Antibiotic resistance is a global health threat that promises to increase pathogen infection-related deaths to 10 million by 2050. Current methods to treat bacterial infections are effective but exacerbate antibiotic resistance. Treatments that are host-specific and show low susceptibility to bacterial resistance include phage endolysins. Endolysin efficacy is appropriate but expensive, requires heavy machinery, and is unscalable. LyseDevice generates novel lysin amino acid sequences and protein structures to specifically target pathogenic bacteria through deep neural networks. The lysins are artificially synthesized through double-domain automated flow chemistry for rapid patient delivery. LyseDevice will enable the rapid on-demand production of individualized treatments for infections, thus resolving the antimicrobial resistance crisis while saving an estimated $340 billion yearly.
Abridged description: Antibiotic resistance is a considerable threat to global health. The LyseDevice is a rapid, low-cost system that uses deep neural networks to generate novel lysin amino acid sequences to target specific and personalized human pathogens. A Convolutional Neural Network will predict distinct lysin protein domains. Artificial lysin synthesis will result in precision drug administration. This work may diminish increased antimicrobial resistance by reducing generalized antibiotic use.