Recognition of distinct sleep states in Drosophila uncovers previously obscured homeostatic and circadian control of sleep
Data files
Mar 12, 2026 version files 2.85 MB
-
FileInformation.xlsx
10.14 KB
-
RawMonitorFiles.zip
2.84 MB
-
README.md
6.39 KB
Abstract
Understanding the mechanisms underlying homeostatic sleep regulation is a central unmet goal of sleep science. Our comprehension of such regulation in mammals has required recognizing distinct sleep stages. Drosophila melanogaster is an important genetic model system for studying sleep. Since the discovery of sleep-like states in the fly 25 years ago, the field has treated sleep as a unitary state consisting of any inactivity lasting 5 minutes or longer, despite convergent work suggesting the existence of multiple sleep states. Here, we establish that three distinct sleep states in flies can be classified based on simple inactivity duration criteria. We show that the daily initiation of these sleep states is temporally distinct, with long sleep occurring immediately following the largest daily period of wakefulness. We also report that rebound in response to mechanical sleep deprivation is present only in long sleep and comes at the expense of shorter sleep states. Deprivation-induced decreases in shorter sleep states obscure homeostatic sleep rebound, but only when sleep is measured using traditional methods. We observe distinctly timed ultradian oscillations of fly sleep states, reminiscent of mammalian sleep cycles. Our results indicate that the recognition of such sleep states will be necessary to fully realize the promise of the Drosophila model system for identifying conserved genetic mechanisms underlying such regulation.
Dataset DOI: 10.7959/dryad.stqjq2cfr
Description of the data and file structure
Files and variables
RawMonitorFiles.zip
Description
The zipped folder contains all Raw Trikinetics scanned Drosophila Activity Monitor files. These are text files and can be accessed through any text editor or spreadsheet program. Note that the raw monitor files were passed through the DAMScan software and only scanned monitor files are provided in the repository. Each monitor file has various metadata information from columns 1 through 10. Individual fly activity counts are recorded in columns 11 through 42, totalling 32 flies per DAM system. Additional details on the metadata and format of scanned monitor files may be found at the manufacturer's website: https://trikinetics.com.
The individual file names are labelled according to the following convention: Genotype_EnvironmentalCondition_ReplicateExperimentNumber_ExperimentalCondition.txt. The three genotypes in the associated manuscript are Canton-S (CS), loss-of-function period mutant (per01), and Shaker[minisleep] mutants. Canton-S flies are wild type flies. Environmental conditions are labeled either LD or DD. LD stands for Light/Dark cycles and DD stands for constant darkness. Experimental conditions are in the filename only for data collected in sleep deprivation experiments. The experimental conditions are either labeled Control (for undeprived flies) or Deprived (for sleep deprived flies). Filenames that do not have environmental conditions in their names were all done under LD cycles. Filenames that do not have run numbers do not have replicates for which data has been uploaded. Filenames that do not have experimental conditions reported do not have different treatment/experimental conditions under which the experiments were conducted.
FileInformation.xlsx
Variables
The spreadsheet file has 5 columns.
- The first column (Figure Details) lists the figure (in the associated manuscript) and environmental condition or genotype for which the data were used.
- The second column (Filename) lists the name of the text files in the zipped folder which were used for making said figure in the associated manuscript.
- The third column (Analysis Start Date)
- The fourth column (Analysis Start Time) report the analysis start date and time used for corresponding analyses.
- The fifth column (Data Repository) contains doi information of data that have been previously published but were re-analysed for this study.
Description of figures referenced in Figure Details column.
Figure 1: Durations of inactivity can be used to identify sleep states.
Sleep bout durations and % change in metabolic rate were extracted from Stahl et al., 2017. Physiological changes previously reported in relation to sleep bout duration that informed bout duration cutoffs are also noted. Bout length-based classification of sleep reveals waveforms of provisional sleep states under Light/Dark cycles for wildtype flies. These are remarkably similar in shape to the Drosophila sleep states identified previously using Hidden Markov Models. Also analyzed are the waveforms of the sleep states, during the first day of constant darkness (DD) following entrainment to LD cycles. The data under LD and DD are means from three independent runs. Filenames for data associated with this figure are CS_LD_DD_Run1.txt, CS_LD_DD_Run2.txt, CS_LD_Run3.txt, CS_DD_Run3.txt.
Figure 2: Probability of initiating long sleep is highest immediately following major peaks of wakefulness.
Profiles of activity, standard sleep and the three provisionally identified sleep states for wildtype flies under Light/Dark cycles (LD), constant darkness (DD), and UV- and Blue-blocked ramped light cycles (ramps). Also analyzed are initiation probabilities of standard, short, intermediate, and long sleep under LD, DD, and ramps. Filenames for data associated with this figure are CS_LD_DD_Run2.txt.
Figure 3: Rose plots clearly demonstrate the temporal sequencing of sleep states in wildtype flies.
Rose plots (averaged over flies) of locomotor activity and sleep states under constant darkness (DD) and UV- and Blue-blocked ramped light cycles (ramps). Polar plots representing the gating of the four sleep/wake states under DD and ramps are generated. Filenames for data associated with this figure are CS_LD_DD_Run2.txt.
Figure 4: Long sleep displays substantial homeostatic sleep rebound following deprivation in wildtype flies.
Sleep rebound for standard sleep and each of the three provisional states of sleep were analyzed. Statistical comparisons were made against 0 for each of the metrices and sleep stages separately using single sample t-tests. Estimated sleep rebound for standard sleep and each of the three provisional sleep states for a “homeostatic” mutant and its background control, CS flies. Note that the data for CS flies is reused. Data for the sh[mns] are pooled from five independent runs (n = 156). Filenames for data associated with this figure are shaker_Run1_Control.txt, shaker_Run1_Deprived.txt, shaker_Run2_Control.txt, shaker_Run2_Deprived.txt, shaker_Run3_Control.txt, shaker_Run3_Deprived.txt, shaker_Run4_Control.txt, shaker_Run4_Deprived.txt, shaker_Run5_Control.txt, and shaker_Run5_Deprived.txt.
Figure 5: Sleep states display distinct circadian and ultradian regulation.
Normalized, averaged scalograms of standard sleep, and the three states of sleep under constant darkness. Filenames for data associated with this figure are CS_LD_DD_Run2.txt.
Figure 6: Ultradian rhythm amplitudes for each sleep state are distinctly gated by the circadian clock.
Normalized scalograms for short, intermediate, and long sleep in wildtype and loss-of-function per01 mutant flies depicting amplitude values in the ultradian period range used for downstream analyses. Results of chi-squared periodogram analyses carried out on wildtype and per01 flies were extracted for short, intermediate, and long sleep. Filenames for data associated with this figure are CS_LD_DD_Run2.txt, and per01_DD.txt.
All data were collected using Drosophila Activity Monitors (as described in the methods section of our manuscript). Processed data are uploaded as a compressed file. Some figures were generated by re-using data from other publications - that have been cited appropriately. Readers are redirected to cited studies for those raw data and associated methods.