Accelerated Antibacterial Coating And Antimicrobial Discovery Via High-throughput Method

Antibiotics are useful for bacterial infection treatment.However,issues like infections of medical equipment and drug-resistant bacteria are hardly treated by antibiotics directly.Particularly,the first issue can be solved by developing antibacterial coating.While the second issue depends on discovery of new antibacterial agents.Unfortunately,the low-efficient trial-and-error method limits the development of antibacterial coating and new drug discovery.The date-driven method,derived from the Materials Genome Engineering,gives us a new sight about accelerating material development.This new method consists of 2parts.Firstly,high-throughput experiment is used,replacing traditional trial-and-error one,to acquire amounts of data.Secondly,data mining is used to build the structure-function relationship of material,based on the data acquired in the first step.As a result,this new method can accelerate material design.For accelerating antibacterial material design,this work explores the establishment of experimental high-throughput screening method and data-driven computing method:1.Experimental high-throughput screening chip for accelerating antibacterial drug-loaded coating design.This work explores the development of a micro-droplet drug-loaded coating array chip based on ultrasonic spray technology for high-throughput screening of drug-loaded antibacterial coating.The chip consists of continuous superhydrophobic substrate and discrete arrays of hydrophilic coatings with gradient drug loading.The superhydrophobic substrate(～154°)is formed by ultrasonic spraying PDMS/PMMA mixed solution on a PMMA surface.Then,a metal template with hollow array pattern is covered on the superhydrophobic substrate,and poly(lactic-co-glycolic acid)(PLGA)is sprayed to obtain hydrophilic coating array.By controlling the times of spraying drug(colistin)solution,the chip obtains gradient drug loading among different rows of coating arrays.The chip is used to screen the appropriate dosage of antibacterial coating for killing P.aeruginosa.It is indicated that at drug loading of 140μg/cm~2,the antibacterial coating can kill bacterial without causing cytotoxicity.2.Data-driven calculating method for accelerating antimicrobial peptide(AMP)design.This work explores the establishment of a machine learning pipeline for AMP design with integrating the process of expert knowledge,classification model,ranking model,regression model,incremental learning correction,and wet experiment verification.The aim of this task is to discover new AMP from hexapeptides including64,000,000 candidates.Before the task,we trained the classification model,ranking model,regression model with about 7000 peptide data labeled with antibacterial properties(minimum inhibitory concentration,MIC)in database.And in the task,in order to avoid the interference of a large number of non-antibacterial peptides in candidates and reduce the computational pressure of the machine learning models,we set up two terminal conditions for peptide screening.3,930,000 peptides are acquired in this step.Then,they are classified(antibacterial or non-antibacterial)by a classification model with precision of 93%and 560,000 peptides are predicted positive.Thereafter,560,000 peptides are ranked by their antibacterial ability via a ranking model and the top-500 peptides are chosen.Considering the data noise of labeled AMP,we randomly synthesize 67 AMPs and measure their MIC in our laboratory for model modification via incremental learning.Finally,the MICs of the top-500 peptides are predicted by the modified regression model and the top-10 peptides are selected for wet experiment verification.It is indicated that all 10 peptides are antibacterial with MIC<160μg/μL.Especially,the MICs of 3 best peptides(CRRI AMPs)are lower than 10μg/μL,better than that of best antibacterial hexapeptide MP196(32μg/m L).These results evidence the performance of machine learning pipeline adequately.3.Characterization of antibacterial and other biological functions of antibacterial peptides.This work explores the antibacterial properties and other biological functions of 3 best CRRI AMPs in detail.Cytotoxicity and hemolytic toxicity tests prove biocompatibility of CRRI AMPs.The antibacterial activity test against various clinically isolated multi-drug resistant bacteria(including Gram-positive bacteria S.aureus,MRSA and Gram-negative bacteria:P.aeruginosa,A.baumannii)indicates the broad-spectrum antibacterial property of CRRI AMPs.Compared with MP196,which is currently recognized as the strongest AMP among hexapeptides,CRRI AMPs show stronger antimicrobial activity against a variety of bacterias,which further evidences the performance of our machine learning pipeline above.The antimicrobial activity test of AMPs in a complex environment characterizes the tolerance of CRRI AMPs to high-salt environment.Flow cytometry,SEM and TEM tests characterize the antibacterial mechanism of CRRI AMPs of destroying cell membrane.Drug resistance test demonstrates that CRRI AMPs are not easy to cause bacterial resistance.Finally,we use mouse acute pneumonia and chronic pneumonia bacterial infection models to prove that CRRI AMPs have almost the same therapeutic effect as penicillin,which can kill more than 95%of lung bacteria and reduce lung inflammation.Above results evidence the biological functions and clinical potential of CRRI AMPs forcefully.