Comments from curator
Phenotyping data comes from
Dataset S6 ("Embryonic heat shock survival data in the DGRP", pnas.2518358123.sd06.txt) in the
Supporting Information of the paper (also on
PMC, and as SUPP_TABLES/TableS6_OK_EB_data.Aug16.2026.txt in the authors'
GitHub repository). For 64 DGRP lines (written line_NNN), it gives the number of embryos that hatched after a heat shock, the number of embryos scored and their ratio. Only these per-line counts were published (no per-plate data), so the study has a summary dataset only.
Design: the DGRP lines (Bloomington Drosophila Stock Center) were kept at 25 °C in 12:12 h light:dark on cornmeal-yeast-molasses food. About 50 pairs of 3- to 5-day-old flies of a line laid eggs for 1 h at 25 °C on grape juice agar plates. The plates with the embryos (0-1 h old) were wrapped in Parafilm and heat-shocked for 45 min in a water bath at 35 °C (a heating ramp of about +0.6 °C/min), following Lockwood et al. 2018. The embryos were then arranged in 5x4 grids on the agar and returned to 25 °C, and the hatched embryos were counted after 48 h. The sex of the embryos is unknown (sex NA).
The summary dataset has:
-
mn_EmbSurv_HS35C: embryonic heat-shock survival, the proportion of the embryos of the line that hatched (0 to 1), as printed in Dataset S6 ("Proportion" = Hatched / Sample_size, 6-7 digits). It ranges from 0 (DGRP_714, 0 of 49 embryos) to 0.66 (DGRP_732), as in the paper ("0% to 66%").
-
NHatched_HS35C and
NEmbryos_HS35C (obsolete design columns): the number of embryos that hatched and the number of embryos scored (16 to 652 per line; 10,968 embryos in total).
Two lines are missing from Dataset S6: the paper and Dataset S6 say 64 lines, but Fig. 5A (original and corrected) plots 66: the 64 lines of Dataset S6, DGRP_088 and DGRP_235. The counts of these two lines are in the authors' GitHub file
CORRECTION/DGRP.ProperlyRelabeled.data.txt (DGRP_088: 2 of 16 embryos hatched; DGRP_235: 7 of 37), and the points and 95% binomial confidence intervals of Fig. 5A match them exactly. The paper's ANOVA (interaction of the 2R and X SNPs, F
1,55 = 11.22, P = 0.0014) is also reproduced with these 66 lines (F
1,55 = 11.23, P = 0.0015, with the authors' corrected
SP70 and
sog genotypes and the DGRP Wolbachia, In(2L)t and In(2R)NS annotations as factors), but not with the 64 lines of Dataset S6 (F
1,53 = 10.36). The genotype group sizes given in the text (23 + 13 + 8 + 22) also add up to 66 lines. These two lines are therefore uploaded from the GitHub file.
Of note, the GitHub file is another version of the data (67 lines). It has the same counts as Dataset S6 for 60 lines, but other counts for DGRP_391 (7/73 instead of 7/72) and DGRP_714 (3/108 instead of 0/49), it lacks DGRP_383 and DGRP_385, and it has 3 lines that are neither in Dataset S6 nor in the figure (DGRP_309: 44/100; DGRP_371: 3/41; DGRP_703: 10/80). Fig. 5A shows the Dataset S6 values of DGRP_391 and DGRP_714, so Dataset S6 is used for all its lines, and these 3 extra lines are not uploaded.
Correction: a
correction (June 2026) states that the allele labels of the DGRP lines were swapped in Fig. 5A-G and SI Appendix Fig. S6B-G. After correction, the lines homozygous for the temperate alleles at both loci (2R
AX
A, near
SP70 on 2R and
sog on X) have the highest embryonic survival, and the two loci interact epistatically. The survival data are not affected.
Paper's analyses: the DGRP lines were used to test the two top SNPs of an introgression mapping between a tropical (St Kitts) and a temperate (Vermont) line: survival was compared among the four homozygous genotype combinations (ANOVA with Wolbachia and inversion effects, not significant). No GWAS was run. The adult CTmax compared in Fig. 5E-G comes from Lecheta et al. 2020 (
DGRPool study 195).
Not uploaded: the introgression lines (backcrosses of the St Kitts and Vermont lines, not DGRP lines), their pooled sequencing and embryo RNA-seq, and the allele ages (Zenodo).