Orphelia

Orphelia predicts protein-coding genes in short DNA sequences from metagenomic sequencing projects to enable gene-level characterization of uncultivated microbial communities.

Key Features:

• Machine learning models: Uses machine learning to construct fragment length-specific prediction models for short DNA sequences.
• Sequencing-technology tailoring: Provides separate models tailored to chain termination sequencing and pyrosequencing read characteristics.
• Training data: Models are trained on a broad array of annotated genomes.
• High specificity: Optimized for high specificity in gene prediction from short DNA fragments.
• Metagenomic focus: Designed specifically for short reads derived from metagenomic sequencing where assembly into longer contigs is often not possible.

Scientific Applications:

• Gene prediction in metagenomes: Predicts protein-coding genes in short sequencing reads from metagenomic datasets.
• Characterization of uncultivated microbes: Enables exploration of the genetic makeup of diverse, uncultivated microbial communities.
• Annotation of fragmented data: Supports gene-level annotation of fragmented sequencing data generated by chain termination sequencing and pyrosequencing.

Methodology:

Employs machine learning to train fragment length-specific prediction models on a broad set of annotated genomes, with separate models tailored to chain termination sequencing and pyrosequencing and optimized for high specificity.