Nubeam-dedup

Nubeam-dedup performs reference-free de-duplication of sequencing reads to identify and remove duplicate reads and improve sequencing data quality for downstream analyses such as variant calling, transcriptome analysis, and metagenomics.

Key Features:

• Reference-Free De-Duplication: Operates independently of genomic references, enabling use with non-model organisms or novel genomes where reference sequences are unavailable.
• Efficient Resource Utilization: Uses 50–70% less CPU time and 10–15% less RAM compared to other state-of-the-art reference-free de-duplication tools.
• Matrix-Based Representation: Transforms sequencing reads into products of matrices and assigns each read a unique identifier to facilitate efficient duplicate identification.
• Collisionless Hash Function: Applies a collisionless hash function to ensure accurate identification and removal of duplicate reads without hash collisions.

Scientific Applications:

• Variant Calling: Improves input read sets for variant calling by removing duplicate reads that can bias allele frequency estimates.
• Transcriptome Analysis: Reduces technical duplicates in RNA-seq data to improve quantification of transcripts.
• Metagenomics: Enables de-duplication in metagenomic datasets without reliance on reference genomes for diverse microbial communities.
• Non-Model Organism Genomics: Supports analyses in species lacking reference genomes by performing de-duplication independently of reference sequences.

Methodology:

Sequencing reads are transformed into products of matrices and assigned unique identifiers using a collisionless hash function to perform precise reference-free de-duplication.