Comments from curator
Phenotyping data comes from the
Supplementary Tables of the bioRxiv preprint of the paper (version 2 of 6 May 2022,
10.1101/2020.06.28.175356, file media-2.zip), which was published as
Sleep 46(4):zsad016. The data availability statement says that the data are available on request, but the per-fly and per-hemisphere tables are in this supplement:
-
S1 Table: raw sleep data, one row per fly (73 DGRP lines, females and males), uploaded as raw dataset 1.
-
S5 Table: raw mushroom body (MB) measurements, one row per brain hemisphere (40 DGRP lines, females), uploaded as raw dataset 2.
-
S4 Table: percentages of gross α- and β-lobe phenotypes per line (the same 40 lines, females), uploaded as the summary dataset.
The line means of S2 Table (sleep) and S6 Table (MB) are reproduced exactly by the means of the raw datasets, so they are not uploaded. The other tables hold variance analyses, GWAS results, gene lists and RNAi knockdown experiments (GAL4 > UAS-RNAi flies, not DGRP lines). The adult size measurements of the paper were made on Oregon-R flies only.
Design. Larvae of each DGRP line were reared from egg laying on normal food (
_NF: 10% brewer's yeast, 5% sucrose) or on restricted food (
_RF: 20% of the yeast and sucrose of normal food, i.e. 2% yeast and 1% sucrose), at 25 °C and 12 h light : 12 h dark. Newly eclosed adults (virgin females and males) were then kept on normal food for 3 days before the sleep recordings (the methods do not say whether this was also done before the MB dissections of the DGRP lines, made at 3-5 days), so the food suffix is the early-life (larval) diet. Each fly was reared on one food only: in each row, the phenotypes of the other food are NA.
Sleep (raw dataset 1: 6,468 flies of 73 lines). Single flies were recorded in DAM2 monitors (TriKinetics) for 7 days. Sleep is any period of at least 5 min without a beam crossing, and the values (computed with pySolo) are those of days 3 and 4 of the recording, given per day. Each replicate is 8 flies of each line, sex and food in one monitor; there are 1 to 5 replicates per line, sex and food (3 for 144 of the 292 groups; the paper says three for most lines), i.e. 6 to 40 flies. Flies that died during the recording were discarded by the authors (all flies of S1 Table are alive). The nine traits are:
-
TotalSleep: minutes of sleep per day;
-
DaySleep and
NightSleep: minutes of sleep in the 12 h light and dark phases;
-
WakeActivity: waking activity (beam crossings per minute awake);
-
DayBoutLen and
NightBoutLen: average sleep bout length (minutes) in the light and dark phases;
-
DayBoutNum and
NightBoutNum: number of sleep bouts in the light and dark phases (per day);
-
SleepLatency: minutes from lights off to the first sleep bout.
Replicate,
DAM_Monitor and
DAM_Channel are kept as obsolete design columns (monitor numbers are reused across replicates).
Mushroom bodies (raw dataset 2: 1,743 hemispheres of 40 lines, females only). Brains of 3-5 day old females were stained with anti-Fasciclin 2 (1D4) and imaged by confocal microscopy. The length and width of the α and β lobes of each hemisphere were measured with Fiji and divided by the distance between the α-lobe heels, so they have no unit (the authors give arbitrary units). There are 10 to 26 hemispheres per line and food: the paper says 20-24, but 13 of the 80 line and food groups have fewer than 20. A value of 0 means that the lobe was missing in that hemisphere (96 values: length and width of 48 lobes, in 41 hemispheres of 15 lines); these zeros are included in the authors' line means (S6 Table) and analyses (Table 2), so they are kept.
Brain (brain number within the line and food) and
Hemisphere (1 or 2) are kept as obsolete design columns; brain numbers are repeated within the line and food for DGRP_303 (RF), DGRP_306 (NF), DGRP_774 (NF) and DGRP_820 (RF), so they do not always identify a single brain.
Gross MB defects (summary dataset: 40 lines, females). Percentages of MBs with a normal, missing, thin or otherwise abnormal α lobe, and with a normal, fused, missing, thin or otherwise abnormal β lobe, per line and food (S4 Table), stored as means (mn_) with the unit %. The number of MBs scored per line is not given: the percentages imply up to 34 per line and food, sometimes more than the hemispheres measured in S5 Table (e.g. DGRP_712 on normal food: at least 34, vs 24 measured hemispheres). In S4 Table, the α-lobe categories of DGRP_707 on normal food sum to 99.94% (normal given as 97 for 97.06); mn_AlphaAbnormal_RF is 0 for all lines. The text of the paper says that 14 and 6 lines had fewer α- and β-lobe defects on restricted food, and 9 and 17 lines more; S4 Table gives 14 and 10 lines with fewer, and 10 and 18 lines with more (100% minus normal).
Lines. 74 DGRP lines in total, as in the methods: 73 for sleep and 40 for MB, 39 of them in both (DGRP_335 only has MB data). 36 of the 40 MB lines were also studied by Zwarts et al. 2015 (DGRPool study 38) on standard food; the data here are new measurements on both diets.
GWAS. The authors ran the DGRP2 web pipeline on the line means of each sleep trait, sex and food, and focused on the difference between the restricted and normal food line means (mean of X_RF minus mean of X_NF, per line and sex, from the uploaded phenotypes). Day bout length was left out of the GWAS (little genotype-by-nutrition interaction), and the results of night bout number and day bout number in females, day sleep in males, total sleep and waking activity in both sexes and all MB traits were not used (Q-Q plots). The top variants are in S8 Table.