Comments from curator
Phenotyping data come from
Table S1 of the paper's Supporting Information (evo0068-1871-SD1.docx,
PMC4257566;
journal page) and from the authors' Dryad deposit (
doi:10.5061/dryad.d4s1k). Table S1 gives, for each of the 50 DGRP lines, the incidence of tapping by the DGRP male and by the focal male paired with it.
Design: in each 3-min trial, a virgin 6-8-day-old male of a DGRP line (the "interacting" male) was paired, in a horizontal 16 × 95 mm vial, with a virgin 3-5-day-old male of a common laboratory strain carrying the
yellow-body mutation on a wild-type background (Hmr2, Bloomington Stock Center; the "focal" male), so that the two males could be told apart. Three evenly spaced 1-min observations were made per pair (08:00-13:00, 19.4-24.9 °C) by the same observer, and the occurrence (yes/no) of each behaviour was recorded for each male: tapping (a male brings the tarsus of a foreleg into contact with the other fly), orienting, following, licking, courtship singing and abdomen curling, plus general activity (total number of behaviours). Flies were reared at 23 °C (12:12 h light:dark), collected as virgins and kept individually until the trial. 40 trials were run per line (2,000 trials; 5 run at about 17 °C were discarded, leaving 1,995), and 8 lines (DGRP_038, 075, 149, 223, 321, 336, 373, 897) were re-tested blind with 40 more trials each, giving 2,315 trials. All phenotypes are of males (sex M).
Indirect genetic effect:
mn_FocalTapping and
FocalTapping are phenotypes of the focal
yellow-body male, not of the DGRP fly. Focal males all came from the same strain, so the variation of their tapping among DGRP partners (from 10% to 54% of trials) is an indirect genetic effect of the DGRP genotype. It is stored under the DGRP partner line, as in the authors' GWAS (line means, one ANOVA per SNP with the DGRP website's GWAS tool; 13 SNPs with P ≤ 10
-5, the top one near
slowpoke). The DGRP male's own tapping (mn_Tapping) did not predict focal male tapping (Spearman r = 0.048, P = 0.741). In a multiple regression on line means, only the male startle response of the DGRP line (from Mackay et al. 2012,
DGRPool study 21) did (ψ = −0.486): focal males tapped less the males that took longer to settle after being startled.
Summary dataset, Table S1: incidence = proportion of trials in which the male tapped, as published (4 decimals). Each value is a whole number of trials out of 40 (39 for DGRP_021, 088, 392 and 639, one trial discarded each), or out of 80 for the 8 re-tested lines (79 for DGRP_336): as stated under Table S1, the values of these lines combine the original and validation trials. They reproduce the paper: 2,315 trials, focal males tapped in 28.16% of them, and r = 0.048 (P = 0.741) between the two columns. mn_FocalTapping is
obsolete, as the per-trial raw data (raw dataset 1) reproduce it.
Raw dataset 1 (Tmodelisoall_SAS_DATA.txt, "original plus validation data combined"): one row per trial (2,315 trials, 50 lines).
FocalTapping is 1 if the focal male tapped the DGRP male during the trial, 0 if not; DGRPool's line means are the incidences of Table S1. The file also has 5 trials with a missing response (run at about 17 °C, discarded by the authors): not uploaded.
TrialTemp (°C), the covariate of the authors' models, is kept as an obsolete column, and so is
Block (0 = original screen, 1 = blind validation re-test); for the 8 re-tested lines, the rows come from Tmodelisovalidation_SAS_DATA.txt, which holds the same trials with their block. Of note, 4 trials of DGRP_301 and 365 were run at 19.3 °C (the paper says 19.4-24.9 °C); one trial of DGRP_336 is at 234 °C in the combined file and 23.4 °C in the validation file, the value uploaded. The per-trial data of the DGRP male's behaviours were not deposited (only their line means).
Summary dataset, line means of the DGRP male's other behaviours (Line means for multiple regression.xlsx, values as in the file): mn_Orienting, mn_Following, mn_Licking, mn_Singing, mn_AbdCurling, mn_Activity. In this file, the line is called "Focal line" and the behaviours of the DGRP male are the "X focal (ISO)" columns (its "T focal (ISO)" column is the DGRP male's tapping of Table S1, and "Isolated T model" the focal male's). Incidences are proportions of trials, like the tapping of Table S1;
mn_Activity is the mean number of behaviours of any type per trial (counted over the three 1-min observations). The behaviour called abdomen curling in the methods is called "Mounting" in Table 1 of the paper (column "M focal (ISO)"). The file's line means of chill coma recovery, survival on menadione sodium bisulfite and paraquat, startle response and starvation resistance were taken from the DGRP website (Mackay et al. 2012, Weber et al. 2012;
DGRPool study 21 and
study 36) and are not uploaded, nor is its
Wolbachia column.
Raw dataset 2 (Startlevalidation_SAS_DATA.txt, File S2): startle response of 163 males of 6 lines (one row per fly), re-measured blind in this study to check the startle values of Mackay et al. 2012 between laboratories: the 3 lines of the study with the highest (DGRP_093, 358, 897) and the 3 with the lowest (DGRP_405, 535, 805) startle response in Mackay et al. 2012. Startle response = total time (s) the fly was mobile in the 45 s after its vial was tapped on the bench (males 3-7 days old, 25 °C, all tested on one day). The file's block 0 rows are the Mackay et al. 2012 values (DGRPool study 21) and were not uploaded. File S2 names RAL_353 instead of RAL_535 among the low lines, but the data file has 535, one of the 50 lines of Table S1. Medians per line in this study: DGRP_093 33 s, 358 15 s, 897 42 s, 405 8.5 s, 535 25.5 s, 805 22 s (one line, DGRP_358, switched from high to low, as File S2 reports).
Not uploaded: the
Wolbachia infection status of the lines (a DGRP annotation, used for 43 of the 50 lines; no effect on focal male tapping) and the GWAS results (Figure 6 only).