Glossary

  • Anchor round - A one-time round taken, usually on a chosen "anchor channel" that has a high Signal-to-Noise Ratio (SNR). All genes of interest are given the same fluorescing dye probe. The anchor round is essential for detecting all spots at once in the same microscope image.

  • Channel - A combination of excitation light of a certain wavelength and specific emission filter. We use multiple channels to distinguish every dye colour (almost always the number of channels is equal to the number of unique dyes). But, a dye can have "bleed through", i.e. brightness in multiple channels from the same dye.

  • DAPI - A dye that fluoresces the nuclei of all cells. It is used to register the round images to one another. The DAPI is also an overlay in the Viewer.

  • Gene code - A sequence of dyes that are assigned to a gene for each sequencing round. Each gene has a unique gene code. For example, if the dyes are labelled 0, 1, 2 and there are 2 sequencing rounds, some example gene codes are 0, 1 (i.e. dye 0 in first round, dye 1 in second round), 1, 2, 0, 2.

  • Notebook - A write-once1 compressed file that stores all important outputs from coppafish. The notebook is used to plot many diagnostics. The notebook contains notebook pages. There is a notebook page for each method section. A notebook can be loaded by from coppafish import Notebook; nb = Notebook("path/to/notebook.npz"). Variables from the notebook can be directly read. For example, you can read the use_tiles variable from the basic_info page by print(nb.basic_info.use_tiles). Each variable also has a description, which can be printed. For example, nb.basic_info.describe("use_tiles").

  • OMP - Stands for Orthogonal Matching Pursuit. It is the final section of the coppafish pipeline. It is coppafish's most sophisticated algorithm for gene calling and is used as a way of untangling genes that overlap on images by assuming that the pixel intensity is a linear combination of each gene intensity. There is no reason to believe that gene intensity would combine non-linearly.

  • Point cloud - A series of spatial pixel positions. Typically used to represent detected spot positions during find spots.

  • PSF - Stands for Point Spread Function and is used during image filtering. The Wiener deconvolution requires a PSF to remove blurring caused by frequencies with a low signal-to-noise ratio. See the Wikipedia article for more details.

  • Sequencing round - An image of the tissue, made up of multiple tiles and sequencing channels. Before each imaging round, the tissue is treated with various solutions to remove the previous DNA probes and then hybridise new ones. Each spot will bind to a specific bridge probe and then fluorescing dye probe, causing it to fluoresce in specific channel(s). The colour of each spot in each round is dictated by its gene identity (identities) and their corresponding gene code(s).

  • Spot - An amplified ball of DNA with a unique barcode specific to each gene. The gene can be determined by looking at the same spot in all sequencing rounds to reveal the gene code. Coppafish takes the raw images of the spots as input and outputs the identity of each gene in situ.

  • Tile - A cuboid subset of the microscope image of size \(n_z \times n_y \times n_x\) in z, y, and x, where \(n_y = n_x\). Typically, \(n_z\sim55\). Usually, all adjacent tiles overlap by \(10\%-15\%\) to give coppafish information on how to best align tiles (see stitch for details).

  1. There are some minor cases of a notebook being "rewritten", see advanced usage. ↩