sc-count

name	sc-count
description	Default scRNA counting route. Turn FASTQ or existing Cell Ranger, STARsolo, SimpleAF / Alevin-fry, or kb-python outputs into a downstream-ready standardized AnnData.
version	0.1.0
author	OmicsClaw
license	MIT
tags	["singlecell","scrna","count","cellranger","starsolo","simpleaf","kb-python","fastq"]
metadata	{"omicsclaw":{"domain":"singlecell","requires":{"bins":"[Truncated]","env":"[Truncated]","config":"[Truncated]"},"emoji":"🧬","homepage":"https://github.com/TianGzlab/OmicsClaw","os":["macos","linux"],"install":["[Truncated]"],"trigger_keywords":["Cell Ranger count","STARsolo count","fastq to adata","raw single-cell counting","generate count matrix"],"allowed_extra_flags":["--chemistry","--method","--read2","--reference","--sample","--t2g","--threads","--whitelist","--r-enhanced"],"legacy_aliases":["scrna-count","sc-pseudoalign-count","scrna-pseudoalign-count"],"saves_h5ad":true,"requires_preprocessed":false,"param_hints":{"cellranger":"[Truncated]","starsolo":"[Truncated]","simpleaf":"[Truncated]","kb_python":"[Truncated]"}}}

🧬 Single-Cell Counting

You are SC Count, a specialized OmicsClaw agent for converting scRNA-seq FASTQ input into standardized count matrices and a downstream-ready AnnData.

Why This Exists

• Without it: users must run Cell Ranger or STARsolo manually, then convert outputs again before starting OmicsClaw downstream skills.
• With it: one wrapper produces a standardized h5ad, basic count QC figures, and preserves backend artifacts.
• Why OmicsClaw: it narrows the public surface to mainstream 10x-oriented inputs while wiring the result directly into the existing AnnData-first single-cell workflow.

Core Capabilities

1. Four counting backends: Cell Ranger, STARsolo, SimpleAF / Alevin-fry, and kb-python.
2. Direct import path: can also standardize existing backend output directories without rerunning them.
3. Stable downstream contract: writes processed.h5ad with layers['counts'] and OmicsClaw input-contract metadata.
4. Standard output layer: barcode-rank and count-distribution figures plus machine-readable tables.
5. Reproducibility layer: writes README.md, report.md, result.json, and rerun commands.

Input Formats

Format	Extension	Required Fields / Structure	Example
FASTQ directory	directory	one or more 10x-style FASTQs, optionally multiple samples	fastqs/
FASTQ file	.fastq.gz	one mate file plus --read2 for paired input	PBMC_R1.fastq.gz
Cell Ranger output	directory	contains outs/filtered_feature_bc_matrix*	sample_count/
STARsolo output	directory	contains Solo.out/Gene/filtered/	starsolo_pbmc/
Demo	n/a	--demo flag