OmicVerse Single-Cell kb Alignment

Goal

Build a kb reference from FASTA/GTF inputs, then quantify single-cell FASTQs into count matrices with OmicVerse alignment wrappers. This skill stops at the count matrix. The notebook's later QC, preprocessing, scaling, and PCA cells are generic follow-on work and should be handled separately.

Quick Workflow

1. Choose the reference workflow and sequencing technology.
2. Build the reference with ov.alignment.single.ref(...).
3. Run ov.alignment.single.count(...) on matching FASTQs.
4. Validate the returned paths and, if requested, load the emitted adata.h5ad.
5. Hand the count matrix to your downstream single-cell preprocessing workflow if needed.

Interface Summary

• ov.alignment.single.ref(index_path, t2g_path, fasta_paths=None, gtf_paths=None, cdna_path=None, workflow='standard', d=None, k=None, threads=8, overwrite=False, temp_dir='tmp', make_unique=False, include=None, exclude=None, dlist=None, dlist_overhang=1, aa=False, max_ec_size=None, nucleus=False, f2=None, c1=None, c2=None, flank=None, feature=None, no_mismatches=False, distinguish=False, **kwargs) builds the reference bundle.
• ov.alignment.single.count(index_path, t2g_path, technology, fastq_paths, output_path='.', whitelist_path=None, replacement_path=None, threads=8, memory='2G', workflow='standard', overwrite=False, temp_dir='tmp', tcc=False, mm=False, filter_barcodes=False, filter_threshold=None, loom=False, loom_names=None, h5ad=False, cellranger=False, gene_names=False, report=False, strand=None, parity=None, fragment_l=None, fragment_s=None, bootstraps=None, em=False, aa=False, genomebam=False, inleaved=False, batch_barcodes=False, exact_barcodes=False, numreads=None, store_num=False, long_read=False, threshold=0.8, platform='ONT', c1=None, c2=None, nucleus=False, **kwargs) quantifies reads and returns discovered output paths.
• ov.alignment.ref and ov.alignment.count are aliases; ov.alignment.single is the convenient namespace used by the notebook.