iMPP

iMPP integrates hybrid assembly graph generation, graph-based gene calling, and peptide assembly-based refinement to improve de novo prediction of genes and peptides from metagenomic FASTQ reads for recovery of novel protein families and reconstruction of metabolic pathways in complex microbial communities.


Key Features:

  • Hybrid Assembly Graph Generation Module: Creates a comprehensive assembly graph by integrating fragmentary reads to provide a foundation for downstream gene calling and peptide assembly.
  • Graph-Based Gene Calling Module: Leverages the assembly graph to predict genes from unassembled fragmented reads, reporting recall of 92% to 97% while maintaining precision above 90%.
  • Peptide Assembly-Based Refinement Module: Refines peptide assembly using graph-based predictions to recover both reference proteins and hypothetical protein sequences.
  • Implementation and I/O: Implemented in C++, tested on 64-bit Linux, accepts FASTQ input and outputs predicted genes (nucl), predicted peptides (prot), gene predictions in GFF format, and assembled peptide sequences (prot).

Scientific Applications:

  • Recovery of novel protein families: Enables discovery and recovery of novel protein families from complex metagenomic datasets.
  • Metabolic pathway reconstruction: Facilitates reconstruction of metabolic pathways from community sequence data.
  • Functional profiling of microbial communities: Supports de novo functional analysis of high-complexity microbial communities and their interactions with environments or hosts.
  • Detection of hypothetical proteins: Improves sensitivity for recovering hypothetical protein sequences absent from reference databases.

Methodology:

Hybrid assembly graph generation by integrating fragmentary reads; graph-based gene calling on the assembly graph from unassembled fragmented reads; peptide assembly-based refinement of graph-based predictions; implemented in C++; accepts FASTQ input and outputs predicted genes (nucl), predicted peptides (prot), gene predictions in GFF, and assembled peptide sequences (prot); tested on 64-bit Linux.

Topics

Details

License:
GPL-3.0
Cost:
Free of charge
Tool Type:
command-line tool
Operating Systems:
Mac, Linux, Windows
Programming Languages:
C++, Perl, Python, Shell
Added:
2/24/2022
Last Updated:
2/24/2022

Operations

Publications

Thippabhotla S, Liu B, Yooseph S, Yang Y, Zhang J, Zhong C. Integrated <i>de novo</i> Gene Prediction and Peptide Assembly of Metagenomic Sequencing Data. Unknown Journal. 2021. doi:10.1101/2021.09.20.461079.