Vulture

Vulture enables scalable cloud-based mining of microbial reads in public single-cell RNA sequencing (scRNA-seq) data for meta-analysis of host-microbial interactions.

Key Features:

• Scalable microbial calling: Performs microbial calling for scRNA-seq data with comparable microbial read identification to existing tools.
• Cloud-based architecture: Operates on cloud computing systems with reduced cost compared to local and other cloud-based alternatives.
• Speed optimization: Achieves 66% to 88% faster processing than local tools (PathogenTrack and Venus) and 41% faster than the cloud-based tool Cumulus.
• Cost efficiency: Reduces cloud computing costs by 83% compared to alternative approaches.
• Copy number variation integration: Supports cell subtype-associated microbial enrichment analysis based on inferred copy number variations.

Scientific Applications:

• Pathogen detection in human scRNA-seq data: Identifies pathogens previously overlooked in human single-cell sequencing datasets.
• Host-microbial interaction meta-analysis: Enables meta-analysis of host-microbial studies from large-scale public scRNA-seq data repositories.
• Viral infection profiling: Discovers cell type-specific enrichment of severe acute respiratory syndrome coronavirus 2 in coronavirus disease 2019 patient cohorts.
• Hepatitis B virus analysis: Identifies hepatocyte-only enrichment of hepatitis B virus in hepatocellular carcinoma cohorts with cell subtype-associated patterns.
• Bacterial pathogen detection: Detects Helicobacter pylori-positive cells in gastric cancer patient cohorts.

Methodology:

Vulture is implemented as a cloud-based pipeline for microbial calling in scRNA-seq data; benchmarking experiments compared performance against PathogenTrack, Venus, and Cumulus using public sequencing reads data from scRNA-seq experiments.