varseek 0.1.1

Efficient variant screening of RNA-seq and DNA-seq data using k-mer-based alignment against a reference of known variants.

varseek

varseek is a free, open-source command-line tool and Python package that provides variant screening of RNA-seq and DNA-seq data using k-mer-based alignment against a reference of known variants. The name comes from "seeking variants" or, alternatively, "seeing k-variants" (where a "k-variant" is defined as a k-mer containing a variant).

The two commands used in a standard workflow are varseek ref and varseek count. varseek ref generates a variant-containing reference sequence (VCRS) index that serves as the basis for variant calling. varseek count pseudoaligns RNA-seq or DNA-seq reads against the VCRS index and generates a variant count matrix. The variant count matrix can be used for downstream analysis. Each step wraps around other steps within the varseek package and the kb-python package, as described below.

The functions of varseek are described in the table below.

Description	Bash	Python (with import varseek as vk)
Build a variant-containing reference sequence (VCRS) fasta file	vk build ...	vk.build(...)
Describe the VCRS reference in a dataframe for filtering	vk info ...	vk.info(...)
Filter the VCRS file based on the CSV generated from varseek info	vk filter ...	vk.filter(...)
Preprocess the FASTQ files before pseudoalignment	vk fastqpp ...	vk.fastqpp(...)
Process the variant count matrix	vk clean ...	vk.clean(...)
Analyze the variant count matrix results	vk summarize ...	vk.summarize(...)
Wrap vk build, vk info, vk filter, and kb ref	vk ref ...	vk.ref(...)
Wrap vk fastqpp, kb count, vk clean, and vk summarize	vk count ...	vk.count(...)
Create synthetic RNA-seq dataset with variant-containing reads	vk sim ...	vk.sim(...)

After aligning and generating a variant count matrix with varseek, you can explore the data using our pre-built notebooks. The notebooks are described in the table below.

Description	Notebook
Preprocessing the variant count matrix	3_matrix_preprocessing.ipynb
Sequence visualization of variants	4_1_variant_analysis_sequence_visualization.ipynb
Heatmap visualization of variant patterns	4_2_variant_analysis_heatmaps.ipynb
Protein-level variant analysis	4_3_variant_analysis_protein_variant.ipynb
Heatmap analysis of gene expression	5_1_gene_analysis_heatmaps.ipynb
Drug-target analysis for genes	5_2_gene_analysis_drugs.ipynb
Pathway analysis using Enrichr	6_1_pathway_analysis_enrichr.ipynb
Gene Ontology enrichment analysis (GOEA)	6_2_pathway_analysis_goea.ipynb

You can find more examples of how to use varseek in the GitHub repository for our preprint GitHub - pachterlab/RLSRWP_2025.

If you use varseek in a publication, please cite the following study:

PAPER CITATION

Read the article here: PAPER DOI

Installation

pip install varseek

🪄 Quick start guide

1. Acquire a Reference

Follow one of the below options:

a. Download a Pre-built Reference

• (optional) View all downloadable references: vk ref --list_downloadable_references
• vk ref --download --variants VARIANTS --sequences SEQUENCES

b. Make custom reference – screen for user-defined variants

• vk ref --variants VARIANTS --sequences SEQUENCES ...

c. Customize reference building process – customize the VCRS filtering process (e.g., add additional information by which to filter, add custom filtering logic, tune filtering parameters based on the results of intermediate steps, etc.)

• vk build --variants VARIANTS --sequences SEQUENCES ...
• (optional) vk info --input_dir INPUT_DIR ...
• (optional) vk filter --input_dir INPUT_DIR ...
• kb ref --workflow custom --index INDEX ...

2. Screen for variants

Follow one of the below options:

a. Standard workflow

• (optional) fastq quality control
• vk count --index INDEX --t2g T2G ... --fastqs FASTQ1 FASTQ2...

b. Customize variant screening process - additional fastq preprocessing, custom count matrix processing

• (optional) fastq quality control
• (optional) vk fastqpp ... --fastqs FASTQ1 FASTQ2...
• kb count --index INDEX --t2g T2G ... --fastqs FASTQ1 FASTQ2...
• (optional) kb count --index REFERENCE_INDEX --t2g REFERENCE_T2G ... --fastqs FASTQ1 FASTQ2...
• (optional) vk clean --adata ADATA ...
• (optional) vk summarize --adata ADATA ...

Examples for getting started: GitHub - pachterlab/varseek Manuscript: ... Repository for manuscript figures: GitHub - pachterlab/RLSRP_2025