328 - smFISH Analysis using Big FISH library in python

328 - smFISH Analysis using Big FISH library in python

Code from this video is available at:
1. https://github.com/bnsreenu/python_for_microscopists/blob/master/328a_smFISH_analysis_using_Big_FISH_singleplex.ipynb

2. https://github.com/bnsreenu/python_for_microscopists/blob/master/328b_smFISH_analysis_using_Big_FISH_multiplex.ipynb

This video tutorial is a walkthrough of smFISH (single-molecule fluorescence in situ hybridization) analysis using the Big-FISH library. The code demonstrates various steps involved in the analysis, including image reading, normalization and filtering, spot detection, segmentation of nuclei and cells, extraction of cell-level results, and computation of features for each cell - in a singleplex (and multiplex) dataset.

By analyzing the extracted features and the spatial distribution of spots within cells, researchers can gain insights into various aspects of cellular processes, including gene expression, RNA localization, and spatial organization. These insights can contribute to understanding the functional organization of cells and uncovering potential relationships between gene expression patterns and cellular phenotypes.

The "spots" in smFISH analysis refer to individual mRNA molecules that have been labeled with fluorescent probes and can be visualized as discrete signals in the images. Each spot represents the presence of a specific mRNA molecule within the cell.

The brightness of a spot generally corresponds to the abundance or level of the mRNA molecule it represents. Bright spots indicate a higher concentration of the mRNA molecule, suggesting higher expression levels of the corresponding gene. Conversely, dimmer spots may indicate lower expression levels.

"Clusters" refer to groups of spots that are in close proximity to each other. Clusters can arise due to various reasons, such as multiple mRNA molecules originating from the same gene or co-localization of mRNA molecules from different genes. The presence of clusters may indicate co-regulation or co-localization of specific mRNA molecules within the cell.

Spots that are classified as "inside" the nucleus are localized within the nuclear boundary, indicating that the corresponding mRNA molecules are likely involved in nuclear processes, such as transcription, splicing, or RNA processing. On the other hand, spots classified as "outside" the nucleus are located in the cytoplasm, suggesting that the corresponding mRNA molecules have been transported out of the nucleus and are involved in cytoplasmic processes, such as translation.

Analyzing the distribution of spots inside and outside the nucleus can provide insights into gene expression regulation and mRNA localization. For example, certain genes may exhibit preferential nuclear localization, indicating their involvement in nuclear processes. On the other hand, cytoplasmic localization may be associated with mRNA molecules that are actively being translated or are involved in cytoplasmic functions.

Researchers often work with fluorescence in situ hybridization (FISH) images that involve multiple channels containing spots from multiple RNA molecules. FISH techniques can be designed to target specific RNA molecules of interest using fluorescently labeled probes. Each RNA molecule can be labeled with a different fluorophore, allowing researchers to distinguish and visualize multiple RNA species simultaneously.

There are multiple data sets to play with but for this exercise we will be using two datasets.

1. The first one (singleplex) involves example images provided via the big-fish library.

Running this line: stack.check_input_data(path_input, input_segmentation=True) will place four images in your directory of choice.
experiment_1_dapi_fov_1.tif experiment_1_smfish_fov_1.tif example_nuc_full.tif example_cell_full.tif

experiment_1_dapi_fov_1.tif
experiment_1_smfish_fov_1.tif
example_nuc_full.tif
example_cell_full.tif
We will be primarily working with the first two images where dapi is used to segment nuclei and smfish image gets used to segment cells and spot detection. As mntioned above, in real situations you may be working with multichannel images representing signals from multiple RNA molecules.

2. The second one (multiplex) can be downloaded from: https://github.com/LieberInstitute/dotdotdot/blob/master/images/Mouse1.czi

The z-stack consists of 14 z slices, each 201 x 201 pixels and 4 channels: "Cy5", "DsRed" (red), "EGFP" (green), and "DAPI" (blue) in that order. Scaling is 0.31 um x 0.31 um x 0.40 um. The data was collected on ZEISS LSM700, AxioObserver microscope with plan Apochromat objective at 40x/1.3 oil DIC.

Other datasets of use:
This is a good reference paper that mentions a few datasets: https://static-content.springer.com/esm/art%3A10.1038%2Fs41592-022-01669-y/MediaObjects/41592_2022_1669_MOESM1_ESM.pdf