Comments from curator
Phenotyping data comes from the
Dryad dataset of the paper (
doi:10.5061/dryad.d4d19; file taken from its
Zenodo mirror, identical md5):
-
Miyagi et al_Appendix 1.xlsx, titled "Table S1 Raw pigmentation scores of the 20 DGRP strains used in this study": one row per female fly, with the raw image measurements and the photograph IDs of the body images, which are also on Dryad (one zip per line, 5.3 GB, not used here).
The paper (
journal page) is not open access. The methods below come from the Dryad README, which quotes the paper's Materials and Methods, and from the companion paper
Sunaga et al. 2016 (Genes Genet Syst 91:127; same lab and imaging set-up). Its
Supplementary Table S1 gives three per-line values of the 20 lines "from Miyagi et al. (2015)", which were used to check the data here.
Phenotyping: only adult females of 20 DGRP lines were measured. The abstract's "seven body positions" are most likely the four brightness positions plus the dark-pigmented area of three abdominal segments:
-
% darkness (females 3-6 days post-eclosion): flies were fixed overnight in 10% glycerol in ethanol at 4°C, and head, legs and wings were removed. The headless bodies were kept in 10% glycerol/PBS and photographed dorsally with a reference gray card (brightness 128) under identical settings. The mode grayscale brightness (0-255, ImageJ) was measured at four positions marked in Fig. 1B of the paper: thoracic center (TC), thoracic flank (TF) and thoracic scutellum (TS) on the thorax image, and abdominal light area (AL) on the abdomen image of the same fly. Appendix 1 also gives the brightness of the gray reference (B) of each image.
-
% dark-pigmented area (other females, 6-7 days post-eclosion): abdomens were cleared in 10% KOH (2 h, 65°C), and the cuticle was flattened and mounted in Hoyer's medium. The dark stripe of tergites A4, A5 and A6 was selected with a threshold at the midpoint between the stripe and non-stripe brightness peaks. Appendix 1 gives the area (pixels) of each segment and of its stripe.
Values computed per fly by the curator with the authors' formulas (Sunaga et al. 2016 methods): % brightness = (brightness × 128 / B) / 255 × 100, using the gray reference of the same image, and
% darkness = 100 − % brightness (higher = darker).
% dark-pigmented area = 100 × stripe area / segment area. Values are written at full precision. The raw brightness values and pixel areas are not uploaded (they are in Appendix 1).
Datasets (sex F only):
-
raw_1: one row per fly imaged whole.
DarkTC,
DarkTF,
DarkTS and
DarkAL (% darkness of the four positions), and
DarkAvg (mean of the four, per fly), for 206 flies (4 to 13 per line, 10 for most lines). The same four traits with suffixes
_3d and
_6d come from an age series of 5 flies exactly 3 or 6 days post-eclosion, for DGRP_357, 358, 820 and 852 only. The authors used these flies for the age-dependent change of pigmentation, not for the line values. The age series is stacked next to the main flies: row k holds the k-th fly of each group, and different groups on one row are different flies.
FlyNo,
FlyNo_3d and
FlyNo_6d (obsolete) are the authors' fly numbers. They match the photograph IDs RAL-<line>_<number>_1 (thorax) and _2 (abdomen), or RAL-<line>_3days_<number>_1 for the age series.
-
raw_2: one row per dissected abdomen.
DarkAreaA4,
DarkAreaA5 and
DarkAreaA6 (% dark-pigmented area of each segment), 5 flies per line, 100 flies in all.
FlyNo_DarkArea (obsolete) is the authors' fly number (photograph RAL-<line>_<number>_1).
-
summary:
mn_DarkAreaA6Auth, the authors' per-line % dark-pigmented area of A6 as published in Sunaga et al. 2016 Table S1 (see below).
Checks against the authors' per-line values (Sunaga et al. 2016, Table S1, 1 decimal):
- The line means of DarkAvg (main flies only) equal the published "average % darkness" for 20/20 lines (max difference 0.05). Adding the age-series flies breaks 4 lines, so the authors did not include them.
- The line means of DarkAreaA4 equal the published A4 values for 20/20 lines (max difference 0.05).
- The line means of DarkAreaA6 equal the published A6 values for only 11/20 lines. DGRP_208, 324, 335, 358, 360, 365, 380, 517 and 555 differ by 0.2 to 2.8 points (DGRP_555: 80.8 here vs 78.0), although the correlation is 0.998. For 7 of these 9 lines, the published value is the mean of the first 4 of the 5 flies, as if the fifth A6 measurement had not been used. DGRP_380 and 555 fit neither, and no other simple alternative (median, ratio of summed areas, A5) explains them. The published values are therefore also uploaded as
mn_DarkAreaA6Auth.
Sunaga et al. 2016 gives no per-line values for the four separate positions or for A5, and the paper itself could not be read.
Not uploaded: the allele-specific expression (ASE) of four melanin-pathway genes (the DDBJ project lists
ebony,
tan,
black and
pale) in F1 hybrids between the Mel6 strain and each DGRP line. These are molecular data, and their per-line values are not in Appendix 1 and could not be checked. The 454 amplicon reads are at DDBJ (BioProject PRJDB4270). The paper found that the ASE of
ebony was associated with the pigmentation intensity of the thorax.
The companion paper
Sunaga et al. 2016 (not in DGRPool) correlated these per-line values with starvation resistance and chill coma recovery (
Ayroles et al. 2009,
Mackay et al. 2012). It also measured thoracic brightness, the A6/A7 dark area and gene expression on 4 of these lines reared at 20, 25 and 29°C.