MethScope — pretrained models

Pretrained models for methscope-cli, ultra-fast analysis of sparse DNA methylomes via MRMP (Most Recurrent Methylation Pattern) encoding. Each file is a single self-contained bundle — classifier (.clfx), upscale decoder (.updecx) or deconvolution reference (.msdref) — that carries its own feature definition, so a query .cg runs directly.

methscope fetch hg38/models          # every human model, digest-checked
methscope fetch mm10/models          # every mouse model
methscope fetch -c hg38/models/hg38_celltype_full.clfx   # one file, into the current dir

methscope fetch verifies each file against a digest compiled into the binary, so a given release fetches exactly the models it documents; methscope --version prints the tag it pins. Tags here are additive and never moved: SHA256SUMS at a tag is the file list, and an older methscope keeps fetching the tag it was built against.

What each model is, how it was built, and how accurate it is — the model table, bank vs routing-tree classifiers, the deconvolution references, and the parameterless --framework violation rule — lives in the methscope documentation, Models tab, which is kept in step with the binary. This card is deliberately minimal so that there is one place to maintain.

Small query .cg test fixtures live in zhou-lab/methscope_data.