Study

Curated
Bai et al., 2026
Outbred Drosophila populations reveal diet-dependent genetic effects on development time
[10.64898/2026.06.05.730474]

Description

Body weight, glucose, glycogen and triglyceride of larvae and adult females, development time, larval survival and female lifespan were measured in 32 DGRP lines on low- and high-sugar diets. High sugar delayed development in every line, with genotype-dependent effects on most other traits.

Comments from curator

Phenotyping data comes from the Supplementary Material of the preprint (bioRxiv 10.64898/2026.06.05.730474, file media-1.zip of the supplementary material):
- Supplementary Data 1 (sheet All_trait_line_summary, identical to its 11 per-trait sheets): for 11 traits and 32 DGRP lines, the line mean and standard error of the mean (SEM = SD/√n over 5 biological replicates) on each diet, the relative change (HSD − LSD)/LSD with its standard error, and a Welch t-test P value.
- Supplementary Table 1 (Supplementary Information): the data code the lines #1 to #32; this table gives their Bloomington stock and RAL numbers (#1 = RAL-440 ... #32 = RAL-859), which all match DGRPool.
Only these per-line values were published (replicate data are "available from the corresponding author upon request"), so the study has a summary dataset only.

Design: the 32 DGRP lines, which are among the 64 founders of the authors' outbred Drosophila Recombinant Populations (DRP), were tested on a low-sugar diet (LSD) and a high-sugar diet (HSD): Bloomington semi-defined medium with 0.15 M or 1.0 M sucrose, at 25 °C and 12:12 h light:dark, after two generations on standard medium at controlled density. Every trait has 5 biological replicates per line and diet.
- Adult females (sex F): flies reared on standard medium were moved 2 days after eclosion to LSD or HSD for 5 days (50 males and 50 females per bottle). Body weight and whole-body glucose, glycogen and triglyceride were measured on pools of 10 females (one pool = one replicate). Lifespan was measured on virgin females moved to the diets 2 days after eclosion, 5 per vial and 5 vials per line and diet, transferred every 2 days without CO2 and counted until all died; the Methods say that the median lifespan was recorded.
- Larvae (unsexed, sex NA): body weight, glucose, glycogen and triglyceride were measured on pools of 10 early wandering third-instar larvae reared on LSD or HSD from the embryo. For development time and larval survival, 70 first-instar larvae were put per vial on LSD or HSD and white pupae were counted every 12 h. Development time is the time from egg deposition to 50% of the total pupae, on a sigmoid fitted to the cumulative pupation curve (authors' code); larval survival is the proportion of the larvae that pupated (0-1).
Metabolites are normalised by the weight of the pool (µg per mg of tissue; Figure 2b and Supplementary Figure 1A). Body weight is in mg per animal (Figure 2b; values of 0.7-2.1 mg).

Summary dataset (per line, one row with sex F for the adult traits and one row with sex NA for the larval traits; 66 phenotypes): for each trait, the mean (mn_) and SEM (se_) on LSD and on HSD, e.g. mn_DevTime_LSD and se_DevTime_HSD, and the authors' per-line relative change (HSD − LSD)/LSD (mn_RelChg_..., computed from the line means; Figure 2d) with its standard error (se_RelChg_..., computed from the replicate-level relative changes). The traits are BodyWt_Ad, Gluc_Ad, Glycog_Ad, TG_Ad and Lifespan (adult females), and BodyWt_L3, Gluc_L3, Glycog_L3, TG_L3, DevTime and LarvSurv (larvae). Values are copied at full precision. The per-line P values are not uploaded. Development time was the trait that the authors mapped in the DRP.

Missing values (NA; Supplementary Data 1 gives a trait for a line only when both diets were measured): adult triglyceride was not measured in lines #1-#8 (DGRP_440, 761, 897, 790, 045, 703, 776 and 228). Larval triglyceride is missing for #4, #12, #15, #17 and #24 (DGRP_790, 796, 508, 321 and 639), larval glucose and glycogen for #4, #12, #15 and #24, larval weight for #4, #12, #15 and #17, development time for #4, #12 and #17, and larval survival for #11 and #14 (DGRP_879 and 822) and for #4 on HSD (its LSD survival, 0.886, is uploaded). The paper says that lines #4, #12, #15, #17 and #24 failed to complete development on HSD and that line #15 also failed on LSD, but #15 and #24 have a development time and a survival on both diets (and #24 larval weights), while #12, #17 and #24 have a low survival on HSD (0.045-0.116). The values are uploaded as published.

Errors in the source:
- Larval glucose and glycogen of lines #17-#20 (DGRP_321, 531, 217 and 380) are identical, on both diets and to the last digit (means and SEMs), to the adult glucose and glycogen of the same lines. Line #17 has no larval weight, which is measured on the same pools, and is one of the lines that failed to develop, so these are most likely copies of the adult values, not larval measurements. They are uploaded as published (larval glucose and glycogen of 28 lines), because the authors' analyses included them: the variance components of Supplementary Table 4 are reproduced exactly from the published means and SEMs only with them. For these four lines, the larval glucose and glycogen phenotypes (mn_, se_, mn_RelChg_ and se_RelChg_ of Gluc_L3 and Glycog_L3, 48 values) therefore equal the adult ones and should be treated with caution.
- Three possible duplicates are kept as published: the adult weights on LSD of #7 and #23 (DGRP_776 and 229; 1.386 ± 0.0133 mg) and of #8 and #24 (DGRP_228 and 639; 1.454 ± 0.0068 mg), and the larval weights on HSD of #24 and #25 (DGRP_639 and 716; 1.526 ± 0.0333 mg), have the same mean and SEM.
- The two correlation sheets of Supplementary Data 1 have swapped names: 'HSD_LSD_combined_corr_r' holds the Pearson correlations of the relative changes, and 'Relative_change_pairwise_corr_r' the HSD/LSD correlations of the line means.

Checks: the uploaded line values reproduce the trait summaries of Supplementary Data 1 (numbers of lines, means across lines, mean and SD of the relative changes, number of lines with P < 0.05) and the variance components, heritabilities and coefficients of variation of Supplementary Table 4 for all 22 trait and diet combinations (n = 5 replicates; with the copied larval values of #17-#20).

Not uploaded: the genetic mapping uses 16 outbred advanced intercross populations founded by 64 DGRP lines (Supplementary Table 1, #1-#64), with flies from the earliest and latest 5% of eclosion sequenced individually (Supplementary Data 2, 5 and 6), so it has no per-line phenotype. The functional validation (Supplementary Data 3, 4 and 7, Supplementary Figures 6 and 7) uses RNAi, overexpression and mutant stocks, not DGRP lines. The hemolymph glucose of lines #1-#8 (Supplementary Figure 1B) is only shown in a plot. The preprint gives no sequencing accession.
Back to studies