Machine learning & research
BioInfo Repo Pulls
Cloned/vendored GitHub research repos, wrapped with a Flask frontend on top of the original code.
Repos
4ST-Net
He et al. 2020 (Nature Biomed Eng) — predicts spatial gene expression from H&E slides. Upload a slide, get a 250-gene expression map across the tissue.
Panoptes
Hong et al. 2021 (Cell Rep Med) — multi-resolution InceptionResNet predicting endometrial cancer subtype, molecular subtype, and gene mutations from H&E. Workflow writeup, source browser, and a real train-and-test run in the repo’s own TF1.13/Keras 2.3 environment.
STalign
Clifton et al. 2023 (Nat Commun) — diffeomorphic metric mapping for aligning spatial transcriptomics data. Runs the real vendored LDDMM registration on bundled Visium-capture-to-Visium-capture pairs, with before/after overlays and downloadable registered points.
MEA-seqX
Emery et al. 2025 (Adv. Sci.) — relates HD-MEA electrophysiology to 10x Visium spatial transcriptomics on the same mouse hippocampal slices. Runs the authors’ own official Zenodo demo dataset (real spaceranger capture, real BrainWave-processed recording) through the vendored spatial-alignment and gene-expression code, essentially unmodified.