples is also seasonally dependen as honey and pollen
would no be ound s o ed in su icien amoun s h oughou he yea and opening
hi es du ing colde mon hs o collec samples can isk he su i al o honeybee
colonies. Addi ionally, sample collec ion o honeybee ood sou ces (honey and
pollen) can also weaken hi es o e ime i long- e m moni o ing is unde aken.
A need he e o e exis s o explo e al e na i e op ions o sample eNA om honey-
bee hi es ha a e easy o non-specialis s o collec , a e conside ed non-in asi e
o he ongoing pe sis ence o hi es and is applicable h oughou he yea .
A p omising echnique is he use o swabs o collec non-in asi e samples
o eNA om inside honeybee hi es. To da e, his me hod has been explo ed in
he con ex o iden i ying axa ela ed o honeybee heal h, wi h a ying success
(Čukano á e al. 2023; Héna e al. 2023; Boa dman e al. 2024; Mackay e al.
2025; Robe s e al. 2025). Howe e , o ully unde s and he sui abili y o swabs
o he collec ion and de ec ion o axa o biodi e si y and biosecu i y impo ance,
471
Me aba coding and Me agenomics 9: 469–499 (2025), DOI: 10.3897/mbmg.9.165436
E in Hill e al.: Su ace swab eco e y o eDNA and eRNA
a compa ison be ween he e ec i eness o swabs o o he commonly used hi e
sample ypes such as pollen, bees, and honey o de ec a ange o axonomic
g oups is necessa y. Fu he mo e, he e is s ill li le known abou how hese con-
en ional sample ypes pe o m ela i e o each o he and he po en ial implica-
ions o hese di e ences on biomoni o ing and su eillance ou comes.
Mos eNA applica ions ocus on cap u ing DNA o species de ec ion, how-
e e he e is a g owing in e es in how eRNA can be e ec i ely collec ed and
used in biomoni o ing (Veilleux e al. 2021). RNA is expec ed o be less s able
in he en i onmen han DNA and may be able o p o ide in o ma ion on local
axa ha ha e been ecen ly ac i e in he en i onmen (Ya es e al. 2021; Sc i e
e al. 2023). Impo an ly, many i uses o in e es o biosecu i y su eillance
a e RNA i uses and he e o e unde ec able by eDNA me hods. Honeybee RNA
i uses can be impo an pa hogens, and many s udies ha e used samples o
bees and hi e p oduc s o in es iga e he hi e i ome (Robe s e al. 2018a;
Kadlečko á e al. 2022; Lee e al. 2023). Mul iple plan RNA i uses a e also
signi ican c op pa hogens, wi h many being pollen ansmi ed and commonly
de ec ed in hi e samples (Fe e s and Ashman 2023). Howe e , he e has been
li le examina ion o non-in asi e me hods o collec hi e eRNA o i us de ec-
ion (Čukano á e al. 2023).
The aims o his s udy we e o: (1) es he pe o mance o swab samples as
an ideal non-in asi e hi e eNA collec ion me hod, and (2) compa e swab sam-
ples o o he hi e sampling me hods including honey, pollen, and bees. We did
his by examining he axonomic composi ion o DNA and RNA collec ed om
swabs and hi e de i ed samples including honey, pollen, and bees, coupled
wi h a me aba coding app oach o iden i y bac e ial, animal, ungal, and plan
di e si y, and a me agenomic app oach o iden i y RNA i uses p esen wi hin
ou chosen hi e samples, o explo e he ull b ead h o axonomic di e si y pos-
sible wi hin a hi e. Replica e samples we e also collec ed o DNA me aba cod-
ing o examine he sampling e o equi ed o maximise axonomic di e si y
de ec ions. Sample ypes wi h he mos di e si y cap u ed we e iden i ied o
each axonomic g oup, and op imal sampling s a egies discussed o u u e
biomoni o ing and biosecu i y su eillance e o s.
Me hods
S udy si es and sampling
Se en sample ypes we e collec ed h ough he cou se o his s udy: hi e en-
ance swabs, whole bees collec ed om inside he hi e ( op box bees), honey,
in e nal pollen s o ed in pollen cells, hi e bo om swabs (hi e deb is), pollen
collec ed in hi e en ance pollen aps, and e u ning o age bees. All se en o
hese sample ypes we e compa ed o DNA me aba coding (bac e ia (16S),
animal (COI), ungi (ITS1), plan (ITS2)), while only a subse o hese sample
ypes we e used o he RNA i us compa isons (hi e en ance swabs, in e nal
pollen s o ed in pollen cells, op box bees, and honey). A summa y o he sam-
ples collec ed o his s udy can be ound in Table 1.
Each hi e chosen o sampling comp ised wo boxes, we e non-adja-
cen in he apia y whe e possible, and had s o es o pollen and honey. All
sampling and ield ools we e s e ilised p io o each sampling e en i s
472
Me aba coding and Me agenomics 9: 469–499 (2025), DOI: 10.3897/mbmg.9.165436
E in Hill e al.: Su ace swab eco e y o eDNA and eRNA
using 10% bleach solu ion, ollowed by deionised H2O, and hen 70% e ha-
nol. Equipmen used be ween hi es (e.g. hi e ool, weeze s, mic o spa ulas)
was cleaned wi h 70% e hanol in he ield p io o a ending each hi e. All
samples collec ed du ing each ip we e immedia ely placed on ice in he
ield, hen s o ed a -80 °C un il DNA/RNA ex ac ion. No s o age bu e s
we e used o any o he samples collec ed.
DNA sampling
Hi e samples collec ed o me aba coding analysis we e loca ed a
he Gungahlin Homes ead in C ace, Aus alian Capi al Te i o y (ACT)
(-35.223300°S, 149.121971°E), which a he ime o his s udy was a o me
pas o al p ope y occupied by he Commonweal h Scien i ic and Indus ial
Resea ch O ganisa ion (CSIRO). The si e is su ounded by g assland na u e
ese es o he no h and sou h, and esiden ial a eas o he eas and wes
o he p ope y. The apia y a his si e comp ised 12 hi es o al, o which
h ee hi es we e u ilised o his s udy. Sampling was unde aken on he 20 h,
24 h, and 27 h o Janua y 2023, wi h sample collec ions spli in o mo ning and
a e noon sessions. Hi es we e sampled in a di e en o de each ip o min-
imise any iming e ec s, pa icula ly om pollen apping.
In he mo ning sampling session, he hi e en ance was swabbed using a
s e ile nylon FLOQSwab (Copan Diagnos ics Inc.) dipped in deionised H2O.
Each swab was ubbed ac oss he en i e leng h o he hi e en ance ou imes,
wis ing i h oughou , and a emp ing o each a ange o dep hs inside he
hi e. Swabbing o he hi e en ance was epea ed wo mo e imes wi h a new
swab each ime. The hi e was hen opened, and bees p esen in he op box on
an ou e honey ame we e collec ed in a s e ile 50 mL alcon ube, collec ing
app oxima ely 50 bees. Following his, he op box and ames om he bo om
box we e emo ed o allow swabbing o he bo om o he hi e. Swabbing he
bo om o he hi e was ca ied ou as desc ibed abo e, a oiding any emaining
bees and concen a ing on a eas whe e hi e en ance swabs we e unlikely o
ha e eached. Th ee consecu i e swabs we e also collec ed o he bo om o
he hi e. The hi e was hen e-assembled, and a s e ilised pollen ap was a -
ached o he on o he hi e o collec pollen om he hind-legs o e u ning
o age bees. The ime he pollen ap was a ached was eco ded.
Table 1. Numbe o samples collec ed ac oss each sample ype and sampling day
du ing his s udy. Values in pa en heses ep esen he numbe o ex ac ion eplica es
aken om each sample. Ex ac ion eplica es we e no pe o med o any swab sam-
ples. RNA samples we e pooled o sequencing.
DNA samples RNA samples
Sample ype Day 1 Day 2 Day 3 To al samples Day 1 To al samples
Hi e en ance swab 9 9 9 27 6 3
Hi e deb is swab 9 9 9 27 - -
Honey - - 9 (×2) 18 3 (×2) 3
In e nal pollen - - 6 (×2) 12 3 (×2) 3
Pollen ap 3 (×2) 3 (×2) 3 (×2) 18 - -
Re u n o age bees 3 (×2) 3 (×2) 3 (×2) 18 - -
Top box bees 3 (×2) 3 (×2) 3 (×2) 18 3 (×2) 3
473
Me aba coding and Me agenomics 9: 469–499 (2025), DOI: 10.3897/mbmg.9.165436
E in Hill e al.: Su ace swab eco e y o eDNA and eRNA
On he inal sampling ip du ing he mo ning session, samples o honey and
in e nal pollen we e also collec ed om each hi e p io o swabbing he hi e bo -
om. These samples we e only collec ed on he inal sampling day due o hei
p esumed long-s o age pe iod, he e o e unlikely o show signi ican day- o-day
a ia ion. Th ee honey samples we e collec ed, all om he op box bu ac oss
sepa a e ames. App oxima ely 50 mL o honey was collec ed in a 100 mL speci-
men collec ion con aine om each ame. In e nal pollen was simila ly collec ed
ac oss wo ames; a o al o 10 pollen cells we e sampled om each ame, by
scooping ou he con en s o indi idual pollen cells using a s e ile mic o spa ula
and placing in o a 100 mL specimen collec ion con aine o each ame.
In he a e noon sampling session, he pollen collec ion ay unde nea h he
pollen ap was emo ed and he ime no ed, and pollen collec ed in o a 50 mL
alcon ube. Th oughou he expe imen , pollen aps we e closed o app oxi-
ma ely 5 hou s du ing he day o allow o pollen collec ion. Pollen aps we e
le open o e nigh o educe impac on he hi es. Fo age bees e u ning o he
hi e wi h isible pollen on hei hindlegs we e hen collec ed wi h s e ile wee-
ze s and placed in o a 50 mL alcon ube, wi h he aim o collec 20 e u ning
o age s pe hi e, howe e , his was no always possible. On he second sam-
pling ip o his s udy, a pollen ap was imp ope ly i ed o one o he hi es,
esul ing in no pollen p esen in he ap in he a e noon session. To ensu e a
sample was s ill collec ed om his hi e a oughly he same ime as he o he
hi es, he pollen ap was eposi ioned and le on he hi e o e nigh , wi h he
pollen collec ed a midday he ollowing day. Re u ning o age bees we e s ill
collec ed om his hi e a he in ended a e noon sampling session.
RNA sampling
Hi e samples collec ed o RNA sequencing we e loca ed a he CSIRO Black
Moun ain si e in Ac on, ACT (-35.275051°S, 149.112159°E). This si e bo de s
Black Moun ain and he Aus alian Na ional Bo anic Ga dens o he wes , he
Aus alian Na ional Uni e si y immedia ely eas , and esiden ial a eas o he
no h and sou h o he p ope y. The apia y a his si e comp ised a o al o i e
hi es, o which h ee hi es we e sampled. RNA sampling was unde aken in a
single session on he 11 h o Ap il 2024, wi h sampling me hods and o de he
same as abo e o he DNA samples, howe e , only wo swab eplica es we e
collec ed o RNA analysis. Samples aken o RNA analysis included hi e en-
ance swabs, op box bees, in e nal pollen cells and honey.
Sample p ocessing
Honey, pollen, and bee samples equi ed p e-p ocessing o p oduce iable sam-
ples o DNA and RNA ex ac ions, including a bead bea ing s ep o ensu e
samples we e homogenised and allow lysis o ough pollen, bac e ial, and un-
gal pa icles ha may ha e been p esen in he samples.
Honey samples collec ed om each hi e we e aliquo ed in o app oxima ely
10 mL olumes and each sample was weighed. Two eplica es om each ame
we e hen chosen o ex ac ions, wi h equal s a ing ma e ial weigh s conside ed
when choosing eplica es. Ul apu e H2O was hen added o each sample o c ea e
inal olumes o 50 mL and incuba ed a 40 °C o 30 minu es wi h egula mixing
474
Me aba coding and Me agenomics 9: 469–499 (2025), DOI: 10.3897/mbmg.9.165436
E in Hill e al.: Su ace swab eco e y o eDNA and eRNA
o homogenise he solu ion and dissol e suga s p esen . Samples we e hen cen-
i uged a 5000 pm o 20 minu es and he supe na an disca ded. Remaining
pelle s we e hen esuspended in 500 μL o 1 × PBS bu e and ans e ed o a
2 mL BeadBug™ homogenize ube con aining 1.5 mm zi conium beads.
All pollen samples we e weighed and homogenised p io o subsampling
o ex ac ions. Fo he in e nal pollen samples, 5 mL o 1 × PBS bu e was
added. Pollen ap samples we e weighed and homogenised using a mo -
a and pes le, and 1 g o pollen subsampled and 5 mL o 1 × PBS bu e
added. Some pollen ap samples weighed < 1 g so he en i e sample was
used, howe e , olumes o 1 × PBS bu e we e adjus ed o main ain a sim-
ila pollen:PBS bu e a io o he o he pollen ap samples. Samples we e
hen homogenised u he using a he momixe se a oom empe a u e o
1 hou a 1000 pm o pollen ames and pollen ap samples. Following his,
500 μL o solu ion was aliquo ed in o a 2 mL BeadBug™ homogenize ube
con aining 1.5 mm zi conium beads. Two eplica es we e aken om each o
he pollen ap and in e nal pollen ame samples.
The numbe o e u n o age bees pe sample was coun ed and placed in o
an ex ac ion bag (Sewa d) o homogenisa ion. Ca e was aken o ensu e pol-
len did no de ach om he bees’ legs du ing his p ocess. App oxima ely 10 mL
o PBS bu e was hen added o each bag, and he mix u e was ini ially c ushed
using a pes le on he ou side o he s aine bag. Following his, s aine bags
we e passed h ough a pas a olling machine (Biale i) mul iple imes o ensu e
bees we e ho oughly homogenised. The supe na an was hen collec ed, and
500 μL o solu ion was ans e ed o a 2 mL BeadBug™ homogenize ube con-
aining 1.5 mm zi conium beads. This p ocess was epea ed o he op box
bees, ensu ing an equi alen numbe o op box bees we e used in compa ison
o he e u n o age bees collec ed om he same hi e/sampling ip. Fo he
RNA samples, 20 op box bees we e used as s a ing ma e ial. Two eplica es
we e aken om each o he homogenised bee samples.
BeadBug ubes con aining sample and 500 μL o 1 × PBS bu e we e cen-
i uged a 13,000 pm o 1 minu e o pelle each sample, and he supe na an
disca ded. 550 μL o bu e CF (Mache ey-Nagel Nucleospin Food Ki ) was hen
added o he DNA samples, and 500 μL o Lysis Bu e (Maxwell RSC simply
RNA Tissue Ki (P omega)) was added o he RNA samples. Bead bea ing was
pe o med using a TissueLyse (Qiagen), comp ising wo ounds o bead bea -
ing a 30 Hz o 90 seconds.
DNA ex ac ions and ampli ica ion
DNA ex ac ions o honey, pollen, and bee samples we e pe o med using
he Mache ey-Nagel Nucleospin Food Ki ollowing he p o ocols desc ibed
abo e, ollowing he manu ac u e ’s ins uc ions. Samples we e le o si a
oom empe a u e o 1 hou in a inal elu ion o 100 μL o Bu e CE be o e he
las cen i uga ion s ep. Swab samples we e ex ac ed using a QIAamp DNA
In es iga o Ki (Qiagen) ollowing he p o ocol o isola ion o o al DNA om
su ace and buccal swabs. The p ocedu e was ollowed o he use o co on
o Dac on swabs, wi h he op ional ca ie RNA and QIA sh edde s eps includ-
ed. Samples we e le o si a oom empe a u e o 1 hou in a inal elu ion o
100 μL o Bu e ATE be o e he las cen i uga ion s ep.
475
Me aba coding and Me agenomics 9: 469–499 (2025), DOI: 10.3897/mbmg.9.165436
E in Hill e al.: Su ace swab eco e y o eDNA and eRNA
Nega i e ex ac ion con ols we e un h oughou he DNA ex ac ion p o-
cess o all sample ypes. All samples we e quan i ied using a Nanod op 8000
Spec opho ome e (The mo Scien i ic) and hen no malised o 5 ng/μL o
PCR ampli ica ion o ensu e di e ences in OTU composi ion we e no a ec ed
by s a ing concen a ions o samples. Any samples wi h an ini ial concen a-
ion < 5 ng/μL we e kep a a nea concen a ion.
Samples we e ampli ied using ou uni e sal p ime pai s a ge ing bac e i-
al (16S), animal (COI), ungal (ITS1), and plan (ITS2) species (Table 2). PCRs
we e ca ied ou in 15 μL eac ions, con aining 7.5 μL o GoTaq G een Mas e
Mix (P omega), 0.6 μL each o o wa d and e e se p ime s (10 uM), 4.3 μL o
ul apu e H2O, and 2 μL o DNA. PCR ampli ica ion comp ised an ini ial dena-
u a ion a 95 °C o 5 minu es, ollowed by 30 cycles o 95 °C o 30 seconds,
50 °C o 30 seconds, 72 °C a 50 seconds, and a inal ex ension o 72 °C o
5 minu es. The cycle numbe was inc eased o 35 cycles o he ampli ica ion
o he ungal ITS1 and animal COI p ime s. Animal COI eac ions included a
PNA clamp designed o educe Apis melli e a ampli ica ion (Suppl. ma e i-
al 1). Modi ica ions o he p o ocol abo e included he addi ion o 0.75 μL o
PNA clamp (20 uM) o he eac ion and educ ion o ul apu e H2O o 3.55 μL,
he use o GoTaq En i o qPCR Sys em (P omega) in lieu o he Taq desc ibed
abo e o imp o e PCR ampli ica ion, and he ini ial dena u a ion s ep changed
o 2 minu es as ecommended o he GoTaq En i o qPCR Sys em. Nega i e
PCR con ols we e un h oughou he ampli ica ion p ocess, and a subsample
o PCR p oduc s we e isualised using a 2% aga ose gel o con i m success ul
ampli ica ion. Th ee eplica e PCRs we e comple ed o each ma ke and hen
pooled a equal amoun s o a inal olume o 30 μL. PCR p oduc s we e hen
quan i ied using a Fluo oskan (The mo Fishe Scien i ic) and a Quan -iT dsDNA
High Sensi i i y Assay Ki (The mo Fishe Scien i ic), and all samples/ma ke s
no malised o 10 ng/μL. Samples we e hen pooled a equal amoun s ac oss
he ou ma ke s, and sen o he Biomolecula Resou ce Facili y in Canbe a,
Aus alia, o indexing and Illumina MiSeq sequencing using a 3 600 cycle ki .
RNA ex ac ions
RNA ex ac ions o all sample ypes we e pe o med using he Maxwell RSC
simplyRNA Tissue Ki (P omega). Fo swab samples, swab ips we e added o a
1.5 mL Eppendo ube con aining 500 μL o Lysis Bu e (P omega) and placed
Table 2. Me aba coding p ime s used o ampli ica ion in his s udy. PNA clamp sequence de eloped o his s udy o
educe Apis melli e a ampli ica ion is lis ed wi h he animal (COI) p ime s.
Region P ime Name Sequence (5’-3’) Ta ge Taxa Re e ence
16S 515-FY GTGYCAGCMGCCGCGGTAA Bac e ia (App ill e al. 2015; Pa ada e al. 2016)
806R GGACTACNVGGGTWTCTAAT
COI FwhF2 GGDACWGGWTGAACWGTWTAYCCHCC Animal (Vamos e al. 2017)
FwhR2n GTRATWGCHCCDGCTARWACWGG
Amel_PNA CATTCTTCACCTTCAGTAGA
ITS1 ITS1 CTTGGTCATTTAGAGGAAGTAA Fungi (Whi e e al. 1990; Ga des and B uns 1993)
ITS2 GCTGCGTTCTTCATCGATGC
ITS2 S2F ATGCGATACTTGGTGTGAAT Plan (Chen e al. 2010)
4 e TCCTCCGCTTATTGATATGC
476
Me aba coding and Me agenomics 9: 469–499 (2025), DOI: 10.3897/mbmg.9.165436
E in Hill e al.: Su ace swab eco e y o eDNA and eRNA
on a he momixe se o shaking a 900 pm o 1 hou a oom empe a u e.
Following bead bea ing o he honey, pollen, and bee samples and shaking o
swab samples, 350 μL o lysa e was collec ed and added in o a 1.5 mL Eppendo
ube con aining 350 μL o Lysis Bu e . Samples we e hen p epa ed o use on
a Maxwell RSC Ins umen , o au oma ed RNA ex ac ion. Two ex ac ion ep-
lica es we e unde aken o each hi e sample, and nega i e ex ac ion con ols
we e un h oughou he RNA ex ac ion p ocess. Ex ac ion eplica es we e
hen pooled by sample ype and hi e and quan i ied using a 4200 TapeS a ion
Sys em (Agilen ). Samples we e hen sen o he Biomolecula Resou ce Facili y
(BRF) in Canbe a, Aus alia, o lib a y p epa a ion wi h he Illumina S anded
mRNA P ep ki and sequencing using a Nex Seq 2000 P2 2 × 150 bp un.
Me aba coding analysis
The demul iplexed pai ed FASTQ eads we e quali y immed o Q20 using
bbduk (Bushnell 2014) and me ged using USEARCH. Gene-speci ic p ime se-
quences we e used o spli eads in o ma ke s using cu adap (Ma in 2011)
and low-quali y eads we e il e ed ou using he USEARCH as q il e unc ion.
Fo each ma ke , sequences we e denoised and ze o- adius ope a ional axo-
nomic uni s (ZOTUs) we e chosen using he USEARCH unoise3 unc ion using
a minimum clus e size o 8 eads. Fo bac e ial 16S, we con inued o assign
axonomy o ZOTUs. Fo he o he h ee ma ke s, we clus e ed in 97% iden i y
OTUs using USEARCH clus e _ as unc ion.
To assign axonomy o p edic ed ZOTU/OTUs, we used he USEARCH sin-
ax unc ion wi h an 90% boo s ap cu o agains he co esponding ma k-
e da abases. The sin ax command uses he kme -based SINTAX algo i hm
(Edga 2016) o make axonomic p edic ions. Fo animal COI, we downloaded
he GB255 e sion o he SINTAX- o ma ed MIDORI2 da abase (Le ay e al.
2022) om he MIDORI websi e (www. e e ence-mido i.in o). Fo plan ITS2,
we downloaded ull S ep ophy ina sequences om he ITS2DB (Selig e al.
2007; Koe schan e al. 2010, 2012; Me ge e al. 2012; Ankenb and e al. 2015)
websi e (h p://i s2.bioapps.biozen um.uni-wue zbu g.de) on Decembe
2023. We used axonki (h ps://bioin .shenwei.me/ axonki ) o add axonom-
ic lineage in o ma ion and e o ma he FASTA heade o SINTAX o ma . Fo
ungal ITS1, we downloaded he SINTAX- o ma ed ile u ax_ e e ence_da a-
se _25.07.2023. as a.gz om he UNITE 9.0 (Aba enko e al. 2024) da abase
websi e (h ps://uni e.u .ee/ eposi o y.php). Fo bac e ial 16S, we download-
ed he RDP aining se 19 om he RDP websi e (h ps://sou ce o ge.ne /
p ojec s/ dp-classi ie / iles/RDP_Classi ie _T ainingDa a) and e o ma ed
i o SINTAX o ma using a cus om py hon sc ip . We gene a ed sea chable
da abases o all ma ke s wi h he USEARCH 11.0.667_i86linux32 (Edga
2010) makeudb p og am. Seconda y iden i ica ions we e assigned using
blas n agains he ull NCBI n da abase, e u ning he op en ma ches. The
op BLAST esul s we e pa sed using he py hon sc ip axonomy_assignmen _
BLAST_V2.py (gi hub: Joseph7e/Assign-Taxonomy-wi h-BLAST), which as-
signs he consensus lowes axonomic le el possible om all hi s ha ma ch
he p o ided iden i y pe cen cu -o s and e- alue. We assigned he consensus
BLAST esul s wi h a minimum o 97% iden i y and e- alue o 1e-10 o missing
axonomic assignmen s no ound in ma ke -speci ic da abases.
477
Me aba coding and Me agenomics 9: 469–499 (2025), DOI: 10.3897/mbmg.9.165436
E in Hill e al.: Su ace swab eco e y o eDNA and eRNA
Once ZOTU/OTUs we e iden i ied, we used he R package mic oDecon
(McKnigh e al. 2019) o educe con aminan eads o emo e ZOTU/OTUs
ha we e abundan in nega i e con ols. Addi ionally, we excluded COI iden-
i ica ions ha we e assigned o phyla Ascomyco a and Basidiomyco a, 16S
iden i ica ions ha we e no bac e ia o we e assigned o chlo oplas o mi o-
chond ia, and ITS2 iden i ica ions ha we e no assigned o phyla S ep ophy a
(ITS2DB) o Vi idiplan ae (NCBI). Finally, we used he a e y unc ion o he R
package egan (Oksanen e al. 2019) o gene a e a e ac ion cu es and iden-
i y samples wi h 100 o ewe eads and excluded hem om u he analysis.
F om he il e ed da ase , we ca ied ou he ollowing compa isons: (a)
axonomic co e age ac oss sample ypes; and (b) e ec o eplica es (sam-
ple and DNA ex ac ion) on axonomic co e age. To iden i y axonomic co -
e age ac oss sample ypes, he numbe o obse ed axa was plo ed o
each sample ype and me aba coding ma ke . To keep sample eplica ion
consis en ac oss all sample ypes, only he i s wo swab eplica es we e
included in his analysis. Any signi ican di e ences be ween numbe o axa
de ec ed ac oss sample ypes we e no ed o each ma ke using - es s,
pe o med using he R package s a ix 0.7.2 (Kassamba a 2019). The op
10 gene a p esen in he mos samples ac oss all sample ypes we e hen
iden i ied o each axonomic g oup and plo ed o iden i y consis ency o
common gene a ac oss sample ypes. NMDS was calcula ed using egan o
each ma ke o iden i y i axonomic composi ions di e ed be ween sample
ypes. Analysis o mul i a ia e homogenei y o g oup dispe sions ( a iances)
was calcula ed om NMDS esul s using he be adispe unc ion in egan,
and signi ican dispe sion di e ences iden i ied using ANOVA. Shannon di-
e si y indices we e also calcula ed o es ima e he ichness and e enness
o sampling o each ma ke and sample ype using he es ima e_ ichness
unc ion o he R phyloseq package (McMu die and Holmes 2013). Species
accumula ion cu es we e calcula ed o each ma ke and sample ype us-
ing he specaccum unc ion in egan o iden i y he sampling e o neces-
sa y o app op ia e axonomic co e age o each sample ype. We an a
gene alised linea model wi h gamma dis ibu ion o explo e he e ec s o
a iables including indi idual hi es, sample ype, collec ion da e, and com-
bina ions o hese, on he di e si y o axa de ec ed. Analysis o de iance
was hen calcula ed using he Ano a unc ion o he R ca package (Fox and
Weisbe g 2018) ac oss all sample ypes o bac e ia (16S), ungi (ITS1), and
plan s (ITS2), o iden i y a iables ha may ha e signi ican ly in luenced he
di e si y o axa de ec ed using he Shannon di e si y index.
To assess he e ec s o eplica es on axonomic eco e y, he pe cen age
o e lap o axa de ec ed ac oss ex ac ion eplica es was calcula ed o iden i y
how simila each eplica e was in axonomic composi ion. Fo his analysis,
only he i s wo swab eplica es we e e ained o p o ide consis ency when
compa ing pe cen age o e lap wi h he o he sample ypes. Following his,
mo e in-dep h analysis o all h ee swab eplica es was pe o med. The num-
be o unique axa obse ed in each swab eplica e was iden i ied o con i m
i subsequen swabbing esul ed in ewe de ec ions o unique axa. This was
expec ed as swab eplica es sampled he same su aces in he hi es and we e
p edic ed o emo e DNA du ing he swabbing p ocess, lea ing less deposi ed
DNA a ailable o sampling in subsequen swabbing eplica es.
484
Me aba coding and Me agenomics 9: 469–499 (2025), DOI: 10.3897/mbmg.9.165436
E in Hill e al.: Su ace swab eco e y o eDNA and eRNA
echnique on he hi d eplica e, and no new axa appea ing om a second hi e
deb is swab eplica e. In he animal (COI) da ase , all ou axa we e de ec ed
om i s swabs o he hi e en ance and hi e deb is. The ungi (ITS1) swab-
bing eplica es con inued o de ec addi ional unique axa ac oss each epli-
ca e o bo h he hi e en ance and he hi e deb is, wi h 31 and 19 addi ional
axa de ec ed in eplica es wo and h ee o he hi e en ance swabs, and 42
and 17 addi ional axa in eplica es wo and h ee o he hi e deb is samples. In
he plan (ITS2) da ase , eplica es one and wo de ec ed mos o he obse ed
axa o bo h he hi e en ance swabs and he hi e deb is swabs, wi h se en ad-
di ional unique axa de ec ed in he second swab eplica e o he hi e en ance
68.8 69.8 61.7 39.2 43.3 83.8 86.5
0
25
50
75
Hi
e Deb is Swabs
Hi e En ance Swabs
Honey
In e nal Pollen
Pollen T ap
Re u n Fo age Bees
Top Box Bees
Pe cen o e lap (%)
Bac e ia (16S)
67.6 63.9 53.7 033.381.
58
1.5
0
25
50
75
100
Hi e Deb is Swabs
Hi e En ance Swabs
Honey
In e nal Pollen
Pollen T ap
Re u n Fo age Bees
Top Box Bees
Pe cen o e lap (%)
Animal (COI)
32.9 40.6 29.5 28.3 27.8 25.5 19.1
0
10
20
30
40
Hi
e Deb is Swabs
Hi
e En ance Swabs
Honey
In e nal Pollen
Pollen T ap
Re u n Fo age Bees
Top Box Bees
Pe cen o e lap (%)
Fungi (ITS1)
61.9 58 65.3 77.9 70.9 61.8 65.6
0
25
50
75
Hi e Deb is Swabs
Hi
e En ance Swabs
Honey
In e nal Pollen
Pollen T ap
Re u n Fo age Bees
Top Box Bees
Pe cen o e lap (%)
Plan s (ITS2)
Figu e 4. Pe cen age o e lap o axa de ec ed ac oss DNA collec ion and ex ac ion eplica es o each o he sample
ypes sequenced o bac e ia (16S), animal (COI), ungi (ITS1) and plan (ITS2) me aba coding ma ke s. Mean o e lap
pe cen ages a e shown below each boxplo . Fo hi e en ance and deb is swabs only he i s wo eplica es we e com-
pa ed o consis ency wi h he o he sample ypes.
485
Me aba coding and Me agenomics 9: 469–499 (2025), DOI: 10.3897/mbmg.9.165436
E in Hill e al.: Su ace swab eco e y o eDNA and eRNA
samples and 11 unique axa de ec ed in he second swab eplica e o he hi e
deb is samples. Only one unique axon was de ec ed in he hi d swab eplica e
o he hi e deb is samples in his da ase , and no addi ional axa in he hi d
swab eplica e o he hi e en ance swab samples.
RNA sequencing
RNA sequencing esul ed in 872 million aw eads e u ned ac oss 13 sam-
ples. Raw assignmen s showed ha mos eads we e assigned o euka yo e
species in he hi e en ance swabs, in e nal pollen, and op box bee sam-
ples, including Apis melli e a, pa icula ly in he op box bee samples (Fig. 6).
0
10
20
30
Replica e 1 Replica e 2 Replica e 3
Numbe o new axa de ec ed
Bac e ia (16S)
0
1
2
3
4
Replica e 1 Replica e 2 Replica e 3
Numbe o new axa de ec ed
Animal (COI)
50
100
150
Replica e 1 Replica e 2 Replica e 3
Numbe o new axa de ec ed
Fungi (ITS1)
0
10
20
30
40
50
Replica e 1 Replica e 2 Replica e 3
Numbe o new axa de ec ed
Plan s (ITS2)
Hi e Deb is Swabs Hi e En ance Swabs
Figu e 5. S ep plo s showing he numbe o axa de ec ed by he i s swab eplica es aken om he hi e en ance and
hi e deb is, and he numbe o addi ional unique axa de ec ed by pe o ming addi ional swab eplica es o bac e ia
(16S), animal (COI), ungi (ITS1) and plan (ITS2) me aba coding ma ke s.
486
Me aba coding and Me agenomics 9: 469–499 (2025), DOI: 10.3897/mbmg.9.165436
E in Hill e al.: Su ace swab eco e y o eDNA and eRNA
The op 10 euka yo e species (excluding A. melli e a) assigned in his g oup
we e p edominan ly plan species, wi h a single ungal species (Au eobasidium
melanogenum) de ec ed in he hi e en ance swab samples (Suppl. ma e ial 2:
ig. S3). In he honey samples, he highes p opo ion o eads assigned we e o
la ge subuni ibosomal ibonucleic acid (LSU RNA). The p opo ion o all aw
assigned eads ha we e iden i ied as i uses anged om 0.005% in he op
box bees and honey samples, 0.016% in he hi e en ance swabs, and 0.056%
in he in e nal pollen samples.
A e quali y con ol and il e ing s eps, 67 i uses om a combined o al
o 17 million eads assigned o con igs we e iden i ied (Suppl. ma e ial 2: a-
ble S4). To al numbe o eads assigned o con igs was highes in he in e -
nal pollen samples, ollowed by he hi e en ance swabs, op box bees, and
he honey samples. A e age con ig leng hs we e highes in he in e nal pol-
len samples (3865.82 ± 3226.87 SD), ollowed by he op box bee samples
(3544.36 ± 2925.95) and he hi e en ance swabs (2833.81 ± 2743.46), and
lowes in he honey samples (1298.83 ± 917.29 SD).
Honey samples had he highes obse ed numbe o i uses ac oss all
sample ypes, wi h an a e age numbe o 23 i uses pe eplica e (Fig. 7).
0.00
0.25
0.50
0.75
1.00
Hi
e En ance Swabs
Honey
In e nal Pollen
Top Box Bees
P opo ion o eads
Bac e ia
CO1
O he Euka yo es
Honeybee
Low quali y eads
RNA
Unassigned eads
Vi uses
Read assignmen s pe sample ype
Figu e 6. RNA aw ead ou pu s and he p opo ions assigned o di e en axonomic g oups o each sample ype.
487
Me aba coding and Me agenomics 9: 469–499 (2025), DOI: 10.3897/mbmg.9.165436
E in Hill e al.: Su ace swab eco e y o eDNA and eRNA
This was ollowed by hi e en ance swab and op box bee samples, wi h a e ages
o 16and 14 i uses obse ed, espec i ely. In e nal pollen samples con ained he
lowes obse ed numbe o i uses o any sample ype wi h an a e age o 7.73.
Assigning de ec ed i uses o expec ed hos species, all ou sample ypes
we e able o de ec ungal and plan i uses, albei a di e en a es (Fig. 8).
Honey samples comp ised he highes numbe o plan i uses wi h 23 plan
i uses de ec ed, ollowed by hi e en ance swabs and op box bees (10 i uses
each), and in e nal pollen samples (se en i uses). Fungal i uses we e mos
abundan in swab samples (10 i uses), wi h wo ungal i uses each in he
bee and honey samples, and one ungal i us in he pollen samples. Honeybee-
speci ic i uses we e mos abundan in he honeybee samples (se en i uses),
ollowed by honey and hi e en ance swab samples (six and h ee i uses, e-
spec i ely). O he insec i uses we e also de ec ed in hese sample ypes. No
insec i uses we e p esen in he in e nal pollen samples.
Fi e o he op 10 i uses de ec ed ac oss all sample ypes we e associa ed
wi h he plan genus Camellia (Fig. 9). These we e all de ec ed in each sample
ype, wi h he p opo ion ac oss samples much highe in he in e nal pollen sam-
ples han he o he sample ypes. Hi e en ance swabs, honey, and op box bees
showed posi i e de ec ions o he emaining i e mos abundan i uses; hese in-
cluded wo ungi-associa ed i uses (Lep osphae ia biglobosa be a lexi i us 3 and
Bo yosphae ia do hidea na na i us 1), wo i uses associa ed wi h clo e s (whi e
clo e mosaic i us and clo e yellow mosaic i us), and he honeybee i us black
queen cell i us (T ia o i us nige eginacellulae). Howe e , o hese i e i uses only
one o he ungi i uses was de ec ed in he in e nal pollen samples, esul ing in
only six o he op 10 i uses being de ec ed in he in e nal pollen samples.
Shannon di e si y es ima es indica e ha honey samples comp ise he high-
es i us alpha di e si y measu es, ollowed by bee samples (Suppl. ma e ial
2: ig. S4). All hi e en ance swab and pollen samples comp ise i us alpha
di e si y alues below 1.
16 23 7.33 14
0
10
20
30
Hi
e En ance Swabs
Honey
In e nal Pollen
Top Box Bees
Numbe o axa de ec ed
Di e ences in subs a e o RNA da a
Figu e 7. Numbe o axa de ec ed ac oss each sample ype o RNA i uses. Mean alues o each subs a e a e shown
below each boxplo .
488
Me aba coding and Me agenomics 9: 469–499 (2025), DOI: 10.3897/mbmg.9.165436
E in Hill e al.: Su ace swab eco e y o eDNA and eRNA
Discussion
Ou s udy ep esen s he mos comp ehensi e compa ison o hi e eNA sample
ypes o da e. Expanding on he wo k o Boa dman e al. (2024) and Čukano á
e al. (2023) we examined all majo axonomic g oups p esen in hi es (bac e ia,
0
10
20
30
40
Hi
e En ance Swabs
Honey
In e nal Pollen
Top Box Bees
To al
Hos
ungi
honeybee
o he insec
plan
Vi uses de ec ed in RNA da a by hos ype
Figu e 8. Numbe o i uses de ec ed in di e en hos ca ego ies ac oss he ou sample ypes sequenced.
0.00
0.25
0.50
0.75
1.00
Hi
e En ance Swabs
Honey
In e nal Pollen
Top Box Bees
Rela i e p opo ion o samples
Species
Camellia japonica associa ed be a lexi i us
3
Camellia ingspo associa ed i us 3
Camellia japonica associa ed be a lexi i us
1
Camellia japonica associa ed be a lexi i us
2
Camellia ingspo associa ed i us 2
Lep osphae ia biglobosa be a lexi i us 3
T ia o i us nige eginacellulae
Whi e clo e mosaic i us
Clo e yellow mosaic i us
Bo yosphae ia do hidea na na i us 1
O he
Top species de ec ed in RNA da a
Figu e 9. Top 10 i uses de ec ed ac oss all sample ypes and hei de ec ion p opo ions.
489
Me aba coding and Me agenomics 9: 469–499 (2025), DOI: 10.3897/mbmg.9.165436
E in Hill e al.: Su ace swab eco e y o eDNA and eRNA
animal, ungi, plan s, i uses) o iden i y he mos sui able sample ypes o hon-
eybee pes and pa hogen de ec ion, ege a ion moni o ing and plan pa hogen
su eillance. We also d aw on commonly u ilised hi e sample ypes (honey, pol-
len, bees) o p o ide a obus compa ison o a non-in asi e swabbing me hod
o samples equen ly used in he li e a u e. Ou esul s demons a e ha col-
lec ing swab samples om honeybee hi e en ances is an e ec i e me hod o
de ec ing b oad axonomic assemblages o bac e ia, animal, ungi, plan s, and
i uses. Fu he , we showed ha he mos common axa and o e all axonomic
composi ion a e highly simila ac oss sample ypes, indica ing ha swabs can
e eal mo e di e si y and he axa iden i ied a e consis en wi h mo e in usi e
sampling collec ion me hods. F om hese esul s, we ecommend hi e en ance
swabbing as he p e e ed me hod o sample collec ion om honeybee hi es
o me aba coding s udies, and a sui able me hod o RNA collec ion whe e
he goal is b oad-scale moni o ing o i al di e si y in he landscape. This wo k
p o ides an imp o ed honeybee hi e sampling p o ocol ha can be implemen -
ed o a ange o biosecu i y su eillance ac i i ies, including moni o ing o
honeybee ela ed pes s and pa hogens, as well as plan pa hogens and weeds.
The e a e clea bene i s o hi e en ance swabbing, being a simple me h-
od o non-specialis s and enabling non-in asi e yea - ound sample collec ion
wi hou impac ing hi es. The compa able and o en be e eco e y o eNA
om hi e en ance swabs o e mo e di ec sample ypes was somewha su -
p ising. This e ec i eness likely s ems om he accumula ion and pe sis ence
o eNA ac oss he hi e en ance and base. Hi e deb is ypically accumula es a
he base o he hi e as a esul o ou ine colony ac i i ies such as b ood ea -
ing and ood s o age. Addi ionally, high bee a ic h ough he hi e en ance
acili a es pollen, mic obial spo es and o he biological ma e ial being depos-
i ed om e u ning o age s and passi e ai low can also d aw eNA om he
su ounding a ea in o hi es. O he s udies ha e also ecognised he bene i s
o hi e en ance swabbing o eDNA de ec ion. Boa dman e al. (2024) anked
hi e en ance swabbing as one o hei easy o collec sample me hods ha is
ideal o la ge-scale moni o ing, echoing ou expe ience and indings p esen ed
he e. Thei compa ison o swab and sp ay/wash me hods om a ious hi e en-
i onmen sou ces ound simila o e all de ec ion o a h opod and mic obial
axa, al hough ma ginally imp o ed species ichness wi h sp ay/wash su ace
agg ega ion me hods. Hi e en ance swabbing was also ecen ly demons a ed
as an e ec i e ool o a ge ed eDNA de ec ion o honeybee pes s and pa ho-
gens (Mackay e al. 2025; Robe s e al. 2025). While hi e swabbing o e s clea
logis ical ad an ages, i may no ully eplace di ec sampling me hods o spe-
ci ic moni o ing objec i es, such as quan i ying pa hogen load o in es iga ing
hi e o aging p e e ences. Addi ionally, ac o s such as hi e design, clima e,
and seasonal a ia ion a e likely o in luence eNA accumula ion and de ec ion
sensi i i y and wa an u he in es iga ion.
The e we e no able di e ences in he pe o mance o he mo e common
hi e sample ypes (bees, pollen and honey) ac oss axonomic g oups. Honey
samples consis en ly showed high axonomic di e si y, sugges ing hey a e
e ec i e o b oad su eillance, simila o hi e en ance swabs. The high
suga con en o honey helps p ese e eNA and has been used o de ec -
ing plan and mic obial DNA e en a e se e al decades (Ci will and Wi a
2025). Howe e , he honey samples also had low o al ead coun s ac oss axa
490
Me aba coding and Me agenomics 9: 469–499 (2025), DOI: 10.3897/mbmg.9.165436
E in Hill e al.: Su ace swab eco e y o eDNA and eRNA
g oups, likely due o PCR inhibi o s and lowe eNA concen a ions ypical o
his ma ix (Soa es e al. 2023). The ela i ely poo axonomic eco e y om
he pollen samples, pa icula ly o plan s (ITS2), was unexpec ed. This may
e lec he dynamic o aging beha iou o bees, which a ies wi h he colo-
ny’s nu i ional needs (Co ey and B een 1997; De Ve e e al. 2017). Pollen and
honey a e also consumed by hi es a di e en a es, wi h pollen ypically used
mo e quickly han honey s o es (Ande son e al. 2014; Eye e al. 2016; Ca oll
e al. 2017; Roessink and Van De S een 2021). As a esul , honey likely p o-
ides a cumula i e eco d o he local en i onmen , while pollen samples gi e
a con empo a y iew o ecen lowe ing ac i i y. Bee samples also showed
lowe species di e si y, despi e o aging bees being he p ima y mechanism
o eNA en e ing he hi e. Indi idual bees ca y only a ac ion o he o al eNA,
mos ly hei ood sou ces and associa ed mic obes, while su ace-bound eNA
is likely mo e ansien . Bee samples also ep esen a na owe ime window
compa ed o honey and hi e deb is, which accumula e eNA o e longe pe-
iods. These di e ences highligh how each sample ype cap u es a dis inc
empo al and ecological snapsho o he hi e en i onmen , ein o cing he al-
ue o mul i-sample app oaches o comp ehensi e su eillance.
eRNA is inc easingly used o dis inguishing con empo a y species p esence
and is essen ial o e ec i e i us su eillance. Because RNA is mo e suscep-
ible o deg ada ion, di ec sample ypes such as bees o pollen a e gene ally
expec ed o p o ide mo e eliable i us de ec ion. Howe e , ou s udy ound no
op imal sample ype o i us su eillance. Su p isingly, hi e en ance swabs
we e compa able wi h honey and bee samples in cap u ing a b oad snapsho
o i al di e si y, hough hey eco e ed ewe bee-speci ic i uses. Pas in es-
iga ions ha e also epo ed poo i us de ec ion om hi e en ance swabs
(Čukano á e al. 2023). One possible explana ion is ha honeybees ypically
exi he hi e o ‘cleansing ligh s’ o de eca e, po en ially educing i us depo-
si ion on he hi e base. Bee i uses may also deg ade mo e apidly in he en-
i onmen han plan i uses, pa icula ly hose ansmi ed ia pollen (Ca d e
al. 2007). Honey samples p o ided he mos comp ehensi e i us de ec ion in
e ms o di e si y, sugges ing hei sui abili y o holis ic plan -pollina o i us
su eillance. Howe e , honey also yielded he lowes i al ead coun s and con-
ig leng hs, which may limi de ec ion con idence. In con as , bee and pollen
samples p oduced highe ead dep hs and con ig leng hs bu de ec ed ewe
i uses o e all. The lack o honeybee and o he insec i uses de ec ed in he
pollen samples was also su p ising. The eco e y o plan i uses was compa-
able be ween sample ypes, sugges ing he e was no echnical issue wi h he
pollen RNA ex ac ions. Mul iple o he s udies also epo success in de ec ing
honeybee i uses using pollen samples (Pe ei a e al. 2019; Balkanska e al.
2023; Čukano á e al. 2023; Lee e al. 2023; Robe s e al. 2023; As and E oglu
2025). I is likely ou esul is due o he small numbe o hi es sampled and a
gene al low p e alence o honeybee i uses in hese hi es a he ime o sam-
pling. These ade-o s be ween he numbe o i uses de ec ed (quan i y) and
he dep h and comple eness o i al sequences (quali y) a e a key di e ence
om DNA me aba coding and emphasise ha he mos sui able sample ype
depends on he speci ic i us su eillance goal.
Op imising sampling e o is c ucial o maximising axonomic eco e y
in eNA s udies, and ou esul s p o ide speci ic guidance o honeybee hi e
491
Me aba coding and Me agenomics 9: 469–499 (2025), DOI: 10.3897/mbmg.9.165436
E in Hill e al.: Su ace swab eco e y o eDNA and eRNA
moni o ing. Replica ion o swab samples in ou s udy showed diminishing
e u ns in he mean numbe o axa and numbe o unique axa by he hi d
swab eplica e, excep o he ungal da ase . The e o e, we sugges ha
collec ing wo swab eplica es pe hi e may be su icien o cap u e he di-
e si y o axa om hi e eDNA me aba coding. Fo all he sample ypes and
axonomic g oups we assessed ia me aba coding, addi ional sample and
DNA ex ac ion eplica es led o inc eases in axa eco e y. Howe e , he ex-
en o eplica ion necessa y o cap u e he ull b ead h o axa p esen in a
sample o si e will be dependen on bo h he s udy sys em being in es iga -
ed and he axonomic g oups being ampli ied (Fice ola e al. 2015; Shi azi
e al. 2021). Based on species accumula ion cu es, eplica ion in ou s udy
was su icien o he bac e ial and animal communi ies, bu u he eplica-
ion would be needed o eco e addi ional di e si y o he ungal and plan
communi ies in ou s udy hi es. G ea e sequencing dep hs will ypically
inc ease he de ec ion o a e axa in samples (Smi h and Peay 2014; Doble
e al. 2020; Shi azi e al. 2021), al hough ou da a sugges s ha his will s ill
be mo e in luenced by sample ype. Using mul iple p ime pai s o a ge
axa can also ci cum en possible biases and p ime misma ches du ing
PCR ampli ica ion (F eeland 2017; Xiong e al. 2022). Ul ima ely, he le el
o eplica ion necessa y o axonomic eco e y will be dependen on s udy
ques ions, whe e aims o de ec ing a e axa will equi e mo e eplica ion
han s udies in es iga ing b oad-scale biodi e si y. In honeybee hi e su -
eillance, sampling om addi ional hi es is likely o be mo e bene icial han
collec ing mo e han wo swab eplica es pe hi e and educes an agonising
a hi e due o inc eased handling imes.
Fu u e esea ch ocusing on he ecology o eNA in ela ion o pe sis ence
wi hin hi es, a iabili y ac oss hi es, and limi s o de ec ion is necessa y o
u he imp o e he eliabili y and applica ion o hese biomoni o ing me h-
ods. This is especially impo an in a biosecu i y con ex whe e ea ly de ec-
ion o new incu sions is c i ical and con idence a ound alse nega i e and
posi i e de ec ions mus be es ablished. Fu he compa ison o hi e eNA
de ec ion wi h di ec su eys would imp o e ou unde s anding o how well
hi e axonomic di e si y p o ides a obus ep esen a ion o he su ounding
en i onmen . Inclusion o eNA om ex e nal samples, such as om soil, wa-
e o ai , o e a ange o dis ances om hi es would help de e mine some o
he p ac ical limi s o his app oach. Hi e-based de ec ion is also inhe en ly
biased by bee o aging beha iou , which is in luenced by lo al a ailabili y,
seasonali y and landscape composi ion. Conside a ion o hese ac o s will
help e ine when and whe e hi e eNA su eillance is mos e ec i e. Despi e
hese limi a ions in ou unde s anding o hi e eNA, he b ead h o axa de ec -
able in bee hi es is imp essi e and highligh s a deepe complexi y o how
bees in e ac wi h hei en i onmen . One o he key s eng hs o eNA su -
eillance is he abili y o conduc la ge-scale sampling e o s ac oss di e se
en i onmen s. Engaging wi h beekeepe s o cap u e hi e eNA ac oss u ban
and ag icul u al landscapes is an exci ing oppo uni y o suppo he heal h o
hi es and hei o aging en i onmen s. Le e aging hi e ne wo ks and simple
swab me hods o e s a scalable, low-impac app oach o landscape-le el bi-
osecu i y and biodi e si y moni o ing, one ha can be s eng hened h ough
collabo a ion wi h beekeepe s and ci izen scien is s.
492
Me aba coding and Me agenomics 9: 469–499 (2025), DOI: 10.3897/mbmg.9.165436
E in Hill e al.: Su ace swab eco e y o eDNA and eRNA
Conclusions
This s udy demons a es ha non-in asi e swab sampling om hi e en-
ances is an e ec i e eNA me hod o de ec ing he b oad di e si y o axa
wi hin honeybee hi es. Hi e en ance swabs we e compa able wi h com-
mon hi e eNA sampling me hods, consis en ly eco e ing simila axonomic
p o iles while also de ec ing addi ional axa, pa icula ly o bac e ial and
ungal g oups. Sampling eplica ion imp o ed he numbe o axa de ec ed,
wi h wo swab eplica es pe hi e ecommended o u u e s udies. Hi e en-
ance swabbing also p o ides p ac ical ad an ages in being simple, non-in-
asi e, sui able o yea - ound sampling and can be conduc ed by non-spe-
cialis s. These a ibu es, combined wi h b oad axonomic co e age, make
hi e en ance swabbing a powe ul ool o la ge-scale biosecu i y su eil-
lance and biodi e si y moni o ing. U ilising es ablished hi e ne wo ks wi h
his app oach o e s a highly scalable s a egy o he de ec ion o honey-
bee pes s and pa hogens, plan pa hogens, and he su ounding ege a ion
ac oss di e se landscapes.
Acknowledgemen s
The au ho s wish o hank Pe e Jones, Biko Mui a, Alicia G ealy and Ha y Eyck
o hei assis ance wi h sample collec ion and labo a o y p ocessing o sam-
ples. We a e g a e ul o Flo ence B a o and Amy Pa en o hei discussions
ega ding DNA and RNA ex ac ion me hods. This p ojec was unded by he
En i onomics Fu u e Science Pla o ms o he Commonweal h Scien i ic and
Indus ial Resea ch O ganiza ion (CSIRO). We would like o acknowledge he
con ibu ion o Biopla o ms Aus alia in he gene a ion o da a used in his
publica ion. Biopla o ms Aus alia is enabled by NCRIS.
Addi ional in o ma ion
Con lic o in e es
The au ho s ha e decla ed ha no compe ing in e es s exis .
E hical s a emen
No e hical s a emen was epo ed.
Use o AI
No use o AI was epo ed.
Funding
No unding was epo ed.
Au ho con ibu ions
Concep ualisa ion: EH, LM, FEV, BG, MH, JR. Da a cu a ion: EH, LM, JO. Fo mal anal-
ysis: EH, LM, JO. Funding acquisi ion: JR. In es iga ion: EH, LM, JR. Me hodology:
EH, LM, FEV, JO, JR. So wa e: LM, JO. Supe ision: FEV, BG, MH, JR. Visualisa ion:
EH, LM, JO. W i ing – o iginal d a : EH, LM, FEV, JO, JR. W i ing – e iew and edi ing:
EH, LM, FEV, JO, BG, MH, JR.
493
Me aba coding and Me agenomics 9: 469–499 (2025), DOI: 10.3897/mbmg.9.165436
E in Hill e al.: Su ace swab eco e y o eDNA and eRNA
Au ho ORCIDs
E in Hill h ps://o cid.o g/0000-0002-7642-696X
Liz Milla h ps://o cid.o g/0000-0002-6139-4336
F ancisco Encinas-Viso h ps://o cid.o g/0000-0003-0426-2342
James O'Dwye h ps://o cid.o g/0000-0003-3058-6295
Ben Gooden h ps://o cid.o g/0000-0003-0575-9078
Ma iana Hoppe h ps://o cid.o g/0000-0003-3685-473X
John Robe s h ps://o cid.o g/0000-0001-9739-5595
Da a a ailabili y
Raw as q iles, associa ed me ada a and analysis pipelines can be accessed om he
CSIRO Da a Access Po al h ps://doi.o g/10.25919/d2gq- 73.
Re e ences
Aba enko K, Nilsson RH, La sson K-H, Taylo AFS, May TW, F øsle TG, Pawlowska J,
Lindahl B, Põldmaa K, T uong C, Vu D, Hosoya T, Niskanen T, Pii mann T, I ano F,
Zi k A, Pe e son M, Cheeke TE, Ishigami Y, Jansson AT, Jeppesen TS, K is iansson
E, Mik yuko V, Mille JT, Oono R, Ossandon FJ, Paupé io J, Saa I, Schigel D, Suija A,
Tede soo L, Kõljalg U (2024) The UNITE da abase o molecula iden i ica ion and
axonomic communica ion o ungi and o he euka yo es: Sequences, axa and clas-
si ica ions econside ed. Nucleic Acids Resea ch 52(D1): D791–D797. h ps://doi.
o g/10.1093/na /gkad1039
Ande son KE, Ca oll MJ, Sheehan T, Mo BM, Maes P, Co by‐Ha is V (2014) Hi e‐
s o ed pollen o honey bees: Many lines o e idence a e consis en wi h pollen p es-
e a ion, no nu ien con e sion. Molecula Ecology 23(23): 5904–5917. h ps://doi.
o g/10.1111/mec.12966
Ankenb and MJ, Kelle A, Wol M, Schul z J, Fö s e F (2015) ITS2 Da abase V: Twice
as Much: Table 1. Molecula Biology and E olu ion 32(11): 3030–3032. h ps://doi.
o g/10.1093/molbe /ms 174
App ill A, McNally S, Pa sons R, Webe L (2015) Mino e ision o V4 egion SSU RNA
806R gene p ime g ea ly inc eases de ec ion o SAR11 bac e ioplank on. Aqua ic
Mic obial Ecology 75(2): 129–137. h ps://doi.o g/10.3354/ame01753
As Y, E oglu GB (2025) Molecula de ec ion and phylogene ic analysis o i e honey bee
i uses om we pollen in Tü kiye. Biology Bulle in o he Russian Academy o Sci-
ences 52(3): 89. h ps://doi.o g/10.1134/S1062359024612898
Balkanska R, Shumko a R, A seno a N, Salko a D, Dunda o a H, Radosla o G, H is-
o P (2023) Molecula de ec ion and phylogene ic analysis o de o med wing i us
and sacb ood i us isola ed om pollen. Ve e ina y Sciences 10(2): 140. h ps://doi.
o g/10.3390/ e sci10020140
Bane jee P, S ewa KA, Dey G, An ognazza CM, Sha ma RK, Mai y JP, Saha S, Doi H, De
Ve e N, Chan MWY, Lin P-Y, Chao H-C, Chen C-Y (2022) En i onmen al DNA analysis
as an eme ging non-des uc i e me hod o plan biodi e si y moni o ing: a e iew.
AoB PLANTS 14: plac031. h ps://doi.o g/10.1093/aobpla/plac031
Ba chelo KL, Bell KL, Campos M, Webbe BL (2023) Can honey bees be used o de ec
a e plan s? Taking an eDNA app oach o ind he las plan s in a weed e adica ion
p og am. En i onmen al DNA 5(6): 1516–1526. h ps://doi.o g/10.1002/edn3.471
Bio á J, Cha iè e J-D, Dos álko á S, Šk abišo á M, Pe ři alský M, Bzdil J, Danihlík J
(2021) Melissococcus plu onius Can Be E ec i ely and Economically De ec ed Using
