Comments from curator
Phenotyping data come from the
Dryad deposit of the paper (
doi:10.5061/dryad.sn02v6x7w, CC0; files taken from its
Zenodo mirror, identical to the Dryad files):
4b_development_time_keep_outlier_rep1.csv to
rep6.csv, the authors' R input (one file per trial: number of adults of each sex, vial and pupation site collected at each check), with the collection times taken from the authors' R code (Chapter_I_09DEC2023.R). The original collection forms (Collected_data.xlsx) were used to check the counts. The
paper (
PMC) and its supplement give no per-line values (Fig. 1 only).
Design: 21 DGRP lines were reared at 25 °C (12:12 h light:dark). For each line, groups of 20 mated females laid eggs in up to 8 vials (fewer when a line had too few females) for 40 h (Trial I) or 24-32 h (later trials). One day after the females were removed, a dry cotton roll was pushed into the centre of the food; wandering larvae pupated either on the cotton roll or on the vial wall / food surface. Once most larvae had pupated, the rolls were moved to new food vials, and the adults emerging from both vials were sexed and counted 1-3 times per day. Trials I-IV tested 21, 20, 17 and 18 of the lines (DGRP_307 only in Trial I; vial positions randomized from Trial II on; stocks re-ordered from Bloomington before Trial IV). Trials V and VI re-tested 6 lines chosen from Trials I-IV for their sex difference in pupation site (more males on cotton: DGRP_304 and 362; more females: 427 and 732; no difference: 335 and 517; 517 was not in Trial VI).
Uploaded data: the raw unit is
one adult fly: the per-vial counts were expanded into one row per fly (71,259 flies in all, 49.6% males), with its sex, its development time and its pupation site.
-
Raw dataset 1 (Trials I-IV, the paper's main experiment, Table 1 and Fig. 1): 60,827 flies, 21 lines.
DevTime (hours) is the estimated egg-to-adult time: from the midpoint of the egg-laying window to the collection (first collection) or to the midpoint between the previous and the current collection, as in the authors' R code.
CottonPupation is 1 for a fly that pupated on the cotton roll and 0 otherwise, so its line mean is the proportion of flies pupating on cotton.
Trial (1-4) and
Vial (1-8 within line and trial) are design columns (obsolete); the number of rows of a vial is its density (adults emerged), the covariate of the authors' models.
-
Raw dataset 2 (Trials V-VI, 6 lines): 10,432 flies, same phenotypes with the suffix _TrialV_VI. They are kept apart because these lines were selected on their Trials I-IV results and tested for repeatability under different conditions (30-32 h of egg laying, drier food in Trial VI, eclosion from 205 h instead of 240-273 h): pooling them with Trials I-IV would shift the means of these 6 lines only.
Exclusions: all flies are kept. The authors analysed per-vial differences between males and females and removed for that the 4 vials with only one sex (623 vials left) and, for development time, 5 outlier vials (difference beyond the mean ± 3 SD; 618 left), and for pupation site 1 more vial (Trial III, DGRP_313). These exclusions only concern the sex differences, so the flies of these vials are kept here.
Checks: the files reproduce the numbers of the paper and of the R code: 71,259 flies (paper: ~71,000; 49.6% males), 627 vials with flies, 623 with both sexes; vial means of development time 286.3 h (SD 26.7) for males and 282.5 h (SD 27.1) for females; flies pupating on cotton emerged 8.6 h (males) and 7.8 h (females) later; on average 24.2% of the flies of a vial pupated on cotton; 114 adults per vial (SD 74.3; paper 74.2) and 819 per line and trial. In 19/21 lines, males took longer than females in Trials I-IV (DGRP_786 and 360 are the exceptions), as in the paper. The line numbers of the files are DGRP numbers (303 = DGRP_303); all 21 lines are in DGRPool.
Sexual dimorphism: the paper's responses are, per vial, the male mean minus the female mean (development time; proportion on cotton), modelled with density level, trial condition and number of vials per line (Table 1; Fig. 1 shows model estimates and observed differences per line). They can be derived from the raw datasets (group by Trial and Vial). The sex difference in pupation site depends strongly on density, as the paper says: with the per-sex line means of DGRPool, more males than females pupated on cotton in 10 of the 21 lines; with the means of the per-vial differences, in 13 lines (DGRP_732, 427 and 303 being the most female-biased, as in the paper); the paper's 18/21 lines are after removing the density and vial-number effects.
Not available: wing length, wing width and abdominal sternite bristle numbers of ~300 pairs of flies from the 6 lines of Trials V-VI (flies of Trial III) were measured, but their data file (wingsabdomen_1.csv, read by the R code) is not in the deposit; the supplement only gives pooled t-tests (Supplementary Table S1). They are not uploaded.