Comments from curator
Phenotyping data comes from
Supplementary Data 1, the source data file of the paper (
42003_2020_1617_MOESM6_ESM.xlsx;
journal page, also on
PMC):
- sheet
fig 5 a: number of lunges in 20 min of each pair of males, 12 to 24 pairs per line. Used for raw dataset 1 (
Lunges_20min, one row per pair).
- sheet
fig 5 c: locomotor activity (velocity, mm/s) during the 20-min recording, for 10 lines, 18 to 24 values per line. Used for raw dataset 2 (
Velocity_20min, one row per value).
The other sheets only have the Canton-S control strain and a hyper-aggressive strain derived from it (not DGRP lines). All flies are males (sex M). The values were not transformed.
Divider Assay: on the day of eclosion, males were loaded into a 3D-printed chamber of 12 square arenas (13 × 13 mm, 4.5 mm high) set on clear food, one fly on each side of an opaque divider. After 5 days of isolation the dividers were removed, and the 12 pairs (two males of the same line) were videotaped for 20 min, within the first 3 h of the light phase, at ~22.5 °C and ~50% humidity (flies reared on cornmeal-molasses food, 16 h light/8 h dark). The flies were tracked with FlyTracker, and lunges were scored with the authors' JAABA lunge classifier.
Lunges_20min is the number of lunges of the pair.
Velocity_20min is the average velocity over the 20-min recording (mm/s, the unit of Fig. 5c and Supplementary Fig. 4), measured for the 5 lines with the lowest and the 5 lines with the highest mean aggression score (MAS). The file does not say whether a value is one fly or one pair (most likely one fly). The number of values of a line differs from its number of lunge pairs, so the two phenotypes are in separate datasets and cannot be matched row by row. The Canton-S and hyper-aggressive controls of Fig. 5c are not in the file.
Lines: the authors chose 50 DGRP lines spanning the range of the MAS of Shorter et al. 2015 (
DGRPool study 30) and report lunge numbers for 47 of them (Methods list and Fig. 5a; some lines grew poorly). The source file has 49 line columns:
-
DGRP_395 (23 pairs) is in the file but not in the paper. It is uploaded, but it is not part of the paper's analyses.
- Line
471 (24 pairs, mean 0.33 lunges) is in the file but not in the paper, and there is no DGRP line 471 (RAL-471) in DGRPool. It was left out.
In the Methods, RAL 313 is given the Bloomington number of RAL 361 (BL#28180; RAL 313 is BL#25180).
Comparison with the paper: as stated in the paper, no line has a mean above 15 lunges in 20 min (highest: DGRP_392 14.2, DGRP_530 14.2, DGRP_352 13.2). The line medians computed from the raw data match the boxes of Fig. 5a. The mean markers (green squares) of Fig. 5a are drawn about 7% higher than the means of the source data, with the lines in the same order, which looks like a plotting offset. The paper found no correlation between the mean lunge number and the MAS (Fig. 5b, R
2 = 0.0002, n = 47 lines; it can be redone with the means of Lunges_20min and the MAS of study 30), and most of the 10 lines moved less than Canton-S (Fig. 5c). Videos are available from the authors on request.