agreed, TIFFs are not ideal i feel your pain 😅 my background comes more from functional analysis of FISH data so i have avoided dealing with the massive raw imaging data until now. i believe the scverse SpatialData folks would also appreciate OME-NGFF asap 😂 is the reason why your EASI-FISH pipeline is not nf-core yet because of non-standardized file formats? i am almost thinking we could add a barcode decoding module after the EASI-FISH

spot warping

step to support barcode multiplexed spatial data.

We're working on adapting parts of the EASI-FISH pipeline into an nf-core pipeline here: https://github.com/krokicki/nf-lightsheetrecon So far it's just the stitching module, but we plan to continue adding modules. If we end up building an entire "nf-easifish" pipeline and then, as you say, maybe you could use the same modules to build a custom pipeline for your data by just adding the barcode decoding step. I think that's the holy grail of workflows 😁 I think it wouldn't be hard to support TIFF input, if thats what you have. Let me know if you want to experiment with that idea, I'd be happy to discuss.

ah i see! you are making the the easifish stitching module more modular 🤔 and yes TIFF support would be valuable for me. i'd love to chat, i can share some example data with you as well if that helps.

@Clarence Mah, I’m about to leave for vacation, but I’ll be back in a couple weeks and maybe we could chat on Zoom then? If you can share some example data in the meantime, it would definitely help me to start thinking about it.

have a good vacation! the sample merfish and seqfish data found here is a good approximation (if you notice there are many small tiffs as opposed to a few large tiled tiffs, which is not ideal but from what i've seen quite common): https://zenodo.org/record/7909295

Hey @Clarence Mah, we just attended a spatial biology workshop from Stanford (https://www.angelolab.com/spatial-biology-workshop) and David van Valen mentioned that they are also soon releasing a new method for spot calling called Polaris (https://slideslive.com/38970926/polaris-accurate-spot-detection-for-biological-images-with-deep-learning-and-weak-supervision). Just thought this might be interesting for your research and pipeline long-term to see if this can maybe be incorporated in your pipeline for spot calling beside established tools like RS-FISH.

thank you for the great find! disappointed i couldn't attend this one