Configure a run¶

A configuration is a YAML text file containing your input paths, sample names, and analysis settings. One configuration describes one analysis. It does not contain the reads themselves.

Start with setup¶

After installation, run:

Choose either browser setup:

oncotracer setup --project /absolute/path/to/my-study

Or terminal questions for the same project:

oncotracer setup --terminal --project /absolute/path/to/my-study --backend conda

Replace the project path with yours. In the browser, select inputs, choose Conda and Save configuration and check. In the terminal, answer the questions and finish with save. Both save the commented config/run.yml.

Then check and run that configuration from the terminal:

oncotracer check --config /absolute/path/to/my-study/config/run.yml
oncotracer run --backend conda --config /absolute/path/to/my-study/config/run.yml

Skip this run command if you already started the analysis from setup. For another backend, change --backend explicitly. The setup guide covers sample selection; terminal and headless servers covers unattended commands and remote browser access.

Which settings should I change?¶

Setting	Meaning
`mode`	Sequencing platform: `illumina` or `ont`
`outdir`	Where analysis results go
`lpwgs_root`	Where reusable reference files go
`threads`	CPU worker threads to request
`variant_specimen_type`	`fresh` or `ffpe` for optional variant analysis
`variant_callers`	Platform-compatible variant callers; see the variant guide
`methylation`	Add ONT methylation analysis
`methylation_only`	Skip copy-number analysis when `true`
`methylation_classifier`	`marlin` for leukemia or `sturgeon` for CNS research
`methylation_gpu`	Allow GPU work when `true`; default is CPU
`run_cna_classifier`	Add interpretation of copy-number changes; separate from methylation

Keep the generated caller and bin-size settings for your first run. Leave force: false to reuse completed matching stages when you resume. Use a new outdir for a different analysis so you can compare its results with the original.

Other configuration routes¶

For many FASTQ files and a sample table, use batch setup.
For variant callers, Fresh/FFPE processing and matched normals, use the variant guide.
For methylation tools, model files, and CPU options, use the methylation guide.
To edit YAML yourself, see YAML basics, Illumina settings, or ONT settings.
For an individual advanced field, use the parameter reference.

For CNA analysis, tumor and normal samples are analyzed independently; normal rows are not pooled into a reference. Somatic Strelka2 variant calling requires an explicit tumor-to-normal pairing.