Bartender

Bartender clusters barcode sequencing (bar-seq) reads to accurately detect and quantify barcodes and reduce under- and over-clustering artifacts in high-throughput sequencing experiments.

Key Features:

• Modified two-sample proportion test: Uses a two-sample proportion test modified to incorporate cluster size for clustering decisions.
• Sequence similarity integration: Extends traditional sequence similarity-based clustering by combining sequence similarity with statistical tests that consider cluster abundance.
• Cluster size consideration: Incorporates relative abundance information (cluster size) into barcode merging and splitting criteria to reduce spurious merges and splits.
• UMI handling: Supports unique molecular identifier (UMI) processing to distinguish true biological variants from sequencing errors.
• Multiple time point mode: Matches barcode clusters across separate clustering runs to facilitate analysis of time course data.
• Support for pseudo-barcodes: Applicable to neutral random barcodes and pseudo-barcodes such as shRNAs and sgRNAs for screening experiments.

Scientific Applications:

• Microbial and cancer evolution studies: Tracking lineage dynamics using neutral random barcodes in evolutionary experiments.
• High-throughput screening: Quantifying pseudo-barcodes such as shRNAs and sgRNAs in pooled screening experiments.
• Lineage and genotype assays: Assaying large numbers of cell lineages or genotypes within complex cell pools using bar-seq.
• Time course analysis: Following changes in barcode abundance across multiple time points to monitor population dynamics.

Methodology:

Combines sequence similarity-based clustering with a modified two-sample proportion test that incorporates cluster size, includes UMI handling, and provides a multiple time point mode for matching clusters across runs.