Thread by @preibischs

We are excited to share with you RS-FISH: Precise, interactive and scalable smFISH spot detection in @FijiSc using 3D Radial Symmetry:

biorxiv.org/content/10.110…

(1/8)

Radial Symmetry (RS) localizes single, circular, two-dimensional spots by computing the intersection point of image gradients. For RS-FISH, we extended RS to detect smFISH spots in anisotropic 3D image datasets. (2/8)

We combined RS with robust outlier removal (#RANSAC) to identify the set pixels that support a certain localization error for a specific spot (left/middle), separate close detections (right), and ignore outlier pixels with very high noise. (3/8)

To test the speed and localization performance we created a set of simulated images and found that processing speed is significantly improved while achieving similar localization performance as other software solutions. This is especially useful for large datasets. (4/8)

Importantly, RS-FISH brings smFISH detection to the @FijiSc ecosystem, allowing simple installation and running while providing macro-recording functionality to easily automate smFISH detection on many images using local computers or the cluster/cloud #opensource (5/8)

A link to a video explaining RS-FISH can be found below. It highlights the interactive adjustment of parameters, which we believe is important to make smFISH detection as intuitive and straight-forward as possible: drive.google.com/file/d/1h4VDfx… (6/8)

If you want to try it out, there is a tutorial in the supplements, and check out our GitHub page: github.com/PreibischLab/R…
We are happy for any feedback, questions or requests! (7/8)

This project was a great collaboration between @bellonet, @Laura_Breimann, @le0epste1n, @milkyklim, @kisharrington, @successprocess, and @preibischs from @BIMSB_MDC, @MDC_Berlin, @nyugrossman and @HHMIJanelia (8/8)