PlasFlow

PlasFlow predicts plasmid-derived sequences in assembled metagenomic contigs to enable characterization of plasmidomes and study of horizontal gene transfer and microbial adaptation.

Key Features:

• Genomic Signatures: Leverages genomic signatures as discriminative features to identify plasmid sequences within complex environmental metagenomic assemblies.
• Neural Network Classifier: Employs neural network methodologies for sequence identification and classification of contigs as plasmid or chromosomal.
• High Accuracy: Reports up to 96% accuracy in identifying plasmid sequences from assembled metagenomes without requiring prior knowledge of sample taxonomical or functional composition.
• Versatility in Plasmid Types: Capable of recovering both circular and linear plasmid sequences from metagenomic data.
• Taxonomical Classification: Provides initial taxonomical classification of predicted plasmid sequences to aid downstream interpretation.
• Performance on Diverse Samples: Demonstrates ability to recover large plasmids and perform in high-diversity environmental metagenomes.

Scientific Applications:

• Plasmidome Profiling: Characterizing plasmid pools (plasmidomes) in environmental metagenomic samples.
• Environmental Microbiology: Analysis of plasmids from heavy metal-contaminated microbial mats to identify plasmid-borne genes relevant to heavy-metal homeostasis.
• Horizontal Gene Transfer Studies: Enabling investigation of plasmid-mediated gene transfer and microbial adaptation in complex ecosystems.
• Recovery from High-Diversity Metagenomes: Detecting plasmids, including large elements, in high-diversity environmental assemblies.

Methodology:

Uses genomic signatures as input features for a neural network-based classifier to predict plasmid sequences in assembled metagenomic contigs and provides initial taxonomical classification.