Machine Learning Suggests That Small Size Helps Broaden Plasmid Host Range

Area of Science:

• Microbiology
• Genetics
• Bioinformatics

Background:

• Plasmids drive bacterial evolution and gene exchange, impacting antibiotic resistance.
• Understanding plasmid spread is crucial for addressing the antibiotic resistance crisis.

Purpose of the Study:

• To identify key features influencing plasmid host range expansion using machine learning.
• To develop predictive models for plasmid transmissibility across bacterial communities.

Main Methods:

• Assembled a dataset of over 30,000 bacterial plasmids, clustered them, and scored their distribution potential.
• Utilized a model stack including DecisionTreeRegressor, EvoTreeRegressor, LGBMRegressor, and LinearRegressor.
• Incorporated plasmid features such as sequence length, P-loop NTPases, mobility factors, and beta-lactamases.

Main Results:

• Sequence brevity was identified as the most significant factor determining plasmid spread.
• P-loop NTPases, mobility factors, and beta-lactamases were also important determinants of plasmid host range.
• Small plasmids may overcome host defenses and utilize alternative transfer mechanisms.

Conclusions:

• Small plasmid size is a critical factor for broad host range and dissemination.
• Identifying plasmid features that promote spread can inform strategies against antibiotic resistance.
• Further research into alternative plasmid transfer mechanisms is warranted.