Comments from curator
Phenotyping data comes from
Table S1 ("Average D7 larvae length for individual DGRP lines") of
Document S1 (Transparent Methods, Figures S1–S4 and Tables S1–S3, file mmc1.pdf) in the
Supplemental Information of the
paper. For 53 DGRP lines, identified by their Bloomington stock numbers (converted here to DGRP line names), it gives the average larval length (mm) at day 7 after egg laying and its standard deviation (SD) for germ-free (GF) larvae and for larvae mono-associated with
Lactobacillus plantarum WJL (
LpWJL), and the
LpWJL/GF ratio of the two means. The lengths of individual larvae were not published, so only this summary dataset is available. The values are copied as printed (3 decimals). Sex was not determined in larvae (sex NA).
Assay: axenic embryos (dechorionated with bleach) were raised at 25°C on a low-yeast diet (chronic under-nutrition: the standard diet of 80 g/L cornmeal, 7.14 g/L agar, 5.12 g/L Moldex and 0.4% propionic acid with only 6 or 8 g/L inactivated yeast instead of 50 g/L; the Transparent Methods do not say which yeast content was used for the DGRP screen, but the main text reports the extra yeast for germ-free larvae only for the RNA-seq and F
2 experiments). For mono-association, 40 freshly laid eggs were inoculated with about 5x10
7 CFU of
LpWJL; GF controls received sterile PBS. On day 7 after egg laying, all live larvae were collected, heat-killed, mounted and imaged, and their length was measured with ImageJ. Per line and condition, there were 3 biological replicates (10 to 40 viable larvae each) per experiment and 2 to 3 repeated experiments; the Table S1 mean and SD are computed on the pooled larvae of all experiments (Figure 1 legend).
GWAS phenotype: the paper's GWAS (Table S2, run on the DGRP2 web server) used the larval length fold change. The per-line values given to the GWAS are in the authors'
Mendeley Data dataset (FC.zip, file FC/raw.adjusted.pheno.txt, column single_raw), whose top associations (gwas.top.annot) reproduce the P-values of Table S2 (19 of 20; the other one, for dpr6, differs by a digit transposition: 2.492e-05 vs 2.94E-05). These values are on a log2 scale and are uploaded as published as
LarvLenLog2FC_GWAS. Although the paper says the GWAS was based on the ratio column of Table S1, they equal log2 of the Table S1 ratio (within 0.002) for only 18 of the 53 lines; for the other 35 lines they differ by 0.004 to 0.18 log2 units (Spearman correlation with the Table S1 ratio 0.97), so they were most likely computed from a different version of the data. The Table S1 ratio is uploaded as
LarvLenRatio_LpWJL. The column single_adjusted of the same file (values adjusted for
Wolbachia and inversions by the DGRP2 server) was not uploaded.
Mismatches with the paper: for DGRP_535 (Bloomington 28208), the Table S1 ratio (1.767) differs from the ratio of the two means (4.14 / 2.339 = 1.770); it was kept as published. All other ratios agree with the means within rounding. The regression of the
LpWJL line means on the GF line means gives a slope of 0.145, as in Figure 1F.
Not uploaded: the RNAi knock-down experiments (not DGRP lines), the experiments on the F
2 progeny of crosses between four DGRP lines (DGRP_335, DGRP_859, DGRP_820 and DGRP_158: larval length, pupariation and emergence times, adult body, wing and eye size, wing patterning anomalies, N-acetylcysteine treatment), which are not DGRP lines and are shown only in figures, and the heritability estimates. The single-larva BRB-seq transcriptomes (36 GF and 36
LpWJL larvae of 3 DGRP lines at day 4) are in ArrayExpress (
E-MTAB-6518). A preprint of this work was posted on bioRxiv (
10.1101/334342).