Kerlin et al. "Transcriptional coregulation in cis around a contact insulation site revealed by single-molecule microscopy"

Data and analysis scripts used in the study Kerlin et al. entitled "Transcriptional coregulation in cis around a contact insulation site revealed by single-molecule microscopy". In this study, we use single-molecule RNA FISH and oligo-based DNA FISH to investigate the local transcriptional correlations between 3 adjacent genes (FOS, JDP2, and BATF) in human MCF7 breast cancer cells.

Reuse policy: The study Kerlin et al. is currently available as a preprint and has not yet been published in a scientific journal. The data and code in this repository may be downloaded and used, but what results from its usage should not be published until the article Kerlin et al. is published in a scientific journal. If you want to enquire about the publication status of this study, please contact us at antoine.coulon@curie.fr and kyra.borgman@curie.fr.

Description of shared data and code:

• The file analyze_transcriptional_correlations.zip (149 MB) contains the Python scripts, Jupyter notebook, and datafiles (Python .pickle files) to reanalyze the transcriptional correlations between FOS, JDP2, and BATF. It can be used standalone. The .pickle datafiles (10 files) contain the output of the single-sample analysis scripts described below.
  
  
• The file data.zip (49.5 GB) contains 10 folders with raw and processed data from the 10 samples used in our study (2 conditions, 5 replicates each). Due to the 50 GB size limit of Zenodo, we only included the first few raw images for each sample. The full dataset of raw data (1.6 TB) is available upon request (antoine.coulon@curie.fr, kyra.borgman@curie.fr).
  Each of the 10 folders contains:
  • the Jupyter notebook to generate the .pickle file used above (this notebook can be modified to re-generate the .pickle file with different parameters, without having to download all the raw data)
  • a few raw image files (.tif files)
  • a folder out, which contains the output of our spot detection and nuclear segmentation scripts, available at https://github.com/CoulonLab/FISHingRod.
  • two parameter files used by the scripts.
    
    
• The file bioinfo_GROseq_Hi-C.zip (11 MB) contains a Jupyter notebook and a .pickle datafile to analyse transcriptional correlations from public GRO-seq and Hi-C data, respectively from Hah et al. and Rodriguez et al.. The part of the notebook that generates 2D histograms of transcriptional correlation vs genomic distance (see our preprint Kerlin et al.) can be re-run directly using the provided .pickle file. Regenerating this .pickle file and plotting Hi-C maps necessitate files that can be downloaded or re-generated using the data provided by Hah et al. and Rodriguez et al., or that we can alternatively provide upon request.

Please refer to the preprint for details on how this data was generated and analyzed. Note that our datasets also include an extra fluorescence channel not use in our publication but included here for potential future reuse: In addition to the DNA FISH channel marking the whole locus (chr14:75,302,947-76,473,730, hg19 coordinates) using a fluorescent primary oligonucleotide library labeled with a 6-FAM Fluorescein dye, we also included a secondary oligonucleotide, labeled with Alexa Fluor 750 dye, targeting a ~20 kb region (chr14:75,762,762-75,783,466) corresponding to the contact insulation site. 

Conventions and nomenclature:

• sc2, sc4, sc5, sc6, and sc7 in file names refer to the 5 single-cell derived cell populations that we used as replicates.
• Q670, CAL610, and Q570 refer to the fluorophores used for RNA FISH (Quasar® 670, CAL Fluor Red 610, and Quasar® 570). FAM and AF750 refer to the fluorophores used for DNA FISH (6-FAM Fluorescein and Alexa Fluor 750).
• The order of the channels in the raw image files are from the red-most fluorophore to the blue-most fluorophore, following the order described by the file name (AF750, Q670, CAL610, ...), and finishing by the DAPI channel.