Genome-wide gene expression noise in Escherichia coli is condition-dependent and determined by propagation of noise through the regulatory network

<p>In this repository we provide raw and processed datasets for the article: &ldquo;Genome-wide gene expression noise in <em>Escherichia coli </em>is condition-dependent and determined by propagation of noise through the regulatory network<strong>&rdquo;&nbsp;</strong>by Arantxa Urchuegu&iacute;a, Luca Galbusera, Dany Chauvin, Gwendoline Bellement, Thomas Julou &nbsp;and Erik van Nimwegen.</p> <p>A preprint is available under the following DOI:&nbsp;<a href="https://doi.org/10.1101/795369">https://doi.org/10.1101/795369</a>.&nbsp;</p> <p>The repository consists of&nbsp;the following datasets:&nbsp;</p> <p><strong>1. preprocessed_datasets.zip(~22GB)</strong></p> <ul> <li>This dataset contains&nbsp;raw data from the flow cytometry experiments (FACS Canto II, BD Bioscience)&nbsp;in all measured&nbsp;conditions&nbsp;in RData format. Raw fcs files&nbsp;were&nbsp;processed with&nbsp;the&nbsp;tools described in the publication&nbsp;&#39;&#39;Using fluorescence flow cytometry data for single-cell gene expression analysis in bacteria&quot; published here:&nbsp;<a href="https://doi.org/10.1371/journal.pone.0240233">https://doi.org/10.1371/journal.pone.0240233</a>. The tools themselves are&nbsp;available here:&nbsp;<a href="https://github.com/vanNimwegenLab/E-Flow">https://github.com/vanNimwegenLab/E-Flow</a>.&nbsp;&nbsp;Included in the files are&nbsp;the outputs of these processing tools together with&nbsp;all raw values&nbsp;that&nbsp;came&nbsp;directly&nbsp;from the flow cytometer. The file&nbsp;<em>directory_structure_in_preprocessed </em>contains information about how the files are organized.</li> </ul> <p><strong>2.&nbsp;info_files:&nbsp;</strong>This is a set of&nbsp;csv files&nbsp;containing&nbsp;detailed information about the experiments done to acquire the&nbsp;preprocessed_datasets&nbsp;as well as annotation files&nbsp;that we&nbsp;used to retrieve promoter information.&nbsp;</p> <p><strong>3. processed_datasets:</strong>&nbsp;These files correspond to the&nbsp;processed datasets from the raw Rdata files&nbsp;under 1 above.&nbsp;&nbsp;The processed data provide&nbsp;mean and variance estimates in fluorescence&nbsp;of&nbsp;E.coli promoters&nbsp;across&nbsp;the&nbsp;different&nbsp;growth&nbsp;conditions.&nbsp;Note that we discarded &nbsp;flow cytometry measurements from&nbsp;promoter/growth-condition combinations that&nbsp; contained&nbsp;abnormal&nbsp;fluorescence&nbsp;distributions (due to contamination) as well as measurements from reporters&nbsp;with annotation mismatches. The folder contains the following clean dataset&nbsp;files that were&nbsp;used in the paper:</p> <ul> <li><strong>FULL_dataset_mean_var_wreplicates:</strong>&nbsp;In this dataset we include the processed means&nbsp;and variances&nbsp;(in&nbsp;both&nbsp;logarithmic&nbsp;and linear scale) of all&nbsp; promoters in each condition. Included as well are&nbsp;replicate measurements&nbsp;for some conditions..&nbsp;We also include the name and&nbsp;Blattner number of the gene immediately downstream of each promoter,&nbsp;the&nbsp;DNA&nbsp;sequence&nbsp;of each promoter,&nbsp;and regulatory information (number of unique inputs for transcription factors sites and their names)&nbsp;which we obtained from&nbsp;RegulonDB v 10.5 (<a href="https://doi.org/10.1093/nar/gky1077">https://doi.org/10.1093/nar/gky1077</a>).&nbsp;</li> <li><strong>dataset_with_noise_estimates:&nbsp;</strong>In this dataset we&nbsp;provide noise estimates for&nbsp;all&nbsp;promoters expressed above an expression&nbsp;threshold&nbsp;(mean GFP fluorescence at least as large as autofluorescence).&nbsp;Note that the noise estimate correspond to the difference between the promoter&rsquo;s variance in log-expression and the minimal variance as a function of its mean expression (i.e. the so called noise floor was subtracted).&nbsp;Apart from the mean, variance, noise and promoter features (sequence, name of gene downstream,&nbsp;number of unique regulatory inputs and&nbsp;name of the TFs binding), we also include the parameters used for fitting the&nbsp;minimal&nbsp;noise, i.e. noise floor,&nbsp;&nbsp;in each of the&nbsp;conditions.&nbsp;</li> <li><strong>time_course_data_SI</strong>: This dataset contains mean and variance measurements of one of the plates of the library measured at different time points during growth in Minimal media 0.4M NaCl: 0h (just after dilution),&nbsp;1h, 2h, 3h, 5h, 6.5h, 8.5h, 10h and&nbsp;11h.&nbsp;</li> <li><strong>growth_curves_SI</strong>:&nbsp;Growth data (OD₆₀₀&nbsp;as a function of time)&nbsp;for&nbsp;a subset of&nbsp;the&nbsp;promoters&nbsp;from&nbsp;the library&nbsp;across&nbsp;different&nbsp;growth&nbsp;conditions.</li> <li><strong>singlecell_areas_SI:&nbsp;</strong>Single-cell areas&nbsp;estimated using agar patches of cells growing in each&nbsp;condition. Each row&nbsp;of the table&nbsp;contains data for&nbsp;a single-cell.&nbsp;</li> <li><strong>synthetic_promoters_dataset:&nbsp;</strong>This dataset contains mean, variance and noise measurements of a set of constitutive promoters from&nbsp; <a href="https://doi.org/10.7554/eLife.05856.001">https://doi.org/10.7554/eLife.05856.001</a>&nbsp;across different conditions.</li> <li><strong>MARA_results:</strong>&nbsp;&nbsp;All transcription factor activities results explaining measured noise levels in each condition. This data has been obtained after performing Motif Activity Response Analysis on the noise levels of all measured promoters in each condition.</li> </ul>