Comments from curator
Phenotyping data come from
Table S4 of the electronic supplementary material (rspb20212707_si_001.pdf,
figshare;
journal page,
PMC9091837), which gives the mean lifespan estimated by the authors, with its 95% credibility interval, for each of the 38 DGRP lines, and from the per-fly data of the authors' Dryad deposit (
doi:10.5061/dryad.k6djh9w7q, lifespan_Ral_R.csv).
Design: one nearly complete haploid genome (the X chromosome and one copy of each major autosome, more than 99% of the genome; not the dot 4th chromosome) of each of 38 DGRP lines was captured with the hemiclone technique: hemiclone males carry this DGRP genome together with a Y chromosome, a translocation between the two major autosomes (marked with the dominant eye-colour mutation
bwD) and the 4th chromosomes of a clone-generator stock (ESM Figure S1). For each line, four sets of 25 hemiclone males (4 days old) were mated for 60 min to 25 virgin DGRP_859 females (7 days old), which then laid eggs in fresh vials on four consecutive days (about 300 eggs per vial). Only the brown-eyed sons, which inherited the translocation from their father ("PIGE-clone" sons), were collected: two sets of 50 sons from each set of females, i.e. 8 vials of 50 males per line over 4 blocks. Lifespan was scored by transferring the flies to fresh food every second day and recording the dead flies at each transfer, until all were dead (25 °C, 60% humidity, 12:12 h light:dark, yeast and sugar medium, no anaesthesia). All phenotypes are of males (sex M).
Paternal indirect genetic effect: the sons whose lifespan was measured do not carry the genome of the DGRP line. They all have the same genotype (an X chromosome from their DGRP_859 mother; the Y chromosome and the translocation of the clone-generator stock from their father), so differences in their lifespan among lines measure the effect that the DGRP haploid genome of their father has on them through genes they did not inherit (paternal indirect genetic effect, PIGE). The values are stored under the DGRP line whose genome the father carried, and the cross with DGRP_859 mothers is given in the phenotype names (
_x859). DGRP_859 is also one of the 38 tested genomes. In the authors' Bayesian mixed model (rstanarm; Lifespan ~ Block + (1|Line) + (1|Vial)), the variance among lines was small but significant: V
G(I) = 2.19 days² (95% credibility interval 1.05-3.52), against 7.35 for vials, 0.42 for blocks and 120.30 residual, for a mean lifespan of 62.79 days (ESM Table S3, Figure S5).
Summary dataset:
mn_LS_PIGE_x859Auth holds the Table S4 means as printed (3 decimals), from 60.129 days (DGRP_427) to 65.743 days (DGRP_437). They are posterior means of the authors' model (the mean over the 4 blocks plus the random effect of the line), not raw means: they are shrunk toward the overall mean and differ from the line means that DGRPool computes from the raw dataset (see below). The 95% credibility intervals of Table S4 are not uploaded.
Raw dataset 1 (lifespan_Ral_R.csv, Dryad): one row per fly (14,296 flies, 38 lines, 304 vials: 8 per line, 2 per line and block; 27-52 flies per vial, median 50), with
LS_PIGE_x859 = lifespan (age at death in days, column LS of the file; whole days, 2-102), and the experimental block (
Block_x859, column block, 1-4) and vial within line and block (
Vial_x859, column rep, A or B) kept as obsolete design columns. The file has no missing values. The line means computed by DGRPool from these data (58.576 to 67.199 days) correlate with the Table S4 estimates at r = 0.998 and all fall inside the Table S4 95% credibility intervals; they differ from them by -1.55 to 1.46 days because Table S4 comes from the authors' model.
Not uploaded: the among-population assays (hemiclones of single genomes from the Congo, Dahomey, LHm and Z53 populations, crossed to Dahomey females: upper part of Table S4 and the Dryad files direct_lifespan_R.csv, indirect_lifespan_R.csv and F2_lifespan_R.csv), which do not use DGRP lines. The authors' R code is on Zenodo (
doi:10.5281/zenodo.6459147).