Companion Data Page Precise quantification of behavioral individuality from 80 million decisions across 183,000 flies Benjamin de Bivort1,*, Sean Buchanan1, Kyobi Skutt-Kakaria1,2, Erika Gajda1, Chelsea O'Leary1, Pablo Reimers1,3, Jamilla Akhund-Zade1,4, Rebecca Senft1,5, Ryan Maloney1, Sandra Ho1, Zach Werkhoven1, Matthew A-Y Smith1,6 1 - Center for Brain Science & Department of Organismic and Evolutionary Biology, Harvard University, Cambridge, MA, USA 2 - Current affiliation: Department of Biology and Biological Engineering, California Institute of Technology, Pasadena, CA, USA 3 - Department of Neurobiology, Harvard Medical School, Boston, MA, USA 4 - Karius, Redwood City, CA, USA 5 - Imaging Platform, Broad Institute of MIT and Harvard, Cambridge MA,USA 6 - Department of Entomology, University of Wisconsin-Madison, Madison WI,USA * - correspondence: debivort * oeb.harvard.edu Preprint PDF
Data
handMetaData.2021.12.01.mat	Matlab data file containing the table variable allHandMetaData, which contains all behavioral data, Levene-transformed behavioral data for analysis of variability effects, and metadata. Each row is a fly. Also included are tables corresponding to the contributions of individual experimenters, the structure variable " "momentEstimates" which contains 1000x20x5 matrix objects from the statistical moment resampling analysis in Figure 2. The dimensions correspond respectively to 1000 bootstrap replicates, 20 standardized moments, and 5 levels of data subsampling.
Analysis scripts
handMetaDistBootstrap.m	Computes a kernel density estimate of a behavioral distribution with bootstrap resampling. Used to make the first column of panels in Figure 2. Input is a data column from allHandMetaData.
handMetaEyeColorAnalysis.m	Performs the analysis of the effect of the white locus on variability in Figure 5C. Input is allHandMetaData. Output has dimensions 6x3xnumReps corresponding to 6 states of the white locus, 3 behavioral variables and the number of bootstrap replicates.
handMetaGenotypeVariabilityAnalysis.m	Computes the distributions of genotype variabilities in Figure 5A. Output objects are the coefficients of variation for DGRP and non-DGRP genotypes. MATLAB's ksdensity function can then be used to compute the violin plots for these values. Input is allHandMetaData.
handMetaGxEVarAnalysis.m	Computes the genotype variabilities under serotonin perturbation in Figure 5B. Two output objects have dimension equal to the number of DGRP genotypes in this experiment by 6 variability values corresponding to the 6 x-axis conditions in each sub-panel. The remaining object is the indices into allHandMetaData of the flies associated with this analysis. Inputs are allHandMetaData and a string indicating the data column of interest, e.g., 'handedness'.
handMetaLeftyAnalysis.m	Conducts the analysis of mean turn bias in the grand data set and broken out by genotype, sex and experimenter (Figure 3A-D). Output object contain the mean turn bias values in those plots. Input is allHandMetaData.
handMetaMoments.m	Performs the boostrapping analysis for standardized moment estimation in the right column of Figure 2. Output matrix has dimensions numReps x numMoments x 5 which by default are 1000 bootstrap replicates, 20 standardized moments, and 5 levels of data subsampling. Input is allHandMetaData.
handMetaReport.m	Prints to the command window a text report on the behavioral data and meta data in allHandMetaData. Input is allHandMetaData.
handMetaSexAndTempVariability.m	Computes the variability of lines in the temperature analysis of Figure 5D. Also plots the subpanels of that panel. Output is a matrix with dimensions equal to the number of genotypes x 2 (the two temperature conditions). Inputs are allHandMetaData and a string indicating the data column of interest, e.g., 'handedness'.
handMetaStackedBar.m	Plots a stacked bar in the style of the five bars of Figure 1A. Outputs a cell array of the ordered labels of the bar segments. Input is a data column of allHandMetaData.