Prediction of multi-drug resistance transporters using a novel sequence analysis method

Area of Science:

• Biochemistry
• Bioinformatics
• Microbiology

Background:

• Protein functional specificity is challenging to predict from sequence alone.
• Transporter proteins exhibit diverse substrate specificities that are difficult to ascertain from sequence data.
• Current methods struggle to accurately predict transporter substrate specificity.

Purpose of the Study:

• To develop a novel linguistic-based approach for identifying functional patterns in unaligned protein sequences.
• To apply this method for predicting multi-drug resistance transporters (MDRs) in bacteria.
• To assess the performance of the new method against existing databases and apply it to environmental microbiome data.

Main Methods:

• A linguistic-based computational method was developed to analyze unaligned protein sequences.
• The method, termed MDRpred, was validated by recreating known PROSITE patterns.
• MDRpred's predictive performance for MDRs was compared against Pfam family-based models.

Main Results:

• The linguistic method successfully recreated known protein motifs from unaligned sequences.
• MDRpred demonstrated higher accuracy and positive predictive value for MDR prediction compared to Pfam models.
• Novel predictions of drug resistance were made for proteins within an environmental microbiome.

Conclusions:

• A linguistic-based approach offers a powerful new tool for predicting protein function, specifically transporter specificity.
• MDRpred provides a more accurate method for identifying bacterial multi-drug resistance transporters.
• This research highlights the potential for discovering drug resistance mechanisms in environmental microbial communities.