Symbiont-Screener

Symbiont-Screener separates host sequencing reads from contaminants in symbiont and biomedical sequencing datasets without relying on reference genomes, enabling reference-free decontamination of long reads and co-barcoded reads.

Key Features:

• Reference-Free Operation: Operates independently of DNA or protein reference genomes, enabling decontamination when reference data are unavailable or incomplete.
• Triobased Classification Methodology: Employs a triobased classification method that targets error-prone long reads and sparse co-barcoded reads to separate host from non-host sequences.
• Haplotype-Specific k-mers Identification: Identifies high-confidence host reads using haplotype-specific k-mers inherited from parental genomes.
• Unsupervised Clustering: Groups remaining reads by unsupervised clustering to segregate contaminant sequences.
• Performance Metrics: Demonstrates up to 97.38% classification accuracy for human long reads with 99.9999% precision, and 79.95% accuracy with 98.36% precision for host co-barcoded reads on artificially mixed data.
• Real-World Decontamination and Assembly Improvement: Successfully decontaminated real algae datasets, reconstructing two haplotypes and improving genome assembly quality by increasing NGA50 and reducing misassemblies.

Scientific Applications:

• Microbiome Research: Facilitates accurate analysis of host–microbe interactions by producing decontaminated sequencing datasets.
• Genomic Studies: Improves the quality and reliability of genome assemblies in organisms with complex symbiotic relationships.
• Biomedical Discoveries: Enhances the precision of genomic data used in medical research, potentially supporting more accurate diagnostics and treatments.

Methodology:

Uses a triobased classification approach, identifies haplotype-specific k-mers from parental genomes to mark high-confidence host reads, and applies unsupervised clustering to group and separate remaining reads.