https://doi.org/10.5061/dryad.547d7wmg3

Data was generated from a study of Drosophila undergoing seasonal adaptation in NJ during summer-fall 2019.

Description of the data and file structure

Data for temperature data loggers are in files called "dl_N.csv", with N = 1-8 to note the different cages. Each column for the data loggers is as follows: S/N = arbitrary sorting, Time = time the measurement was collected, Temperature = temperature in Celsius, Humidity%RH = %relative humidity. 

The following files are for microbiome data (generated using QIIME2 and exported to analyses/visualization in R):
field19_total_meta.tsv is the metadata that relates sequencing library to collected metadata. The columns are as follows: flyid = individual fly, BarcodeSequence = is for demultiplexing, for.sort = an arbitrary sorting variable, plate.pos = position within 96 well plate during DNA extraction and library prep, pcr.id = internal note for when PCR took place, cage = cage from where the fly collected, wolbachia = wolbachia status, timepoint = discrete variable for collection time, datecollected = date collected (each timepoint matches a date), protK = protK treatment during DNA extraction, day_since_start = time since the field experiment started, merge.group = variable that merges together the +/- protK from individual flies, seq.run = sequencing run, arb.sort = another arbitrary sorting variable. 

otu_matrix.csv is the ASV matrix (rows are ASVs, columns are individual flies)

taxonomy_fullnames.tsv links ASV ID to taxonomic assignment. Please note that there are empty cells in this taxonomy file that are used to filter out later. The Rcode will later fill in empties. Rows are ASVs, columns are the different taxonomic ranks. 

tree.nwk is phylogenetic tree

field2019_cleanF.Rdsis the phyloseq object with the cleaned up data (contaminants removed, protK merged, remaining wolbachia removed)

The data for Wolbachia frequency over time:\
wolbachia_screen.csv

For wolbachia_screen.csv, columns are as follows: cage : cage the fly was collected from the field, date = collection date, sample = sample ID from each cage per time point, wolbachia  = y (yes, presence of band during PCR check), n (no, no band). 

The data for the three starvation timepoints:
starv1_clean.csv for Day 96
starv2_clean.csv for Day 116
starv3_clean.csv for Day 127

For all of the starv files columns are as follows: fly.id = individual fly measured, camera = camera used to film, imageplate = plate used to image flies, imagewell = position in the 24 well plate, hour = hour of death, minute = minute of death, timetodeath = adds hour and minute together, cage = cage the fly was collected from in the field, sex = sex of fly, wolbachia = wolbachia status, full.treat = combines cage + sex + wolbachia status, code = 2 for survminer (all dead)

The data for the three starvation timepoints:
isofemales_lifespan.csv contains lifespan measurements at the end of the season (Day 127). 

The columns are as follows: tube = tube ID that matched an individual fly, death day = date that fly died, start. date = date of collection, days.alive = length fly was alive, cage = cage from where the fly was collected, wolbachia = wolbachia status, code = 2 for survival package (all are 2)

The data from statistical model outputs that were later visualized
ph_starv_coxestimates.csv contains the estimates from the Cox models to visualize effect size over time. The columns are as follows: timepoint = timepoint from where the flies were collected, coefficient = model coefficient, se = standard error, group = model term from either Wolbachia or Commensalibacter

tops_bc_modstats.csv contains the variance explained by fixed effects in the analysis of community turnover in the top10 most abundant bacteria. The columns are as follows: model = how many bacteria were included in the statistical model (i.e., top04 = top 4 most abundant bacteria, top05 = top 5 most abundant bacteria, etc), r2 marg = marginal r-squared value from fitted model, r2 con = conditional r-squared value from fitted model, p val season = the p-value for the estimate for time since start of the experiment, p val wolb = the p-value for the estimate for the wolbachia term, p val interact = the p-value for the time * wolbachia interaction term, interact.sig = whether or the not the interaction term was significant (p<0.05). 

FASTQ files:

fastq files correspond to the "field19_total_meta.tsv". files with "-read-1.fastq" are forward, while "-read-3.fastq" are reverse. 

Sharing/Access information

Sequence is uploaded here. 

Code/Software

field2019_qiimecode.txt contains code run to generate ASV and taxonomy assignments using QIIME2 v2020.6
field2019_v4_share.Rmd is the R markdown containing all analyses. The following packages are required: phyloseq, ggplot2, tidyverse, RColorBrewer, ggpubr, scales, decontam, vegan, lubridate, btools, lmerTest, lme4, forcats, survival, survminer, coxme, and mgcv