PlasmidHawk improves lab of origin prediction of engineered plasmids using sequence alignment.

Area of Science:

• Synthetic Biology
• Genomics
• Bioinformatics

Background:

• Advances in synthetic biology and genome engineering raise biosafety concerns regarding potential misuse.
• Previous machine learning methods for identifying DNA lab-of-origin showed promise but had limitations in accuracy and interpretability.

Purpose of the Study:

• To develop an accurate and explainable tool for predicting the lab-of-origin for synthetic plasmid DNA sequences.
• To address the limitations of existing machine learning approaches in terms of accuracy, computational cost, and feature interpretability.

Main Methods:

• Development of PlasmidHawk, a novel computational tool for lab-of-origin prediction.
• Utilizing PlasmidHawk to analyze synthetic plasmid DNA sequences and identify predictive sub-sequence signatures.

Main Results:

• PlasmidHawk achieves 76% accuracy in predicting the depositing lab of unknown synthetic DNA sequences.
• The correct lab is identified within the top 10 candidates 85% of the time.
• PlasmidHawk precisely identifies specific sub-sequences responsible for accurate lab-of-origin predictions, offering explainability.

Conclusions:

• PlasmidHawk provides an accurate and explainable solution for predicting the lab-of-origin of synthetic plasmid sequences.
• The tool enhances biosafety by enabling better tracking and understanding of synthetic DNA origins.
• PlasmidHawk represents a significant improvement over previous machine learning-based methods for this task.