Comments from curator
Phenotyping data comes from the authors'
Zenodo record 10.5281/zenodo.6651851 ("Data and code", cited in the Data accessibility section of the paper; the record uses the bioRxiv title of the paper,
10.1101/2022.06.09.495537):
-
survival.csv (2,400 flies) and
fecundity.csv (1,193 females of the same experiment), used for raw dataset 1.
-
titre.csv (600 flies), used for raw dataset 2.
-
expression_MK.csv (qPCR of
G9a and
upd3), used for raw dataset 3.
The authors' R script of the same record (DCV_stats_ms.R) was read to interpret the columns.
Design: 10
Wolbachia-free DGRP lines, 5 carrying the resistant (R) allele of the antiviral restriction factor
pastrel (
pst; DGRP_059, 075, 379, 502, 738) and 5 the susceptible (S) allele (DGRP_138, 373, 380, 765, 818). Males and females aged 3-5 days were infected systemically by pricking the thorax with a needle dipped in
Drosophila C virus (DCV) at 10
3, 10
5, 10
6, 10
8 or 10
9 DCV RNA copies/ml, or in sterile Tris-HCl (sham). The dose is the suffix of the phenotype names (_Sham, _DCV1e3, _DCV1e5, _DCV1e6, _DCV1e8, _DCV1e9). The
pst allele (R/S) is a genotype of the DGRP lines and is not uploaded.
Raw dataset 1 (one row per fly; 20 flies per line, sex and dose, 5 in each of 4 experimental blocks): each fly was kept alone in a vial and checked daily for up to 78 days.
Lifespan is the number of days from pricking to death.
CumFec (females only) is the number of adult offspring of the same female over the 30 days after pricking, or until her death, and
ReproRate is the authors' reproductive rate, CumFec / min(lifespan, 30). The rows of fecundity.csv have their own numbering, but they follow the order of the females of survival.csv, with the same days survived fly by fly: 7 females (5 sham, 1 at 10
3, 1 at 10
9) have no fecundity record and are NA. The doses are stacked: row k of a line and sex holds the k-th fly (authors' Fly.ID order) of each dose, so a row holds six different flies of the same block (obsolete column
Block).
Censoring: 78 flies are censored in survival.csv (40 at 76-78 days, presumably alive at the end of the experiment; 38 at 7-62 days, reason not given). The authors treated them as censored in their Cox models (Models 1a, 1b) and used their day of censoring as is in the lifespan models of mortality tolerance (Models 3a, 3b) and for fecundity. So Lifespan keeps that day, and the censoring flag is kept in the obsolete
Censored columns (1 = censored), so that survival analyses can be redone. The line means of Lifespan computed by DGRPool include these days.
Raw dataset 2 (one row per fly): viral load of 5 other flies per line, sex and dose, 3 days after pricking, as log10(DCV RNA copies per fly) by qRT-PCR against a DCV standard curve (authors' column Log10.DCV). Sham flies give a background signal of about 10 copies, which the authors call false positives: they analysed log10(copies + 1) including them, and also with these values set to 0. The values are uploaded as is. The doses are stacked as in raw dataset 1 (different flies on a row).
Raw dataset 3 (one row per replicate): flies aged 3-7 days were sham-pricked or pricked with 10
7 DCV RNA copies/ml. 3 days later, 5 pools of 3 flies per line, sex and treatment were assayed by RT-qPCR.
dCt_<gene>_Sham and
dCt_<gene>_DCV1e7 are Ct(gene) - Ct(
rp49) after the authors' correction for between-plate variation (columns dCTC and dCTE; higher values mean lower expression).
FoldCh_<gene>_DCV1e7 is the authors' 2
-ΔΔCt, pairing the infected and sham pools with the same replicate number (obsolete column
qPCR_Rep). Both genes of a row come from the same cDNA. The uncorrected and corrected Ct values, ΔΔCt and log10 fold change of the file are not uploaded: they are the same data on other scales.
Not uploaded: the per-line tolerance estimates exist only in figures. Mortality tolerance (Fig. 3c) is the integral of each line's quadratic reaction norm of lifespan on log10 dose, with the intercept set to 0. Fecundity tolerance (Fig. 4c) is the slope of each line's cumulative fecundity on log10 dose. Both can be derived from raw dataset 1 (the sham dose is log10 dose 0).
Checks: two-way ANOVAs (Sex × Line) on raw dataset 3 reproduce the line and line × sex F values of Table 4 (Models 5b-8b) exactly, and the sex F values of Models 6b and 8b. The sex F values of Models 5b and 7b are not reproduced (136.3 vs 381.7 and 29.1 vs 49.1; P < 0.0001 either way). R lines lived longer and had more offspring than S lines at every dose, and lower viral loads at every DCV dose, as in the paper.
The same laboratory measured DCV lifespan, viral load and shedding on the same 10 DGRP lines in
DGRPool study 116 (Siva-Jothy & Vale 2021).