The role of seasonality in shaping the interactions of honeybees with other taxa

This is a self-archived version of an original article. This version may differ from the original in pagination and typographic details. Author(s): Title: Year: Version: Copyright: Rights: Rights url: Please cite the original version: CC BY 4.0 https://creativecommons.org/licenses/by/4.0/ The role of seasonality in shaping the interactions of honeybees with other taxa © 2023 the Authors Published version Wirta, Helena; Jones, Mirkka; Peña‐Aguilera, Pablo; Chacón‐Duque, Camilo; Vesterinen, Eero; Ovaskainen, Otso; Abrego, Nerea; Roslin, Tomas Wirta, H., Jones, M., Peña‐Aguilera, P., Chacón‐Duque, C., Vesterinen, E., Ovaskainen, O., Abrego, N., & Roslin, T. (2023). The role of seasonality in shaping the interactions of honeybees with other taxa. Ecology and Evolution, 13(10), Article e10580. https://doi.org/10.1002/ece3.10580 2023 Ecology and Evolution. 2023;13:e10580.   | 1 of 16 https://doi.org/10.1002/ece3.10580 www.ecolevol.org Received:10July2023 | Revised:19September2023 | Accepted:20September2023 DOI: 10.1002/ece3.10580 RESEARCH ARTICLE The role of seasonality in shaping the interactions of honeybees with other taxa Helena Wirta1 | Mirkka Jones2,3 | Pablo PeñaAguilera4 | Camilo ChacónDuque5,6 | Eero Vesterinen7 | Otso Ovaskainen2,8,9 | Nerea Abrego1,8 | Tomas Roslin1,4 ThisisanopenaccessarticleunderthetermsoftheCreativeCommonsAttributionLicense,whichpermitsuse,distributionandreproductioninanymedium, provided the original work is properly cited. ©2023TheAuthors.Ecology and EvolutionpublishedbyJohnWiley&SonsLtd. 1DepartmentofAgriculturalSciences, UniversityofHelsinki,Helsinki,Finland 2OrganismalandEvolutionaryBiology ResearchProgramme,FacultyofBiological andEnvironmentalSciences,Universityof Helsinki,Helsinki,Finland 3HelsinkiInstituteofLifeScience, UniversityofHelsinki,Helsinki,Finland 4DepartmentofEcology,Swedish UniversityofAgriculturalSciences, Uppsala,Sweden 5CentreforPalaeogenetics,Stockholm, Sweden 6DepartmentofArchaeologyandClassical Studies,StockholmUniversity,Stockholm, Sweden 7DepartmentofBiology,Universityof Turku,Turku,Finland 8DepartmentofBiologicaland EnvironmentalScience,Universityof Jyväskylä,Jyväskylä,Finland 9DepartmentofBiology,Centrefor BiodiversityDynamics,Norwegian UniversityofScienceandTechnology, Trondheim,Norway Correspondence HelenaWirta,DepartmentofAgricultural Sciences,UniversityofHelsinki,P.O.Box 27,Helsinki00014,Finland. Email:helena.wir[email protected] Funding information AcademyofFinland,Grant/Award Number:322266,336212and345110; H2020EuropeanResearchCouncil,Grant/ AwardNumber:101057437,101059492 and856506;JanejaAatosErkonSäätiö; KoneenSäätiö,Grant/AwardNumber: 2018104360;NorgesForskningsråd, Grant/AwardNumber:223257 Abstract TheEltoniannicheofaspeciesisdefinedasitssetofinteractionswithothertaxa. Howthissetvarieswithbiotic,abioticandhumaninfluencesisacorequestionof modernecology.Inseasonalenvironments,therealizedEltoniannicheislikelytovary duetoperiodicchangesintheoccurrenceandabundanceofinteractionpartnersand changesinspeciesbehaviorandpreferences.Also,humanmanagementdecisionsmay leavestrongimprintsonspeciesinteractions.Tocomparetheimpactofseasonality tothatofmanagementeffects,honeybeesprovideanexcellentmodelsystem.Based onDNAtracesofinteractionpartnersarchivedinhoney,wecaninferhoneybeeinteractionswithfloralresourcesandmicrobesinthesurroundinghabitats,theirhives, andthemselves.Here,weresolvedseasonalandmanagement-basedimpactsonhoneybeeinteractionsbysamplingbeehivesrepeatedlyduringthehoney-storingperiod ofhoneybeesinFinland.Wethenuseagenome-skimmingapproachtoidentifythe taxonomiccontentsoftheDNAinthesamples.Tocomparetheeffectsoftheseason totheeffectsoflocation,management,andthecolonyitselfinshapinghoneybee interactions,weusedjointspeciesdistributionmodeling.Wefoundthathoneybee interactionswithothertaxavariedgreatlyamongtaxonomicandfunctionalgroups. AgainstabackdropofwidevariationintheinteractionsdocumentedintheDNAcontentofhoneyfrombeesfromdifferenthives,regions,andbeekeepers,theimprintof theseasonremainedrelativelysmall.Overall,ahoney-basedapproachoffersunique insightsintoseasonalvariationintheidentityandabundanceofinteractionpartners amonghoneybees.Duringthesummer,theavailabilityanduseofdifferentinteractionpartnerschangedsubstantially,buthive-andtaxon-specificpatternswerelargely idiosyncraticasmodifiedbyhivemanagement.Thus,thebeekeeperandcolonyidentity are as important determinants of the honeybee's realized Eltonian niche as is seasonality. KEYWORDS Apis mellifera,Eltonianniche,honey,jointspeciesdistributionmodeling,management,microbe, plant,wholegenomesequencing 2 of 16 | WIRTA et al. 1 | INTRODUCTION Theecologicalnicheofaspeciescanbecharacterizedfromtwo perspectives:asthespecies'responsetoabioticconditions(the Grinnellianniche;Grinnell,1917;Whittakeretal.,1973) and as its interactions with other taxa in the surrounding community (the Eltonian niche; Elton, 1927). Over the past few decades, there hasbeenasignificantinterestfocusedoncharacterizingspecies' Grinnellian niches due to changes in global abiotic conditions. However, the Eltonian niche is as important as the Grinnellian niche to be understood (see, e.g., Wirta et al., 2022), as environmentaleffectsonbothaspectsofthenicheareequallylikely (Graveletal.,2019; Pellissier et al., 2018).Thus,weshouldfurther ourunderstandingofhowexternalimpactsshapecommunitydynamicsandecologicalinteractionnetworks,namelytheEltonian niche(Graveletal.,2019). Inseasonalenvironments,therealizedEltoniannichesetislikely tovaryacrosstheseason,asdrivenbyperiodicchangesintheoccurrenceandabundanceofinteractionpartnersandbychangesin speciesbehavior.Seasonalityreferstomajorchangesinaspecies' environmentthatarepredictablyrepeatedeachyear.Howspecies' interactionsare influencedby seasonalcycleshas been the focus ofintenseresearch,inparticularinthecontextofchangingspecies' phenologies(e.g.,Ekholmetal.,2019; Kešnerová et al., 2020;Rabeling et al., 2019). Species inhabiting seasonally fluctuating environments experience variations in the intensity of their interactions, which are influencedbythechangingseasons.Inotherwords,seasonalityis likelytoshapedifferentdimensionsoftherealizedEltonianniche differently,wheresomeinteractionsarestronglyaffectedwhereas othersareweaklyaffectedorremainunaffected.Ononeendofthe spectrum, certain interaction partners are only accessible during specifictimewindows,dictatedbytheirownphenologicalpatterns. Thisleadstoasignificantturnoverininteractionpartnersovertime. Ontheoppositeend,anothergroupofinteractionpartnersremains activeconsistentlythroughdifferentseasons,resultinginminimal turnoverininteractions.Nonetheless,it'sworthnotingthattheimpact of seasonality on species interactions has traditionally been examinedforonlyalimitedsubsetofinteractingtaxaatanygiven time.(e.g.,Bauer&Hoye,2014;Hutchisonetal.,2020;Rasmussen et al., 2013, 2014;Rudolf,2019). Beyond seasonal effects, human management decisions may leavestrongimprintsonspeciesinteractionsandtherefore,onthe realized Eltonian niches. By affecting the availability of resource species across landscapes, humans may strongly affect the set of realizedinteractions(Kortschetal.,2023).Fordomesticatedorhalf- domesticatedspecies,theseeffectswillbemostpronounced,asthe human actor will affect both the focal species and which species itinteractswithbyactivelyaltering,forexample,itsaccesstoresources and its pathogen load. A challenge for exploring the wholesale seasonal and anthropogenicdriversoftheEltoniannicheisthecomplexityofresolving largesetsofinteractionsinempiricalsystems.Here,thehoneybee, Apis mellifera,offersauniquestudysystemforassessingseasonal andothereffectsontherealizedEltonianniche(Wirtaetal.,2022). Honeybeeshavebeeninsertedbyhumansinenvironmentscharacterizedbydifferentseasonalityanddifferentmanagementpractices throughouttheworld.Importantly,theseinteractionscanbereconstructed from DNA traces left in honey (Bovo et al., 2018, 2020; Cirtwill et al., 2022;Galanisetal.,2022;Leponiemietal.,2023).Such studiestodatehaveshownthathoneybeesinteractwithamultitude ofothertaxa,mostimportantlyfloweringplants,butalsomicrobes (Aizenberg-Gershteinetal.,2013;Engeletal.,2016; Moran, 2015; Wirtaetal.,2022). While interactions between honeybees and plants tend to be mutualistic in nature, interactions between bees and microbes can be either pathogenic, mutualistic, or neutral in nature (Engel et al., 2016; Morse, 1994).Asanexampleofapathogenicinteraction,theinteractionofhoneybeeswiththebacteriumPaenibacillus larvaewillcauseseverediseaseinhoneybees.Incontrast,interactionsofhoneybeeswithSnodgrassella alvi or Gilliamella apicola can bedescribedasmutualisticsincethesebacterialiveinthehoneybee'sgut,sustainingthehoneybee'shealth(Fünfhausetal.,2018; Raymann&Moran,2018). The strongest effects of seasonality on honeybee niches willlikelyoccurathighlatitudes,whereseasonalenvironmental changes are most pronounced. Here as everywhere else, bees will encounter a wide range of floral resources (Lehmuskallio & Lehmuskallio,2006; Ruottinen et al., 2003;Salonenetal.,2009), butalsointeractwitharangeofmicrobes,includingthosepresent onflowers(Jonesetal.,2018).Bothmicrobeslivinginthehiveand beepathogenshavebeenfoundtochangeseasonally(Donkersley et al., 2018; Runckel et al., 2011),possiblyfollowingthephenologiesofdifferentplantspecies.Thus,seasonalityislikelytoaffect bothmutualisticandantagonisticinteractionsbetweenbeesand othertaxa. Apart from seasonal effects on the honeybees, their Eltonian niche is likely to be shaped by anthropogenic factors with an impact on how the colony explores and utilizes its environment. Of particular interest are management practices, including measures ofdiseasecontrol,overwintering,andhivesize.Theseeffectscan becapturedbytheidentityofthebeekeeper,whowillapplysimilar methodstotheirhives(Morse,1975, 1994; Ruottinen et al., 2003). However,evenwithsimilarmanagementpractices,individualhives sustainedbyanindividualbeekeeperwillalsodifferfromeachother. Thisisduetodifferencesinforagingbehavior,foragingcapacity,and TAXONOMY CLASSIFICATION Agroecology,Appliedecology,Communityecology 20457758, 2023, 10, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10580 by University Of Jyväskylä Library, Wiley Online Library on [12/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | 3 of 16 WIRTA et al. susceptibilitytodiseasesoftheindividualcolony(Wrayetal.,2011). Suchaspectsarestronglyaffectedbythecharacteristicsandhealth ofthe queen, shaping the performance of thecolonyandfurther modifyingitsbehavior(Amirietal.,2017).Theeffectsofthesefactorscanthusbecapturedbytheidentityofthehive.Additionally, thespecificenvironmentinwhichthehiveisplacedalsoaffectsthe behaviorofthecolony. Inthispaper,we usehoneysamplesfromFinlandto compare the role of seasonality to that of management in determining the interactionsofhoneybees with other taxa (plants,bacteria, fungi, andviruses).Forthispurpose,wedrawonagenome-skimmingapproachtotheDNAtracesstoredinthehoney.Toresolvetemporal variationintheinteractionrecordsofhoney,wesampledhivesrepeatedlyduringthehoney-storingperiodofhoneybeesandasked thefollowingquestions: 1. How does time of the season compare to the geographical location (site), management practices (beekeeper), and colony identity (hive) in terms of its influence on the taxa which honeybees encounter and interact with? 2. Howdoestheuseoffloweringplantsbyhoneybeeschangeduringthemainfloweringseason(i.e.,summer)? 3. How do the interactions between honeybees and microbes changeduringthesummer,andhowdopatternsdifferbetween differenttaxonomicandfunctionalgroupsofmicrobes? 4. Willco-occurrencepatternsamongtaxadetectedintemporally- resolvedhoneysamplessuggestinteractionsorphenologicalassociationsamongthetaxathemselves? 2 | MATERIALS AND METHODS 2.1 | Seasonality in NorthEuropean honeybee resources InnorthernEurope,beescanactivelyinteractwithorganismsoutsidetheirhiveforabout6 months(BennoMeyer-Rochow,2008; Ruottinen et al., 2003).PollenforagingtypicallystartsinMarchor April,withthecolonyreachingitsmaximumsizeinMayandJune. From mid-June to mid-August, the bees work on storing honey, andbylate-August,thecolonybeginspreparingforoverwintering byproducingthelastworkersoftheyear(Ruottinenetal.,2003). Duringthis6-monthperiod,thebeeswillinteractwitharangeof flowering plants, each with its own phenology. Floral resources aretypicallymostabundantinlate-JuneandJuly,whenbothearly andlatesummerfloweringspeciesareinbloomsimultaneously. Thisconcernsboththespeciesrichness offloweringplantsand theirfloralabundance.Ofthetypicalplantsusedbyhoneybees in Finland, willows (Salixspp.)anddandelion(Taraxacum spp.) begintobloominMay,thenrapeseed(Brassicaspp.),raspberry (Rubus idaeus),clovers(Trifoliumspp.),andfireweed(Epilobium angustifolium)flowerfromJuneonwards,andthistles(Cirsium spp.) andheather(Calluna vulgaris)begintheirfloweringonlylaterin July (Benno Meyer-Rochow, 2008; Lehmuskallio & Lehmuskallio, 2006;Salonenetal.,2009). 2.2 | Using honey as an archive of interaction partners Alargeproportionoftaxathathoneybeesencounterorinteract withcanbefoundinandidentifiedfromhoney,whereDNAtraces ofthesetaxatendtobewellpreserved.ByidentifyingtheDNA foundinhoney,onecanthustellwhatothertaxahoneybeeshave encounteredorinteractedwith,especiallyfortheirinteractions withmicrobesandplants(Bovoetal.,2020;Wirtaetal.,2022). Addingtotheinformationvalueofhoney,nectarisspreadinto opencombsfordrying,andbeesaddenzymaticsecretionsduringtheprocessingofnectarintohoney(Crane,1979), and these processingstageswouldallowDNApresentinanyformwithin the hive to enter the nectar, turning it into honey. In practice, recently produced honey can be distinguished by its looks and position:onthehoneyframesofahive,thisfreshhoneysitsnext tohoneystilluncoveredbywax,andpartofthecombsareyet toreceiveafullwaxcover.Thisnewhoneyconveysasampleof thehoneybees'interactionsduringthelastweek,corresponding tothetimeduringwhichthisnectarhasbeencollectedandprocessedintohoneybythebees.Ingeneral,thetimetakenbynectartoripenintohoneytendstovaryfrom3to7 days,depending onweather,colonystrength,andnectaravailability(Crane,1979; Morse, 1975, 1994). Overall, the many processing stages involved in converting nectartohoney,therepeatedmanipulationofthenectarbythe bee,andthetimespentdryingin opencombsallowDNApresent in multiple forms within the hive to enter the nectar. Thus, thehoneyofabeehiveoffersawell-preservedrecordofrecent interactionpartnersofitsbees.However,it'scrucialtonotethat notallDNAfrominteractionsiscarriedbacktothehive,andconsequently,someofitdoesnotbecomeapartofthehoney.For instance,whenhoneybeesarepreyedupon,theDNAofthepredator is not included in the honey, leading to the undetected nature oftheseinteractions. 2.3 | Sampling Tocharacterizeseasonalvariationinthemicrobialandfloralinteractionpartnersofhoneybees,weobtainedhoneysamplesdirectly frombeehivesbelongingto14Finnishbeekeepers(Figure 1).Each beekeeperselectedtwoorthreeoftheirhives,totaling41hives forthestudy.Fromeachhive,honeywascollectedatthreetime points during the season, before the final harvest of all honey from the hives. Thus, samples were obtained from late-June to mid-August.Allbeekeeperswereaskedtosampletheirhivesduringthesameweeks,correspondingtothe22ndto28thofJune, the13thto19thofJuly,andthe3rdto9thofAugustof2020. 20457758, 2023, 10, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10580 by University Of Jyväskylä Library, Wiley Online Library on [12/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 4 of 16 | WIRTA et al. Toensurethatthehoneysamplerepresentedthespecifiedtime of the season, the beekeepers were instructed to collect only honeynewlycoveredbywax.Toobtainuncontaminatedsamples, weprovidedthebeekeeperswithDNA-freesamplingequipment. To ensure that the sample was representative of the variety of nectarrecentlycollected,aspoonfulofhoneyfromthreedifferent frames was combined in each sample. Additionally, we also obtainedasampleofhoneyfromtheendofseasontotalyieldof eachbeekeeper.Thissamplewasusedtoassesswhethersucha time-aggregatedsamplewillincludealltheinformationgathered fromtheseparatetime-specificsamples. DuetoadryperiodinearlyJulyof2020,therewasashortage offlowersinpartsofthestudyarea.Therefore,Julysamplescould notbeobtainedforallhives.Furthermore,forsomeofthesamples, notenoughDNAcouldbeextractedfromthe20 gofhoney.Thus,in total,wewereabletosequence115samplesfromindividualhives and13samplesofcompound,endofseasonhoneyasharvestedby individualbeekeepers. 2.4 | Laboratory methods ToidentifythetaxonomicoriginofDNAinhoneysamplesfromdifferentpartsofFinland,weusedaPCR-freemetagenomicapproach. Instead of metabarcoding, where single genes are amplified and sequencedinasampleusingprimerstargetedtothespecificgene region(e.g.,Vesterinenetal.,2018),weutilizedagenome-skimming approachtosequencerandomfragmentsofeachspecies’genome presentinasamplewithoutanylocus-specificPCR(see,e.g.,Coissac et al., 2016). FIGURE 1 Locationsofbeehives sampledforhoneyinFinland,colored bybeekeeperidentity(with2–3hives sampledperbeekeeper).Notethatsome beekeepershadhivesatmorethanone site. To resolve overlapping sites, the locationsofhiveshavebeenslightly jitteredinboththehorizontalandvertical planes. 20457758, 2023, 10, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10580 by University Of Jyväskylä Library, Wiley Online Library on [12/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | 5 of 16 WIRTA et al. To prepare the samples for DNA extraction, two subsamples of10 gwereeachdiluted with30 mL of DNA-free water(double- distilled“MQ-water”).These subsampleswere allowedtodissolve for 1 h at +60°C. To collect all the tissue material and to remove excesswater,thesubsampleswerecentrifugedfor60 minat8000 g (Centrifuge5810R,Eppendorf).Mostofthesupernatantwasdiscarded, and the pellets from the two subsamples were combined into a 2 mL tube. The tube contents were further centrifuged for 5 minat11,000 g(Heraeus Pico 21 centrifuge, ThermoScientific). The remaining supernatant was discarded, and the pellets were storedat−20°CuntilDNAextraction. Total DNA was extracted from each sample with the DNeasy PlantMiniKit(Qiagen)withthefollowingmodificationstotheprotocol:Initially,thepelletwasresuspendedin400 μLofbufferAP1,and then 4 μL RNase, 4 μL proteinase K (20 mg/mL, Macherey-Nagel), and one 3 mm tungsten carbide bead was added to each sample tube.Thesamplewasthendisruptedfor2 × 2 min30 Hz(MixerMill MM400,Retsch).DNAextractionthenfollowedtheprotocol,exceptthattheQIAshreddercolumnstepwasomittedtoavoidDNA loss.Alllaboratorystepsweredoneinalaminarhoodwipedwith ethanolandcleanedofDNAwith1 hofUVlighteverynight.We onlyusedDNA-freetubes,pipettetips,andPCRplates,aswellas DNA-freewater. DNAquantitywasmeasuredwithaQubit4fluorometer(Thermo FisherScientific).Forpreparingthesequencinglibrary,thesamples weredilutedtoaconcentrationof1 ng/μL.SampleswithDNAconcentrations <1 ng/μLwerenotdiluted.Thequalityandquantityof DNAin eachsample were measuredwithgenomic DNA TapestationandD500HSTapestation,beforethepreparationofthelibrary. TheNexteraXTtransposome,providedwiththeIlluminaNextera XTlibraryPreparationKit(Illumina,Inc.),wasusedtofragmentthe DNAinto150-bp-longpiecesandtotagtheDNAwithadaptersequences, following the Nextera XT Protocol. After this, an indexingPCR to anneal sample-specific indexes to theDNAfragments wasrun,andtheindexingPCRproductswerecleaned.Thesample- specificlibrarieswerenormalizedtothesamequantity,afterwhich theywerecombinedintothepooledlibrarytobesequenced.Allthe stepstopreparethesequencinglibraryfromthetotalDNAfollowed theNexteraXTProtocol(IlluminaInc,2019).Thelibrarywasthen sequencedinanIlluminaNovaSeq6000S4flowcell,using80%of one(outoffour)flowcelllane,equaling20%ofthetotalsequencing capacityoftherun.AllsequencingwasperformedbytheFunctional GenomicsUnitattheUniversityofHelsinki,Finland.Todetectpossiblecontamination,wesequencedaDNAextractionblankcontrol inthesameway. 2.5 | Bioinformatic processing Toremove anylow-qualitybasesfromthestart and endofreads and the Illumina adapter sequences, the raw reads were trimmed using Trimmomatic version 0.39 (Bolger et al., 2014) with the ILLUMINACLIP adapter-clipping settings “adapters.fa: 2:30:10 LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:50”. Toassembletrimmedreadsintodenovoscaffolds,weapplieddifferentk-merlengths[k-mer = 21,33,55,77,99,and121;following Nurketal.(2017)]usingtheSPAdesassemblytoolkitversion3.15.0 (Bankevichetal.,2012; https://github.com/ablab/spades) with the— meta flag (recommended for metagenomic data sets). To reduce heterozygosity, we then applied the Redundans pipeline (Pryszcz &Gabaldón,2016)totheassembledscaffolds,withdefaultvalues ofidentity0.51andoverlap0.8,andaligningallreads(alignsubset ofreadswithalimitvalueof1).ThereducedscaffoldswereannotatedtoNCBITaxIDsusingBLASTNsearchesagainsttheNCBInon- redundantnucleotidedatabase(nt)database(November–December 2021),keepingonealignedsequenceperscaffold(max_target_seqs 1),savingonlythebestalignmentforeachquery-subjectpair(max_ hsps 1), and with an E-valuelessthan1 × e−25.Tomapalltheoriginal trimmedandcorrectedsequencestothetaxonomicallyannotated referencescaffolds,BWAMEM(Li&Durbin,2009, 2010) was used, andtheresultsweresortedintobamformatfilescontainingsample,sequence,andmappedreaddatawithSAMtools(Lietal.,2009). Foreachassembly,theassociatedstatisticsatfourtaxonomicranks (phylum,family,genus,andspecies)weregeneratedwithBlobtools (Laetschetal.,2017)basedontheBLASTnsimilaritysearchresults. Tofurtherfilterallreads,withtheintentofremovingpotentially misassignedreadsandfalsepositivesduetotagjumpingorcontamination,wefollowedaconservativeapproach(followinge.g.,Alberdi et al., 2012; Lee et al., 2018).Asasmallnumberofreadsrepresentingalimitednumberoftaxawerefoundinthecontrolsample,we subtractedthesereadsfromthereadnumbersofthecorresponding taxainthehoneysamples.Asafinalfilteringstepaimedatremoving extremelyrareand/orspuriousreads,wecalculatedthemeanrelativereadabundance(RRAhereafter;Deagleetal.,2019)oftaxa(here genera)withinsamplesandremovedanytaxaandreadsassignedto taxawithasample-specificRRAof<0.001%.Fortheanalyses,we onlyincludedgenerawith≥0.01%meanRRAacrossthesamples. 2.6 | Occurrence of taxa and relative read abundances InanalysesbasedonRRA,astrongincreaseintheabundanceofany taxonwill,pernecessity,bereflectedinareductionintheproportionalrepresentationofothertaxa.Threegeneradominatedsome samples: Apilactobacillus(A. kunkeii), Zygosaccharomyces(Z. rougii), and the virus Apis mellifera filamentous virus (AmFV; see Text S1 and Figure S1).Forsomeofthesamples,thesetaxaaccountedfor mostreads(upto85.8%,91.2%,and99.1%forA. kunkeii, Z. rougii, andAmFV,respectively).Thus,torestricttheimpactofthesetaxa onpatternsinothertaxa,wealsocalculatedRRAafteromittingall readsassignedtothethreedominanttaxaidentifiedabove.Inthe analyses,weusedthepresence–absencedataofalltaxa(withmean RRA across samples ≥0.01%), but for abundance data, we omittedthethreetaxawithhighyetvariableproportions(TextS1 and Figure S1). 20457758, 2023, 10, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10580 by University Of Jyväskylä Library, Wiley Online Library on [12/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 6 of 16 | WIRTA et al. The proportions of reads assigned to individual kingdoms of associated taxa (plant, bacterial, fungal, and viral genera) varied substantially between samples even after removing Apilactobacillus, Zygosaccharomyces, and AmFV (Figure S2). Thus, while we includedbothplantsandmicrobesintheanalyses,wealsodescribed thechangesinplantsandinmicrobesseparatelyfromeachother (Figures S 3 – S 5 ). The sequencing of the samples by Illumina NovaSeq s4, with 80%ofaflowcelllane,resultedin3.72billionreadspassingthefilter(thissequencingrunincluded140samples,outofwhich118plus anegativecontrol sampleare partofthisstudy).Forthesamples inthisstudy,2689.8millionreadspassedthequalitycontrols,averaging23.0millionreadsper sample. 85.8% ofthesereadswere assignedtothegenuslevelandthusretainedforfurtheranalyses. Inadditiontoplants,microbes,fungi,andviruses,weidentified11 animalgenerainthesamples,butthesewerenotconsideredinthe analyses. 2.7 | Functional groups of taxa Toresolvetaxaofdifferentfunctionalaffinitiesandofdifferentassociationswithhoneybees,weclassifiedthegenerafollowingWirta etal.(2022).Theliteratureusedinassigningtaxatospecificfunctional groups is shown in Table S1.Whenreadswithinagenuswere primarily(>90%)assignedtoagivenspecies,webasedthefunctional assignmentofthegenusoninformationassociatedwiththisspecific species.Whenreadswithinagenuswereassignedtomultiplespecies,weassessedthefunctionbasedonaspeciesknowntobeassociatedwithhoneybees.Finally,inthecasewherereadswerenot assignedtoanyparticularspecies,weassessedthefunctionbased onthegeneralbiologyofthegenus. Plants were classified into two groups based on their nectar- producingability.Microbescloselyassociatedwithbeeswereclassifiedascommonbeegutmicrobes,asbeehivemicrobes,orasbee pathogens.Microbeswithoutanyknownassociationwiththebees wereclassifiedasplantpathogens,asanimalpathogens,orasmicrobes known to be beneficial or neutral for plants and animals. Those microbe genera, which were known to have multiple roles, were categorized according to their relationship with honeybees. For instance, bacteria in the genus Lactobacilluscould bepresent innectar,butsomespeciesofthisgenusareconsideredubiquitous inhoneybeeguts,andthusweclassifiedLactobacillusasabeegut microbe (Raymann & Moran, 2018; Vannette, 2020). When the functionalattributeofagenuswasuncertain,thenthegenuswas classifiedasunknown. 2.8 | Statistical modeling To examine the strength and patterns of seasonal imprints on honeybee associations and to compare them to the impacts of thebeekeeper,thesite,thehive,andthesampleitself,weapplied the joint species distribution modeling framework of Hierarchical Modeling of Species Communities (HMSC; Ovaskainen & Abrego,2020). Toaccountforthezero-inflatednatureofthedata,weapplied a hurdle modeling approach, modeling presence–absence with probit regression and abundance conditional on presence using a log-normalmodel.Asresponsedata,weusedamatrixofpresence– absencesofallgenerainthepresence–absencemodelsandthematrixoflog-transformedRRA'sinmodelsofabundanceconditionalon presence(henceforthreferredtoasabundancemodels).Sincetaxa withaparticularlyloworhighprevalencecontainlittleinformation onthefactorsaffectingtheiroccurrence,weexcludedgenerathat werepresentinlessthan5%ofthesamplesfrombothmodels.We notethatwhilepresence–absencesandabundancesweremodeled separately, these two models were used simultaneously to make predictions(seebelow). Theexplanatorypartofthemodelswasidentical,asfollows: As fixed effects, we included the sampling period (a categorical variable with three levels) and the log-transformed number ofreadspersample.Thevariableoflog-transformednumberof readsaccountsfortechnicalvariationinsequencingdepthamong samples.Namely,thisvariableismeanttocapturetheeffectof varying sampling effort among samples due to variation in sequencingdepth.Toaccountforthestructureofthestudydesign, weincludedasexplanatoryrandomeffectsthesite(n = 30),the hive(n = 41),thebeekeeper(n = 14),andthesample(n = 115),of whichthesitewasdefinedasaspatiallyexpliciteffect.Wenote thatthesample-levelrandomeffectwasincludednotnecessarily to account for the spatial structure of the data but to estimate the speciestospecies association networks through latent variable modeling(Ovaskainen etal.,2016).Totestwhetherdifferent taxonomic groups respond differently to sampling time, we includedbroadtaxonomic(plants,bacteria,fungi,andviruses)and functionalgroups(describedabove)asgenus-leveltraitvariables. The models were fitted with the R-package Hmsc (Tikhonov et al., 2020),assumingthedefaultpriordistributions(seeOvaskainen&Abrego,2020,pp.184–216).WesampledtheposteriordistributionwithfourMarkovChainMonteCarlo(MCMC)chains,each ofwhichwasrunfor375,000iterations,ofwhichthefirst125,000 wereremovedasburn-in.Thechainswerethinnedby1000toyield 250 posterior samples per chain and 1000 posterior samples in total. We examined MCMC convergence as a function of the potentialscalereductionfactors(Gelman&Rubin,1992)ofthemodel parameters. The explanatory and predictive powers of the presence– absence models were examined through the metrics of Tjur's R2 (Tjur, 2009) and the Area Under the Curve(AUC;Fielding& Bell, 1997). For the abundance models, we used the R2ofthe linear model(Ovaskainen & Abrego, 2020). Tocompute explanatorypower,wemademodelpredictionsbasedonmodelsfitted toallthedata.Tocomputepredictivepower,weperformedtwofoldcross-validation,inwhichthesubstrateunitswereassigned randomlytotwofolds,andpredictionsforeachfoldwerebased 20457758, 2023, 10, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10580 by University Of Jyväskylä Library, Wiley Online Library on [12/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | 7 of 16 WIRTA et al. onamodelfittedtothedataontheotherfold.Toquantifywhat portionoftheexplained variance wasattributedtoeach of the explanatoryfactorsincludedinthemodels,weappliedavariance partitionapproach.Wethenusedthefittedmodelstobuildpredictionsontheresponsesofthegeneratotheseason.Todoso, weusedthefixedeffectpartofthemodelonlyandpredictedfor eachgenusitsoccurrenceprobabilityforeachofthethreetime points.Werepeatedthepredictionforthe1000samplesofthe posterior distribution to compute the posterior probability by whichthegenushadahigheroccurrenceprobabilityinlateseason(August)thanearlyseason(June).Wefurtherconvertedthe genusresponsestotheseasontoatemporalco-occurrencematrix 𝛀 ,withtheelementcorrespondingtogenuspair( j1,j2 )computed as Ω j 1 ,j 2 =𝛽 T2,j 1𝛽 T2,j 2 +𝛽 T3,j 1𝛽 T3,j 2 , where 𝛽T2,j and 𝛽T3,j are the genus responsestotimepointsT2(July)andT3(August),withtimepoint T1(June)beingsetasthereferencelevel.Toexaminethelevelof statisticalsupportbywhicha givengenuspair co-occursatthe sametime,wecomputedtheposteriordistributionof 𝛀 and then evaluated the posterior probabilities by which each matrix elementwaspositiveornegative. 3 | RESULTS Overall,wedetectedatotalof49plantgenera,45bacterialgenera, 23 fungal genera, and three viral genus-level groups with a mean relativereadabundance(RRA)exceeding0.01%.Theproportionsof readsassignedtodifferentkingdoms(plants,bacteria,fungi,andviruses)variedconsiderablybetweensamples(Figure S2).Persample, theaverageproportionsofplants,bacteria,fungi,andviruseswere 51%,38%,7%,and3%,respectively. 3.1 | Model fit statistics The fittedjoint species distribution models showed high explanatory power both for the presence–absence (Tjur's R2 = 0.42 and AUC = 0.93) and abundance conditional on presence (R2 = 0.64) models(Table 1 and Figure 2a).Nonetheless,theexplanatorypower variedwidelyamonggeneraaswellasamongtaxonomicandfunctional groups. Among taxonomic groups, the explanatory power washighest forfungi,explaining65%and91%ofthevariationin thepresence–absenceandabundancemodels,respectively.Among functionalgroups,animal pathogens reachedthehighestexplanatorypower,explaining69%and92%ofthevariation. Thepredictivepowerwasfarlowerthantheexplanatorypower forboththepresence–absence(Tjur'sR2 = 0.11andAUC = 0.63)and abundance models (R2 = 0.02). However, this result seems attributabletothefactthatthesamplingunit-levelrandomeffects(i.e., samplelevel)accountedforalargepartoftheexplainedvariation (22.3% for the presence–absencemodel and 22.6%for the abundancemodel).Theserandomeffectswillcontributetotheexplanatorypowerbutnottothepredictivepowerofthemodels. 3.2 | Seasonal effects on the interactions of honeybees A variance partitioning among the fixed and random effects showedthattheseasonalimprintexplained,onaverage,3.2%of therawvarianceinthepresence–absencesand7.4%intheabundancesofthetaxa.Thestrengthoftheimprintofthebeekeeper, hive, and site on the occurrences of the interactions of honeybeeswassimilartothatoftimeoftheseason(withthebeekeeper, hive,andsiteexplaining3.2%,2.3%,and4.0%ofthevariance,respectively).However,thehivehadastrongereffectontheabundancesofthetaxahoneybeesinteractwith,explaining14.2%of the variance. Theproportionofvarianceattributedtothetimeoftheseason variedgreatlyamongtaxaandamongbothtaxonomicandfunctional groups (Table 1 and Figure 2). Among taxonomic groups, viruseswerethemostinfluencedandfungitheleastinfluenced bythetimeoftheseason(bothintermsofpresence–absenceand abundance). Among functional groups, the no-nectar-producing planttaxawere the most influenced bythe time of the season. Theamountofvariationexplainedbythetimeoftheseasonvariednotonlyamongtaxonomicandfunctionalgroupsbutwithin groups as well (Figure 2). The occurrences and abundances of sometaxawerewellexplainedbythetimeoftheseason,whereas the occurrences and abundances of other taxa were totally unaffectedbythetimeoftheseason.Asexamples,theoccurrence ofsomeplantswasstronglyimpactedbythetimeoftheseason. ForChamaenerion,samplingtimeaccountedfor11.1%ofthevariationinthepresence–absencemodel,whileforLactuca,thetime oftheseasonaccountedforonly0.7%ofthevariation(Table S2). Intermsofabundances,Taraxacumwasthegenusmostimpacted by the time of the year (with time accounting for 31.6% of the variation), while Medigaco and Cicerfellattheoppositeextreme (withtimeaccountingfor1.0%ofvarianceexplained;Table S3). Inregardtomicrobes,thetimeoftheseasonimpactedtheoccurrencesofthetwoviralgroupsthemost(accountingfor8.3%and 10.0%oftheirvariationforthepresence–absencemodel),while thetimeoftheseasonhadtheleastimpactonthefungalgenus Histoplasma (accounting for 0.5% of its variation; Table S2). For theabundances,thetimeoftheseasonhadastrongimpactonthe bacterialgenusAcinetobacter(accountingfor22.4%ofthevariation),whilethebacterialgeneraPantoea and Pectobacterium were theleastimpactedbytime(accountingfor1.2%and1.5%ofvariation, respectively; Figure 2 and Table S3). Furthermore, the temporal patterns of honeybee interactions withdifferentplantgeneradifferedstronglyamonghives.Somebee colonies,thatis,honeybeesfromparticularhives,usedasimilarset ofplantgenerathroughoutthesummer,withonlygradualchangesin theirrelativeproportions(TextS2).Othercoloniesshiftedstrongly to a particular plant genus, such as Brassica,fromonetimepointto theother(Figure S3).Theoccurrencesandrelativeabundancesof microbes differed greatly among colonies and across time points (Figure S4).Curiously,forsomecoloniesandsamples,themicrobe 20457758, 2023, 10, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10580 by University Of Jyväskylä Library, Wiley Online Library on [12/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 8 of 16 | WIRTA et al. TABLE 1 Meanpercentagesofvarianceexplainedbythefixedandrandomeffectsincludedinthemodels,andtheirsummedtotalexplanatorypowerquantifiedbyTjurR2forthepresence– absencemodelandR2fortheabundanceconditionalonpresencemodel. Taxonomic groups Functional groups Total Plants Bacteria Fungi Viruses Nectar producing No nectar producing Bee gut Beehive Bee pathogens Plant pathogens Neutral or positive Animal pathogens Presence– absence Time 3.2 3.9 3.1 1.4 9.2 3.3 7.3 3.3 3.9 3.5 2.2 2.2 2.0 Total_reads 2.9 2.1 4.2 2.1 0.9 1.8 4.0 2.8 3.3 3.6 3.5 2.3 4.2 Random:sample 26.3 15.8 21.0 58.0 14.7 16.0 14.2 11.6 26.9 29.6 46.2 26.1 61.1 Random:site 4.0 7.8 1.7 1.0 0.6 8.0 6.8 1.5 1.3 4.1 0.7 0.9 0.9 Random:hive 2.3 3.0 2.4 0.9 1.5 2.7 4.7 3.9 2.3 1.9 0.7 2.3 0.6 Random:beekeeper 3.2 2.6 4.8 1.5 0.8 2.8 2.0 2.3 5.9 2.9 0.7 21.1 0.5 Explanatorypower Tjur R2 42 35 37 65 28 35 39 25 44 46 54 55 69 Abundance conditional on presence Time 7.4 7.7 5.4 9.8 15.7 8.0 6.0 5.4 2.8 7.7 8.7 7.5 7.0 Total_reads 3.7 2.7 3.4 6.3 3.4 2.8 2.3 2.7 1.3 3.6 5.8 1.5 5.0 Random:sample 22.6 20.4 33.2 8.0 5.5 23.4 2.2 16.3 60.7 16.6 27.6 28.0 38.5 Random:site 5.3 4.8 5.0 7.0 4.0 5.0 3.5 4.8 2.5 6.7 5.4 6.5 4.8 Random:hive 14.2 3.8 5.7 55.1 4.4 4.1 2.3 5.2 3.2 21.1 32.7 13.9 33.4 Random:beekeeper 8.1 10.3 7.2 5.2 2.5 10.6 8.3 7.2 1.7 7.7 4.8 14.9 2.8 ExplanatorypowerR261 50 60 91 36 54 25 42 72 64 85 72 92 Note:Oneachrowwereporttheaveragevaluesoverallthetaxamodeled(column“total”)andperindividualtaxonomicandfunctionalgroup(allothercolumns). 20457758, 2023, 10, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10580 by University Of Jyväskylä Library, Wiley Online Library on [12/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | 15 of 16 WIRTA et al. Jones,L.,Brennan,G.L.,Lowe,A.,Creer,S.,Ford,C.R.,&DeVere,N. (2021). Shifts in honeybee foraging reveal historical changes in floralresources. Communications Biology, 4(37), 1–10. https://doi. org/10.1038/s4200 302001562 - 4 Kešnerová, L.,Emery,O.,Troilo, M.,Liberti, J.,Erkosar, B.,& Engel, P. (2020). Gut microbiota structure differs between honeybees in winter and summer. ISME Journal, 14(3), 801–814. https://doi. o r g / 1 0 . 1 0 3 8 / s 4 1 3 9 6 - 0 1 9 - 0 5 6 8 - 8 Kortsch, S., Saravia, L., Cirtwill, A. R., Timberlake, T., Memmott, J., Kendall,L.,Roslin,T.,&Strona,G.(2023).Landscapecomposition andpollinatortraitsinteracttoinfluencepollinationsuccessinan individual-basedmodel.Functional Ecology, 37,2056–2071.https:// doi.org/10.1111/13652435.14353 Kwong,W.K.,Medina,L.A.,Koch,H.,Sing,K.W.,Soh,E.J.Y.,Ascher, J.S.,Jaffé,R.,&Moran,N.A.(2017).Dynamic microbiomeevolutioninsocialbees.Science Advances, 3(3),1600513.https://doi. org/10.1126/sciadv.1600513 Laetsch, D. R., Blaxter, M. L., Eren, A. M., & Leggett, R. M. (2017). BlobTools:Interrogationofgenomeassemblies.F1000Research, 6, 1287. https://doi.org/10.12688/f1000research.12232.1 Lee,T.,Alemseged,Y.,&Mitchell,A.(2018).Droppinghints:Estimating thedietsoflivestockinrangelandsusingDNAmetabarcodingof faeces. Metabarcoding and Metagenomics, 2, e22467. https://doi. org/10.3897/mbmg.2.22467 Lehmuskallio,E.,&Lehmuskallio,J.(2006).Luontoportti.Retrievedfrom https://luontoportti.com/ Leponiemi,M.,Freitak,D.,Moreno-Torres,M.,Pferschy-Wenzig,E.-M., Becker-Scarpitta,A.,Tiusanen,M.,Vesterinen,E.J.,&Wirta,H.K. (2023).Honeybees'foragingchoicesfornectarandpollenrevealed byDNAmetabarcoding.Scientific Reports, 13, 14753. https://doi. o r g / 1 0 . 1 0 3 8 / s 4 1 5 9 8 - 0 2 3 - 4 2 1 0 2 - 4 Li,H.,&Durbin,R.(2009).Fastandaccurateshortreadalignmentwith Burrows–Wheeler transform. Bioinformatics, 25(14), 1754–1760. https://doi.org/10.1093/bioinformatics/btp324 Li,H.,&Durbin,R.(2010).Fastandaccuratelong-readalignmentwith Burrows–Wheeler transform. Bioinformatics, 26(5), 589–595. https://doi.org/10.1093/BIOINFORMATICS/BTP698 Li, H., Handsaker, B., Wysoker, A., Fennell, T., Ruan, J., Homer, N., Marth,G.,Abecasis,G.,Durbin,R.,&1000GenomeProjectData Processing Subgroup. (2009). The sequence alignment/map formatandSAMtools.Bioinformatics Applications, 25(16),2078–2079. https://doi.org/10.1093/bioinformatics/btp352 Lowe,A.,Jones,L.,Brennan,G.,Creer,S.,&deVere,N.(2022).Seasonal progressionanddifferences in majorfloralresourceuse by bees andhoverfliesinadiversehorticulturalandagriculturallandscape revealedbyDNAmetabarcoding.Journal of Applied Ecology, 59,1– 12. https://doi.org/10.1111/13652664.14144 Moran,N.A.(2015).Genomicsofthehoneybeemicrobiome.Current Opinion in Insect Science, 10, 22–28. https://doi.org/10.1016/j. cois.2015.04.003 Morse, R. A. (1975). Bees and beekeeping (1st ed.). Cornell University Press. Morse,R.A.(1994).The new complete guide to beekeeping.Countryman Press. Muñoz-Colmenero,M.,Baroja-Careaga,I.,Kovačić,M.,Filipi,J.,Puškadija, Z., Kezić, N., Estonba, A., Büchler, R., & Zarraonaindia, I. (2020). Differencesinhoneybeebacterialdiversityandcompositioninagriculturalandpristineenvironments–Afieldstudy.Apidologie, 51, 1018–1037.h t t p s : / / d o i . o r g / 1 0 . 1 0 0 7 / s 1 3 5 9 2 - 0 2 0 - 0 0 7 7 9 - w Nurk, S., Meleshko, D., Korobeynikov, A., & Pevzner, P. A. (2017). metaSPAdes: A new versatile metagenomic assembler. Genome Research, 27(5), 824–834. https://doi.org/10.1101/ GR.213959.116 Ovaskainen,O.,&Abrego,N.(2020).Joint species distribution modelling: With applications in R. Cambridge University Press. https://doi. org/10.1017/9781108591720 Ovaskainen,O.,Abrego,N.,Halme,P.,&Dunson,D.(2016).Usinglatent variable models to identify large networks of species-to-species associations at different spatial scales. Methods in Ecology and Evolution, 7,549–555.https://doi.org/10.1111/2041-210X.12501 Pellissier,L.,Albouy,C.,Bascompte,J.,Farwig,N.,Graham,C.,Loreau, M.,Maglianesi,M.A.,Melián,C.J.,Pitteloud,C.,Roslin,T.,Rohr, R., & Gravel, D. (2018). Comparing species interaction networks alongenvironmentalgradients.Biological Reviews, 93(2),785–800. https://doi.org/10.1111/BRV.12366 Peltonen-Sainio,P.,Rajala,A.,&Seppälä,R.(2005).Viljojen kehityksen ja kasvun ABC. MTT. Pryszcz,L.P.,&Gabaldón,T.(2016).Redundans:Anassemblypipeline for highly heterozygous genomes. Nucleic Acids Research, 44(12), e113. https://doi.org/10.1093/nar/gkw294 Rabeling,S.C.,LeLim,J.,Tidon,R.,Neff,J.L.,Simpson,B.B.,&Pawar, S. (2019). Seasonal variation of a plant-pollinator network in the Brazilian Cerrado: Implications for community structure and robustness. PLoS One, 14(12), e0224997. https://doi.org/10.1371/ JOURNAL.PONE.0224997 Rasmussen, C., Dupont, Y. L., Mosbacher, J. B., Trjøelsgaard, K., & Olesen,J.M.(2013).Strongimpactoftemporalresolutiononthe structureofanecologicalnetwork.PLoS One, 8(12),81694.https:// doi.org/10.1371/journal.pone.0081694 Rasmussen, N. L., Van Allen, B. G., & Rudolf, V. H. W. (2014). Linking phenologicalshiftstospeciesinteractionsthroughsize-mediated priorityeffects.Journal of Animal Ecology, 83(5),1215.https://doi. org/10.1111/13652656.12203 Raymann, K., & Moran, N. A. (2018). The role of the gut microbiome in health and disease of adult honey bee workers. Current Opinion in Insect Science, 26, 97–104. https://doi.org/10.1016/j. cois.2018.02.012 Rudolf, V. H. W. (2019). The role of seasonal timing and phenological shifts for species coexistence. Ecology Letters, 22(8), 1324–1338. https://doi.org/10.1111/ELE.13277 Runckel,C.,Flenniken,M.L.,Engel,J.C.,Ruby,J.G.,Ganem,D.,Andino, R.,&Derisi,J.L.(2011).Temporalanalysisofthehoneybeemicrobiomerevealsfournovelvirusesandseasonalprevalenceofknown viruses, Nosema, and Crithidia. PLoS One, 6(6),e20656.https://doi. org/10.1371/journ al.pone.0020656 Ruottinen, L., Ollikka, T., Vartiainen, H., & Seppälä, A. (2003). Mehiläishoitoa käytännössä osa 1.SuomenMehiläishoitajainLiittor.y. Salonen,A.,Ollikka,T.,Grönlund,E.,Ruottinen,L.,&Julkunen-Tiitto,R. (2009).PollenanalysesofhoneyfromFinland.Grana, 48(4),281– 289.https://doi.org/10.1080/00173130903363550 Subotic,S., Boddicker,A.M.,Nguyen,V.M., Rivers,J.,Briles,C.E.,& Mosier, A. C.(2019). Honeybeemicrobiomeassociatedwithdifferenthiveandsampletypesoverahoneyproductionseason.PLoS One, 14(11),1–15.https://doi.org/10.1371/journ al.pone.0223834 Tikhonov,G.,Opedal,Ø.H.,Abrego,N.,Lehikoinen,A.,deJonge,M.M. J.,Oksanen,J.,&Ovaskainen,O.(2020).Jointspeciesdistribution modellingwithther-packageHmsc.Methods in Ecology and Evolution, 11(3),442–447.https://doi.org/10.1111/2041-210X.13345 Tjur, T. (2009). Coefficients of determination in logistic regression models – A new proposal: The coefficient of discrimination. American Statistician, 63(4), 366–372. https://doi.org/10.1198/ tast.2009.08210 Vannette, R. L. (2020). The floral microbiome: Plant, pollinator, and microbial perspectives. Annual Review of Ecology, Evolution, and Systematics, 51, 363–386. https://doi.org/10.1146/ANNUREV- E C O L S Y S - 0 1 1 7 2 0 - 0 1 3 4 0 1 Vesterinen,E.J.,Puisto,A.I.E.,Blomberg,A.S.,&Lilley,T.M.(2018). Tableforfive,please:Dietarypartitioninginborealbats.Ecology and Evolution, 8(22), 10914–10937. https://doi.org/10.1002/ ece3.4559 Whittaker,R.H.,Levin,S.A.,&Root,R.B.(1973).Niche,habitat,and ecotope. The American Naturalist, 107(955),321–338. 20457758, 2023, 10, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10580 by University Of Jyväskylä Library, Wiley Online Library on [12/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 16 of 16 | WIRTA et al. Wirta,H.K.,Bahram,M.,Miller,K.,Roslin,T.,&Vesterinen,E.(2022). Reconstructingtheecosystemcontextofaspecies:Honey-borne DNArevealstherolesofthehoneybee.PLoS One, 17(7),e0268250. https://doi.org/10.1371/journ al.pone.0268250 Wray,M.K.,Mattila,H.R.,&Seeley,T.D.(2011).Collectivepersonalities inhoneybeecoloniesarelinkedtocolonyfitness.Animal Behaviour, 81(3),559.https://doi.org/10.1016/j.anbehav.2010.11.027 SUPPORTING INFORMATION Additional supporting information can be found online in the SupportingInformationsectionattheendofthisarticle. How to cite this article: Wirta,H.,Jones,M.,Peña-Aguilera, P.,Chacón-Duque,C.,Vesterinen,E.,Ovaskainen,O., Abrego,N.,&Roslin,T.(2023).Theroleofseasonalityin shapingtheinteractionsofhoneybeeswithothertaxa. Ecology and Evolution, 13, e10580. https://doi.org/10.1002/ ece3.10580 20457758, 2023, 10, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/ece3.10580 by University Of Jyväskylä Library, Wiley Online Library on [12/10/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License