Comments from curator
Phenotyping data comes from the
Supplementary data of the paper (
journal page, also on
PMC; file ATE_Wolb&TEs_STables_revised.xlsx in evad036_supplementary_data.zip):
-
Supplementary Table S2 ("TE expression data", the table named in the data availability statement): qPCR Cq of each transposable element (TE) and of the reference gene
EF1, for each biological replicate, TE, line and
Wolbachia status (5,496 rows).
-
Supplementary Table S5: "raw Effect Size" of
Wolbachia removal per line and TE (350 values).
Table S4 (primers and efficiencies) was used to check the data. Table S3 (ANOVA and Kruskal-Wallis tests per line and TE) and Table S1 (PCR validation of predicted TE insertions) are not phenotypes and are not uploaded.
Design: 25 DGRP lines naturally infected with
Wolbachia (wMel; all annotated as infected in the DGRP, none with
Spiroplasma) were used as such (Wolb+) and as a
Wolbachia-free version (Wolb−), obtained by rearing on tetracycline (0.05 mg/mL) for two generations, restoring the gut microbiota with an inoculum from the untreated line, and keeping the flies for at least five generations before the experiments (cure confirmed by
wsp PCR). Flies were reared at 25 °C, 12:12 h light:dark, on cornmeal-molasses food; newly eclosed adults were kept in vials of 10 females and 6 males, and the females were collected at 10 days of age. For each line and status, RNA was extracted from 8 replicate pools of 10 whole females, and the cDNA was quantified by qPCR (SYBR Green) with primers for 14 TEs (412, 1360, blastopia, blood, copia, Cr1a, gypsy5, Idefix, Juan, mdg1, opus, P-element, pogo, Quasimodo) and for
EF1. Each Cq is the mean of two technical replicates; biological replicates whose technical replicates differed by an SD > 0.5 were discarded by the authors (104 of the 5,600 TE measurements; NA in the upload). The lines are given as RAL-xx (= DGRP_0xx). All values are from females (sex F).
raw_1 has one value per biological replicate (pool of 10 females) in a compact layout: row k of a line holds replicate k (Table S2 "Replicate", kept as the obsolete design column
Replicate) of every TE and status. For each TE (P-element written P_element):
-
Cq_<TE>_Wolb and
Cq_<TE>_NoWolb: Cq of the TE in Wolb+ and Wolb− females (PCR cycles; a lower Cq means more transcript);
-
EF1_<TE>_Wolb and
EF1_<TE>_NoWolb: Cq of
EF1 in the same sample, as given on the same Table S2 row (one
EF1 reaction often served several TEs, so several EF1_ columns hold the same values).
A TE Cq and its
EF1 Cq (same TE and status) on one row come from the same sample. Other values on a row do not necessarily: Wolb+ and Wolb− are different pools, and Table S2 numbers the replicates 1 to n after discarding some, so replicate k of one TE is often not the same RNA pool as replicate k of another TE (the
EF1 values show it).
Normalised expression is not uploaded (it is derived): the authors used the Pfaffl method, log2 ratio = (Cq
cal,TE − Cq
TE) × log2(E
TE) − (Cq
cal,EF1 − Cq
EF1) × log2(E
EF1), with E = 1 + efficiency/100 (Table S4;
EF1 98.10%) and the mean of the line's Wolb+ samples as calibrator (cal). It can be computed row by row from Cq_<TE>_<status> and EF1_<TE>_<status>.
The
summary dataset has
mn_WolbRm_<TE>: the effect of
Wolbachia removal on the expression of each TE, Table S5 "raw Effect Size", as published. These values are the mean log2 expression of the Wolb− samples minus that of the Wolb+ samples (recomputed from Table S2 to 1e-6 for 346 of the 350 line × TE pairs), so a positive value means higher expression without
Wolbachia.
Of note:- The paper's text defines the effect size as the
median log2 expression of Wolb+ minus that of Wolb− flies, i.e. with the opposite sign to Table S5: blood in DGRP_595 is "around 2.3" in the text and −2.26 in Table S5, and the medians over TEs of DGRP_021 and DGRP_712 are "around −1.1 and −0.9" in the text and 1.15 and 0.93 in Table S5. The counts in the text (24 line × TE pairs with expression more than doubled after removal, 3 with a strong decrease) match Table S5 (24 values above 1, 3 below −1).
- For P-element, Tables S3 and S5 were computed with a primer efficiency of 112.72%, not the 94.09% given in Table S4 (all 25 P-element values of Table S5 and the Table S3 F-values are reproduced with 112.72%).
- Four Table S5 values differ slightly from those recomputed from Table S2 (DGRP_321: blastopia 1.863 vs 1.928, Quasimodo 1.399 vs 1.363, Cr1a 1.4291 vs 1.4294; DGRP_237 Idefix: 0.077 vs 0.072), and so do their Table S3 F-values: a few Cq values analysed by the authors probably differ from Table S2. Table S5 is uploaded as published.
- Table S3 has 74 of the 350 line × TE pairs with P < 0.05 (ANOVA), as in the paper; 47 have a higher expression in Wolb− and 27 (26 in the text) in Wolb+ flies.
Not uploaded: the
Wolbachia load (
wsp/
actin qPCR ratio) of 5 individual females per Wolb+ line, which is only shown in Fig. 1A (no table); the numbers of novel TE insertions of the lines (Fig. 1B), taken from Rahman et al. 2015 (TIDAL-FLY) and not measured in this study; and the TE insertion validations (Table S1).