Towards Identify Selective Antibacterial Peptides Based on Abstracts Meaning.

Area of Science:

• Biochemistry
• Computational Biology
• Drug Discovery

Background:

• The discovery of antibacterial peptides is crucial for developing new antibiotics.
• Researchers report antibacterial peptide findings in diverse natural language formats, often omitting peptide sequences.
• Existing databases like APD2 may not capture all reported antibacterial peptides.

Purpose of the Study:

• To propose and validate the Identify Selective Antibacterial Peptides (ISAP) approach for discovering antibacterial peptides based on abstract semantics.
• To identify antibacterial peptides reported in literature that are not currently listed in the Antimicrobial Peptide Database 2 (APD2).

Main Methods:

• The ISAP approach involves two stages: training and discovering.
• Natural language processing techniques are used to analyze the meaning of words within research abstracts.
• The ISAP model was trained and tested on a sample of PubMed abstracts.

Main Results:

• ISAP identified antibacterial peptides in 35% of the initial abstract sample.
• ISAP predicted antibacterial peptides in 45% of abstracts upon testing against APD2.
• The study identified 810 potential antibacterial peptides not currently classified or reported in APD2.
• 20% of abstracts showed no semantic relation to the APD2 database.

Conclusions:

• The ISAP approach effectively identifies and predicts antibacterial peptides by analyzing abstract meaning, offering a novel method for discovery.
• ISAP can significantly expand the catalog of known antibacterial peptides, serving as a valuable complement to existing databases like APD2.
• This semantic-based tool has the potential to accelerate the identification of new antimicrobial candidates for drug development.