seqspec

seqspec, short for “sequence specification” (pronounced “seek-speck”), is a file format that describes data generated from genomics experiments. Both the file format and seqspec tool enable uniform processing of genomics data.

Figure 1: Anatomy of a seqspec file.

We have multiple tutorials to get you up and running with seqspec:

1. Learn how to use seqspec to standardize your genomics data preprocessing.

2. Understand how to manipulate seqspec files using the seqspec command-line tool.

Current release¶

seqspec 0.4.0 keeps the Python and Rust implementations aligned around the same core command set.

• seqspec upgrade upgrades 0.3.0 specs to 0.4.0 in both implementations.

• seqspec loads gzipped specs directly, so .yaml.gz works anywhere a spec path is accepted.

• seqspec auth manages host-matched auth profiles for remote resources, and seqspec check / seqspec onlist can use them with --auth-profile.

• seqspec check now emits warnings for valid but risky geometry, such as reads that cover the same declared regions.

• seqspec onlist -s region-type now errors when the same region type appears across multiple reads, so ambiguous joins are explicit.

• seqspec print -f seqspec-html writes a self-contained HTML view of the library and reads.

• seqspec build is deprecated.

Citation¶

The seqspec format and tool are described in this publication. If you use seqspec please cite

Ali Sina Booeshaghi, Xi Chen, Lior Pachter, A machine-readable specification for genomics assays, Bioinformatics, Volume 40, Issue 4, April 2024, btae168.

seqspec was inspired by and builds off of the Teichmann Lab Single Cell Genomics Library Structure by Xi Chen.

Documentation¶

• Install seqspec: docs/INSTALLATION.md

• Learn about the seqspec file format: docs/SEQSPEC_FILE.md

• Learn about the seqspec tool: docs/SEQSPEC_TOOL.md

• Learn about the seqspec specification: docs/SPECIFICATION.md

• Write a seqspec from a simple example: docs/TUTORIAL_SIMPLE.md

• Write a seqspec from a template: docs/TUTORIAL_FROM_TEMPLATE.md

• Write a more complex seqspec: docs/TUTORIAL_COMPLEX.md

• Browse the generated example site: docs/examples/site/assays.html

• Contribute a seqspec : docs/CONTRIBUTING.md

• Watch a YouTube video about seqspec

• Read the manuscript that describes seqspec

Rust implementation¶

• auth : Manage remote authentication profiles.

• build : Deprecated in both CLIs.

• check : Validate seqspec file against specification (verify check)

• find : Find objects in seqspec file

• file : List files present in seqspec file

• format : Autoformat seqspec file

• index : Identify position of elements in seqspec file

• info : Get information from seqspec file

• init : Generate a new empty seqspec file

• insert : Insert regions or reads into an existing spec (TODO: move Input structs to models)

• methods : Convert seqspec file into methods section

• modify : Modify attributes of various elements in seqspec file

• onlist : Get onlist file for elements in seqspec file

• print : Display the sequence and/or library structure from seqspec file

• split : Split seqspec file by modality

• upgrade : Upgrade seqspec file to current version

• version: Get seqspec tool version and seqspec file version

The standalone Rust CLI supports library-ascii, seqspec-ascii, and seqspec-html in seqspec print. seqspec-png remains Python-only for now.