Study

Curated
Romano et al., 2022
Host genetics and pathogen species modulate infection-induced changes in social aggregation behaviour
[10.1098/rsbl.2022.0233]

Description

Social aggregation (median nearest-neighbour distance within same-sex groups of 12 flies) was measured in males and females of 10 DGRP lines pricked with Pseudomonas entomophila or sterile medium. Infected females, but not males, aggregated more closely.

Comments from curator

Phenotyping data comes from the authors' Zenodo deposit doi:10.5281/zenodo.6554320 (record 6554321, CC-BY 4.0), file NND-Boyle.csv: one row per social group of experiment 2 (466 groups), with the DGRP line, the median nearest-neighbour distance of the group (mm), the day of the assay, the sex and the infection status. Rearing and infection details are in the electronic supplementary material.

The paper has two experiments. Experiment 1 (effect of pathogen species and dose: mixed groups of infected and susceptible females, pairwise distances every 30 min for 4 h; file DistanceInds.xlsx) used a large outbred population derived from 113 DGRP lines, not DGRP lines, and is not uploaded. Experiment 2 (host variation) used males and females of 10 DGRP lines (RAL-21, 28, 208, 304, 324, 358, 375, 427, 712 and 852), chosen to span a range of sociality scores in Anderson et al. 2016. Flies were reared in vials at 25 ± 1°C, 12L:12D, on cornmeal-sugar medium. Mated flies 2-4 days old were pricked with a stainless steel pin dipped either in a culture of Pseudomonas entomophila diluted in LB to OD 0.1 (infected; suffix _Pe) or in sterile LB (control; suffix _Ctrl), lightly anaesthetised with CO2 and placed in groups of 12 flies of the same line, sex and treatment in a 55 mm Petri dish with agar. After a 30-min habituation the dish was photographed once, the thorax of each fly was marked in Fiji, and the nearest-neighbour distance (NND) of each fly was computed with the Fiji NND plugin. The phenotype is the median NND of the group, in mm (higher = less aggregated), the variable of Fig 2 and of the paper's model (which used its log10).

Raw dataset 1 (233 rows): each line, sex and treatment was tested in 9 to 13 groups, with exactly one control group and one infected group per line, sex and assay day. Each row therefore holds the two groups of one line and sex tested on the same day: MedNND_Ctrl (median NND of the control group) and MedNND_Pe (median NND of the infected group). They are different flies; the row only reflects the shared day. The day of the assay (1-17, the random effect of the authors' model) is kept as the obsolete design column SA_Day. Males and females were tested in separate groups (sex column). Values are copied as written in the authors' file; no group was dropped.

Checks: the file has the 466 groups of the paper. The authors' model (from their R script on Zenodo: lmer of log10 median NND on line × sex × infection status with the assay day as random effect, type II F tests) fitted on these data gives exactly the F values and p-values of Table 2 (line F = 4.676, sex F = 4.758, sex × infection F = 4.959, ...) and the least-square-means contrasts of ESM Table S2.

Of note:
- The paper says each line, sex and treatment was replicated 11-12 times, but the file has 9 (DGRP_028 and DGRP_852 males) to 13 groups (DGRP_021 and DGRP_852 females, DGRP_375 males); the total (466) is as in the paper.
- The paper says that infected females aggregated more closely than infected males "by 1.15 mm". This is the back-transformed least-square-means difference on the log10 scale (100.05936 = 1.15, a ratio, as in the authors' script), not a distance: the raw means of the infected groups are 4.99 mm (females) and 5.58 mm (males).
- The text gives p = 0.001 for the line effect, Table 2 gives p < 0.0001 (reproduced: 6.3 × 10-6).

The same lab measured social aggregation (median NND) of 10 other DGRP lines after Drosophila C virus infection in DGRPool study 121 (Siva-Jothy & Vale 2019); there the flies were tested 72 h after infection.

Other external data available

10.5281/zenodo.6554320 Zenodo deposit of the paper (per-group nearest-neighbour distances of experiment 2, uploaded; pairwise distances of experiment 1 in a DGRP-derived outbred population, not uploaded; the authors' R script)

2 categories

Behaviour Immunity

2 phenotypes

MedNND_Ctrl MedNND_Pe

Datasets

Dataset Phenotypes DGRP lines
Raw dataset #1 2 phenotypes
MedNND_Ctrl MedNND_Pe
10 DGRP lines
DGRP_021
DGRP_028
DGRP_208
DGRP_304
DGRP_324
DGRP_358
DGRP_375
DGRP_427
DGRP_712
DGRP_852
Download
Back to studies