This repostory contains the scripts used in processing the images and barcode calling of seqFISH+ experiment and the source data which consists of the processed data. The cell type annotations zip file contains the annotations for each single cells with their Louvain clusters number. The gene ID for each transcript location in the mouse brain data or NIH/3T3 cell data corresponds to the codebook. Additional source data such as the mRNA point locations in NIH/3T3 cells are available on Zenodo at http://doi.org/10.5281/zenodo.2669683.

• Download all the contents of the folder and add it to your Matlab path.

• MATLAB

• radialcenter.m by Dr. Parthasarathy, which can be downloaded here.

Overview of the steps for the seqFISH+ pipeline

• points spread function of a bead
• Region of interests (ROI) holds polygon cell segmentations created from ImageJ software
• thresholding values after Preprocessing Step

• BackCorr_After_Analyze_ALL_channels.m
  • a. grab_ims.m - grabs all the hybridization images in the 3 channels
  • b. grab_ims.m - grab the dapi channel in channel 4
  • c. imregcorr.m - calculate the dapi transformation to the first reference dapi image
  • d. deconvlucy.m - deconvolve the image with the point spread function
  • e. grab_ims_Blanks.m - grab the blank images
  • f. back_correct_multi_pos.m - get the background corrections
  • g. findDotsBarcodeV2.m - find the dots using a laplacian of gaussians and a regional maxima
  • h. radialcenter.m - super resolve dots with radial center
  • i. transformPointsForward.m - apply transformation to dots

• Callbarcodes_ALL_channels.m
  • a. gene_calling.m - colocalize all the points and recieve all the found barcodes
  • b. handle_ambiguous_barcodes.m - match points to genes and remove ambiguous barcodes
  • c. pointLocations.m – create the point locations and lists for each gene
  • d. numseeds.m – calculate the number of seeds for each point
  • e. filterseeds_marked_up.m – filter the points that have seeds 3 or 4 (not in function)

Free for non-commercial and academic research. The software is bound by the licensing rules of California Institute of Technology (Caltech)

• Michael Lawson - Writing and implementing code for the project
• Sheel Shah - Developing the algorithm to find the barcodes, finding dots, and much more
• Nico Pierson - Cleaning and updating code

• Contact the corresponding author: Long Cai ([email protected]) for any inquiry.