SpaceHack2022-Xenium

Our team work on Xenium data for the SpaceHack 2022

Resources

• SpatialData quick start

Slides

• Google slides for results

Biohakcathon 2022::Xenium omni-workflow (a potential plan)

infrastructure --> Luca and Gleb | Txsim/MCMICRO

• having a single step-wise workflow
• effect of method choice for each step

Segmentation refinement by integration --> Gleb and Art1m | Txsim/MCMICRO

• Other segmentation methods
• Using the H&E image

Robustness --> Luca and Gleb | Txsim/MCMICRO

• robustness to noise at each step
• Simulation by subseting
• Simulation by purturbing segmentation boundries
• Xenium to ISS transformation simulation

Initial QC metrics --> Elyas and Art0m | Txsim/MCMICRO

• scFFPEseq vs. scRNAseq vs. Xenium
• cell type transferability

Importing Xenium in spatialdata--> Luca, Gleb, and Art1m

• alignment with and deconvolution of visium
• alignment with h&e and ROI selection

Automatic ROI/niche selection from xenium --> Elyas and Art1m

• evaluation with H&E

Gene panel selection --> Elyas and Art0m

• gene module
• spatially variable genes
• Visium deconv.

Cell typing --> Elyas, Art0m, Olga | Txsim/MCMICRO

• defining QC metrics
• maybe look into SpatialTX

Finding clones in space --> Art0m and Olga

• H&E annotation for ducts/domains
• Visium based + Xenium + H&E

Post processing --> Elays and Art0m

• ideas for cell-cell comm.