Dataset DOI: 10.5061/dryad.qjq2bvqv2

Description of the data and file structure

Data overview

This section contains information related to the datasets, scripts, and figures used to analyse and visualise the variation in the expression of host and parasite traits across space:

1. "behaviour" traits: We used behavioural assays to evaluate the behavioural response of T. longispinosus subcolonies to different intruders (i.e. T. americanus workers), and intruder themselves, across space. 

   Of the 450 videos, a representative subset of videos (N = 30; one video per treatment (n = 3) / per population (n = 10)) was selected and stored on Dryad to represent the behavioural assays, as the total video volume was very large (<3 TB). Each video is named with the random number of the host colony, followed by an “_”, and then the number of the intruder’s source colony (e.g., 7_219_noaudio.mp4, where the host colony = 7 and the intruder colony = 219). The data corresponding to these videos can be retrieved from the file behaviour_data.csv using the columns ID_rand_subcol (host colony) and ID_col_intrud (intruder colony)

2.  "CHCs" traits: We characterised the CHC profiles of these species to determine changes of composition within and between species across space.

3. "algorithm": stepwise reduction algorithm used to identify sets of CHCs related to host behavioural aggression from hosts towards conspecifics and social parasite.

4. "colony_structure": We tested how the structure of T. longispinosus and T. americanus colonies varied across space.

Description of the data and file structure

• *_data.csv: Datasets containing the information necessary to perform analyses and generate figures

  *_analyses.Rmd, *_analyses.R: Scripts used to run statistical models based on the datasets (R version 4.2.3)

  *_plots.R: Scripts used to visualise figures (R version 4.2.3)

Example for behavioural traits

• Dataset: behaviour_data.csv
• Analyses script: behaviour_analyses.Rmd
• Figure script: behaviour_plots.R

The file predictors_data.csv contains information about the collection sites (parasite prevalence, climate). It was used by the scripts analysing colony_structure and CHCs traits. It was not used for the behavioural analyses because behaviour_data.csv already included the relevant predictor columns.

For the analyses script algorithm_analyses.Rmd, two datasets were required to calculate which CHCs are associated with host aggression:

1. behaviour_algorithm_data.csv, extracted from the script behaviour_analyses.Rmd (filtered from the original behaviour_data.csv dataset).
2. CHC_data.csv, also used for the CHCs analyses.

Files and variables

File: behaviour_algorithm_data.csv

Description: Filtered dataset used for the stepwise reduction algorithm that identified sets of CHCs related to host behavioural aggression from hosts towards conspecifics and social parasite.

Variables

• ID_subcol_project: Colony ID from the subcolony used across different datasets. It contains all information of the colony in that form: X (species: SM for T. americanus or L for T. longispinosus)_XX (population origin)_X (colony number).

• ID_intrud_project: Colony ID from the intruder used across different datasets. It contains all information of the colony in that form: X (species: SM or L) _XX (population origin)_X (colony number).

• week: The week in which the behavioural assay was recorded (either week 1,2 or 3).
• type: Type of intruder placed in the arena.
• prevalence: Parasite prevalence estimates based on the 10 collection sites.
• population: Sites from which colonies were collected, in reference to the states in Northern America.

• subcol_aggression_prop: Probability of a T. longispinosus worker to respond aggressively towards an intruder.

 

File: colony_structure_data.csv

Description: Each row of the dataset represents the composition of a colony at the time of collection.

Variables

• population: Sites from which colonies were collected, in reference to the states in Northern America.
• Colony_ID: Unique number given to a colony as identifier.
• Slave_Maker: specifies whether a colony is dulotic based on the term “T. americanus”; an empty cell indicates a free-living colony.
• Host1: If filled, contains the identity of the first host species in a nest. A colony is considered free-living if the column Slave_maker is empty, otherwise it refers to the identity of enslaved hosts.
• Host2: If filled, contains the identity of the second host species in a nest. A colony is considered free-living if the column Slave_maker is empty, otherwise it refers to the identity of enslaved hosts.
• SM_queen: gives the number of dulotic queens in the nest.
• SM_worker: provides the number of dulotic workers in the nest.
• Host1_queen: records the numbers of queens of the first host species.
• Host1_worker: records the numbers of workers of the first host species.
• Host2_queen: records the numbers of queens of the second host species.
• Host2_worker: records the numbers of workers of the second host species.

 

File: CHCs_data.csv

Description: Each row of the dataset represents one chemical profile (one worker for the parasite T. americanus, two pooled workers for the host T. longispinosus) from a colony replicate.

Variables

• ReplicateID: indicates the replicate number within each population for each species.
• Filename: corresponds to the raw colony identifier based on the GC-MS data file of the processed sample.
• ID_project: Unique colony identifier combining the species code (L for T. longispinosus and SM for T. americanus), the population origin, and the replicate number; this identifier is also consistently used in behavioural analyses.
• Species: refers to the species identity, either Temnothorax longispinosus or T. americanus.
• population: provides the geographic origin of the colony.
• C29-en: Known CHC compound
• 5-MeC27: Known CHC compound
• 3-MeC27: Known CHC compound
• 4-MeC28: Known CHC compound
• 5-MeC29: Known CHC compound
• 3-MeC29: Known CHC compound
• 7-MeC31: Known CHC compound
• 11,15; 13,17-DiMeC31, traces of 5-MeC31: Known CHC compound
• 3-MeC31: Known CHC compound
• 5,x-DiMeC33: Known CHC compound
• 11-;13-;15-MeC27 : Known CHC compound
• 7-MeC27: Known CHC compound
• 11,15-DiMeC27 (identification tentative): Known CHC compound
• 12-;13-MeC28 : Known CHC compound
• 3-MeC28: Known CHC compound
• 11-;13-;15-MeC29 : Known CHC compound
• 7-MeC29: Known CHC compound
• 11,15-;13,17-DiMeC29 : Known CHC compound
• 5,x-DiMeC29: Known CHC compound
• 3,7-DiMeC29, possibly other 3,x-DiMeC29: Known CHC compound
• 11-;12-;13-;14-MeC30 : Known CHC compound
• 3-MeC30: Known CHC compound
• x,y-DiMeC30: Known CHC compound
• 9-;11-;13-;15-MeC31 : Known CHC compound
• 7,11-DiMeC31: Known CHC compound
• 11,15,19-TriMeC31: Known CHC compound
• 7,15,21-TriMeC31: Known CHC compound
• 3,11-DiMeC31, traces of other 3,x-DiMeC31: Known CHC compound
• 3,25-DiMeC31 (identification tentative): Known CHC compound
• 3,x,y-TriMeC31 (identification tentative): Known CHC compound
• 11-;13-;15-;17-MeC33 : Known CHC compound
• 13,17-;15,19-DiMeC33, possibly also 11,21-DiMeC33: Known CHC compound
• 3,9-;3,11-DiMeC33 : Known CHC compound
• 11-;13-;15-MeC35 : Known CHC compound
• 11,x-;13,x-;15,x-DiMeC35 : Known CHC compound
• 11,x-;13,x-;15,x-DiMeC37 : Known CHC compound
• n-C25: Known CHC compound
• n-C26: Known CHC compound
• n-C27: Known CHC compound
• n-C28: Known CHC compound
• n-C29: Known CHC compound
• n-C30: Known CHC compound
• n-C31: Known CHC compound
• unknown_RT19.90/RI27.81: Unknown CHC compound
• unknown_RT20.34/RI28.09: Unknown CHC compound
• unknown_RT20.74/RI28.33: Unknown CHC compound
• unknown_RT20.88/RI28.41: Unknown CHC compound
• unknown_RT21.02/RI28.50: Unknown CHC compound
• unknown_RT23.98/RI30.38: Unknown CHC compound
• unknown_RT24.50/RI30.71: Unknown CHC compound
• unknown_RT25.66/RI31.48: Unknown CHC compound
• unknown_RT26.38/RI31.97: Unknown CHC compound
• unknown_RT26.56/RI32.09: Unknown CHC compound
• unknown_RT26.98/RI32.37: Unknown CHC compound
• unknown_RT27.14/RI32.47: Unknown CHC compound
• unknown_RT27.28/RI32.56: Unknown CHC compound
• unknown_RT27.34/RI32.60: Unknown CHC compound
• unknown_RT27.84/RI32.93: Unknown CHC compound
• unknown_RT28.44/RI33.35: Unknown CHC compound
• unknown_RT29.16/RI33.85: Unknown CHC compound
• unknown_RT29.48/RI34.08: Unknown CHC compound
• unknown_RT29.70/RI34.25: Unknown CHC compound
• unknown_RT29.78/RI34.31: Unknown CHC compound
• unknown_RT29.88/RI34.38: Unknown CHC compound
• unknown_RT30.06/RI34.52: Unknown CHC compound
• unknown_RT30.36/RI34.74: Unknown CHC compound
• unknown_RT30.50/RI34.85: Unknown CHC compound
• unknown_RT30.66/RI34.97: Unknown CHC compound
• unknown_RT31.22/RI35.38: Unknown CHC compound
• unknown_RT31.58/RI35.63: Unknown CHC compound
• unknown_RT31.74/RI35.75: Unknown CHC compound
• unknown_RT31.88/RI35.85: Unknown CHC compound
• unknown_RT32.08/RI36.00: Unknown CHC compound
• unknown_RT32.16/RI36.05: Unknown CHC compound
• unknown_RT32.30/RI36.15: Unknown CHC compound
• unknown_RT32.48/RI36.29: Unknown CHC compound
• unknown_RT32.78/RI36.50: Unknown CHC compound
• unknown_RT33.06/RI36.71: Unknown CHC compound
• unknown_RT33.34/RI36.91: Unknown CHC compound
• unknown_RT33.70/RI37.18: Unknown CHC compound
• unknown_RT33.80/RI37.28: Unknown CHC compound
• unknown_RT34.34/RI37.69: Unknown CHC compound
• unknown_RT34.68/RI37.96: Unknown CHC compound
• IS_RT7.02: Internal Standard

 

File: behaviour_data.csv

Description: Dataset used to test behavioural variation in subcolonies and intruders. Each row represents a behavioural assay.

Variables

• ID_col_subcol: Unique subcolony ID used during the collection for all colonies (parasite and host).
• ID_rand_subcol: Subcolony random ID, from 1 to 150.
• ID_col_intrud: Unique intruder colony ID used during the collection for all colonies (Parasite and host).
• ID_subcol_project: Colony ID from the subcolony used across different datasets. It contains all information of the colony in that form: X (species: SM or L)_XX (population origin)_X (colony number).
• ID_intrud_project: Colony ID from the intruder used across different datasets. It contains all information of the colony in that form: X (species: SM or L) _XX (population origin)_X (colony number).
• week: The week in which the behavioural assay was recorded (either week 1,2 or 3).
• type: Type of intruder placed in the arena.
• prevalence: Parasite prevalence estimates based on the 10 collection sites.
• population: Sites from which colonies were collected, in reference to the states in Northern America.
• PC1_climate: Values of the 1st PC axis from the population climate data spanning over temperature and precipitations.
• day: The day the behavioural assays was recorded.
• start_assay: The time of the day the behavioural assay record started.
• n_queen_subcol: Number of queens in the arena at the beginning of the video (1 is expected, except for queenless colonies).
• n_workers_subcol: Number of workers in the arena at the beginning of the video before introduction.
• subcol_mandibles: Number of host workers that open their mandibles as a threat response around an introduced worker.
• subcol_biting: Number of host workers that grab with their mandibles the libs of an introduced worker without dragging.
• subcol_stinging: Number of host workers that bit and curve the thorax and stab with their stinger an introduced worker.
• subcol_dragging: Number of hosts workers that grab the libs of an introduced worker and drag it in a direction.
• subcol_carrying: Number of host workers that lift an introduced worker with their mandibles while moving around.
• subcol_pinning: Number of host workers dragging in opposite directions the legs of an introduced worker.
• subcol_antennation: Number of host workers that touch bodyparts of an introduced worker with their antenna.
• subcol_grooming: Number of host workers that gently use their mouthparts on the introduced worker while touching it with the antenna.
• subcol_trophallaxis: Number of host workers showing a mutual mouthpart contact with an introduced ant.
• subcol_brood_carrying: Number of host workers that carry the brood with their mandibles.
• intrud_mandibles: An introduced worker that opens its mandibles around another worker (threat behaviour) .
• intrud_biting: An introduced worker that grabs with its mandibles the libs of another worker without dragging.
• intrud_stinging: An introduced worker that bits and curves its thorax and stabs with its stinger another worker.
• intrud_dragging: An introduced worker that grabs the libs of another worker and drags it in a direction.
• intrud_carrying: An introduced worker that lifts another worker with its mandibles while moving around.
• intrud_antennation: An introduced worker that touches bodyparts of another worker with its antenna.
• intrud_grooming: An introduced worker that gently uses its mouthparts on another worker while touching it with its antenna.
• intrud_trophallaxis: An introduced worker showing a mutual mouthpart contact with another ants.
• intrud_immobile: An introduced worker that stands on its legs, inactive during an interaction with another worker.
• intrud_curled: An introduced worker that adopts a pupae like posture, where the legs are back hidden close to the thorax and the antenna close to the head in an interaction with another worker.
• corpse_begin: Number of dead workers found on the arena floor at the beginning of the video before the introduction of the worker.

** **

File: predictors_data.csv

Description: Each row of the dataset represents one site where colonies were collected. The column population indicates the geographic origin of the colonies. The variables PC1 to PC5 correspond to principal component values summarizing climatic conditions at each site. The column prevalence provides the parasite prevalence values associated with the respective population.

Variables

• population: Sites from which colonies were collected, in reference to the states in Northern America.
• PC1: Principal Component 1 eigenvalues of climate CHELSA Bio-database.
• PC2: Principal Component 2 eigenvalues of climate CHELSA Bio-database.
• PC3: Principal Component 3 eigenvalues of climate CHELSA Bio-database.
• PC4: Principal Component 4 eigenvalues of climate CHELSA Bio-database.
• PC5: Principal Component 5 eigenvalues of climate CHELSA Bio-database.
• prevalence: Parasite prevalence estimates based on the 10 collection sites.

 

File: algorithm_analyses.R

Description: Script that calculates which CHCs are associated with host aggression based on 1) behaviour_algorithm_data.csv, extracted from the script behaviour.Rmd, and 2) CHC_data.csv, also used for the CHC analyses.

File: functions.R

Description: Script used in combination with algorithm_analyses.R, necessary to run the reductive algorithm.

File: behaviour_analyses.Rmd

Description: Script used to analyse behaviour_data.csv. From this script was extracted behaviour_algorithm_data.csv then used for the script algorithm_analyses.R.

File: CHCs_americanus_analyses.Rmd

Description: Script used to perform CHC analyses on the parasite Temnothorax americanus using the CHC_data.csv dataset.

File: CHCs_longispinosus_analyses.Rmd

Description: Script used to perform CHC analyses for the host species Temnothorax longispinosus using the CHC_data.csv dataset.

File: CHCs_bothspecies_analyses.Rmd

Description: Script used to perform comparative CHC analyses between the host and the parasite using the CHC_data.csv dataset.

File: colony_structure_analyses.Rmd

Description: Script used to perform colony structure analyses using the colony_structure.csv dataset.

File: behaviour_plots.R

Description: Script used to perform behavioural analyses using the behaviour_data.csv dataset.

File: CHCs_plots.R

Description: Script used to generate all plots from the analyses in CHCs_longispinosus_analyses.Rmd, CHCs_americanus_analyses.Rmd, and CHCs_bothspecies_analyses.Rmd.

File: colony_structure_plots.R

Description: Script used to generate all plots from the analyses in colony_structure_analyses.Rmd.

File: Wilcoxon_parasite_prevalence.R

Description: Script used to compare long-term prevalence data with the 2021 collection using a Wilcoxon signed-rank test for paired samples.

Files: Reference behavioural videos

• 10_877_noaudio.MP4
• 105_1977_noaudio.MP4
• 115_1838_noaudio.MP4
• 124_1946_noaudio.MP4
• 129_1533_noaudio.MP4
• 130_1563_noaudio.MP4
• 141_2183_noaudio.MP4
• 143_1924_noaudio.MP4
• 149_438_noaudio_noaudio.MP4
• 150_1576_noaudio.MP4
• 24_835_noaudio.MP4
• 25_389_noaudio.MP4
• 28_899_noaudio.MP4
• 30_1681_noaudio.MP4
• 32_1346_noaudio.MP4
• 33_422_noaudio.MP4
• 37_2136_noaudio.MP4
• 4_1178_noaudio.MP4
• 40_2115_noaudio.MP4
• 6_1833_noaudio.MP4
• 61_1893_noaudio.MP4
• 64_73_noaudio.MP4
• 7_219_noaudio.MP4
• 73_744_noaudio.MP4
• 75_576_noaudio.MP4
• 81_1805_noaudio.MP4
• 83_1045_noaudio.MP4
• 9_1628_noaudio.MP4
• 91_1243_noaudio.MP4
• 94_99_noaudio.MP4

Code/software

Models

We performed generalised linear mixed models (GLMMs) with the glmmTMB package in R v.4.2.3. We ensured the correct distribution of the models with the diagnostic tools provided by the DHARMa package. 

In addition for the CHC analyses, we performed Permutational Multivariate Analyses of Variance (PERMANOVAs) using the adonis2 function from the vegan package to examine links between the composition of overall CHC profiles, n-alkanes, or recognition cues, and parasite prevalence or climate.

Packages used in analyses: "_analyses.Rmd" files

Behavioural analyses were conducted using tidyverse (data filtering and manipulation), glmmTMB (generalised linear mixed models with random effects), DHARMa (residual diagnostics), buildmer (stepwise model selection), emmeans (post hoc comparisons and joint tests), and performance (assessment of collinearity, singularity, and convergence).

CHC analyses were conducted using tidyverse (data filtering and manipulation), vegan (multivariate analyses based on Bray–Curtis distances), glmmTMB (generalised linear mixed models with random effects), performance (model diagnostics), DHARMa (residual diagnostics), ggeffects (extraction of model estimates for plotting), emmeans (post hoc comparisons and joint tests), and ggplot2 (visualisation). In addition, for the script CHCs_bothspecies_analyses.Rmd, reshape2 was used to build the heatmap of chemical presence/absence between species.

Colony structure analyses were conducted using tidyverse (data filtering and manipulation), glmmTMB (generalised linear mixed models with random effects), performance (model diagnostics), DHARMa (residual diagnostics), emmeans (post hoc comparisons and joint tests), ggeffects (extraction of model predictions for plotting), and ggplot2 (visualisation). Figures were assembled with gridExtra.

Figure visualization: "_plots.R" files

For each analyses.Rmd script, there is a corresponding _plots.R script. To generate the figures, the analysis script must be run first, as the plotting scripts require the fitted models to be loaded.

Figure visualisations were produced in R using ggplot2 (plotting and custom themes), ggeffects (extraction of model predictions for plotting, used for behavioural and colony structure figures), and gridExtra (assembling multi-panel figures). For CHC figures (script: CHCs_plots.R), additional packages included ggvenn and eulerr (set-overlap diagrams) and VennDiagram (alternative Venn rendering).