Metabolism of novel potential syntrophic acetate-oxidizing bacteria in thermophilic methanogenic chemostats.

Acetate is a major intermediate in the anaerobic digestion of organic waste to produce CH4. In methanogenic systems, acetate degradation is carried out by either acetoclastic methanogenesis or syntrophic degradation by acetate oxidizers and hydrogenotrophic methanogens. Due to challenges in the isolation of syntrophic acetate-oxidizing bacteria (SAOB), the diversity and metabolism of SAOB and the mechanisms of their interactions with methanogenic partners are not fully characterized. In this study, the in situ activity and metabolic characteristics of potential SAOB and their interactions with methanogens were elucidated through metagenomics and metatranscriptomics. In addition to the reported SAOB classified in the genera Tepidanaerobacter, Desulfotomaculum, and Thermodesulfovibrio, we identified a number of potential SAOB that are affiliated with Clostridia, Thermoanaerobacteraceae, Anaerolineae, and Gemmatimonadetes. The potential SAOB possessing the glycine-mediated acetate oxidation pathway dominates SAOB communities. Moreover, formate appeared to be the main product of the acetate degradation by the most active potential SAOB. We identified the methanogen partner of these potential SAOB in the acetate-fed chemostat as Methanosarcina thermophila. The dominated potential SAOB in each chemostat had similar metabolic characteristics, even though they were in different fatty-acid-fed chemostats. These novel syntrophic lineages are prevalent and may play critical roles in thermophilic methanogenic reactors. This study expands our understanding of the phylogenetic diversity and in situ biological functions of uncultured syntrophic acetate degraders and presents novel insights into how they interact with methanogens.IMPORTANCECombining reactor operation with omics provides insights into novel uncultured syntrophic acetate degraders and how they perform in thermophilic anaerobic digesters. This improves our understanding of syntrophic acetate degradation and contributes to the background knowledge necessary to better control and optimize anaerobic digestion processes.

Development of a Novel Typing Scheme Based on the Genetic Diversity of Heme/Hemin Uptake System Hmu in Klebsiella pneumoniae Species Complex.

Iron is essential for the survival and reproduction of Klebsiella pneumoniae. Although K. pneumoniae employs multiple types of siderophores to scavenge iron during infections, the majority of host iron is retained within erythrocytes and carried by hemoglobin that is inaccessible to siderophores. HmuRSTUV is a bacterial hemin/hemoprotein uptake system. However, the genetic background and function of HmuRSTUV in K. pneumoniae remain unknown. We collected 2,242 K. pneumoniae genomes, of which 2,218 (98.9%) had complete hmuRSTUV loci. Based on the 2,218 complete hmuRSTUV sequences, we established a novel typing scheme of K. pneumoniae named hmST, and 446 nonrepetitive hmSTs were identified. In hypervirulent lineages, hmST was diversely distributed and hmST1 mainly existed in ST23 strains. In contrast, hmST was less diversely distributed among multidrug-resistant strains. hmST demonstrated greater genetic diversity in hypervirulent lineages and community-acquired and bloodstream-sourced strains. In vitro and in vivo experiments revealed that an intact hmuRSTUV was essential for hemin uptake, playing an important role in bloodstream infections. This study established a novel typing scheme of hmST based on hmuRSTUV providing new insights into identifying and monitoring the emergence of novel virulence evolution in K. pneumoniae. IMPORTANCE Siderophore is a group of low molecular weight compounds with high affinity for ferric iron, which could facilitate bacterial iron consumption. Similarly, hemin/heme scavenged by the hemin uptake system HmuRSTUV usually act as another critical iron source for K. pneumoniae. This study proved that Hmu system significantly promoted the growth of K. pneumoniae in the presence of hemin and played an important role in bloodstream infections. A novel typing scheme named hmST was established, and the genetic diversity of hmuRSTUV loci was analyzed based on a large number of genomes. This study provides new insights into identifying and monitoring the emergence of novel virulence evolution in K. pneumoniae.

Opposite evolution of pathogenicity driven by in vivo wzc and wcaJ mutations in ST11-KL64 carbapenem-resistant Klebsiella pneumoniae.

AIMS: To investigate the in vivo evolution of the mucoid-phenotype of ST11-KL64 carbapenem-resistant Klebsiella pneumoniae (CRKP) isolated from the same patients and gain insights into diverse evolution and biology of these strains. METHODS: Whole genome sequencing and bioinformatic analysis were used to determine the mutation involved in the mucoid phenotype of ST11-KL64 CRKP. Gene knockout, bacterial morphology and capsular polysaccharides (CPS) extraction were used to verify the role of wzc and wcaJ in the mucoid phenotypes. Antimicrobial susceptibility, growth assay, biofilm formation, host cell adhesion and virulence assay were used to investigate the pleiotropic role of CPS changes in ST11-KL64 CRKP strains. RESULTS: Mutation of wzc S682N led to hypermucoid phenotype, which had negative impact on bacterial fitness and resulted in reduced biofilm formation and epithelial cell adhesion; while enhanced resistance to macrophage phagocytosis and virulence. Mutations of wcaJ gene led to non-mucoid phenotype with increased biofilm formation and epithelial cell adhesion, but reduced resistance of macrophage phagocytosis and virulence. Using virulence gene knockout, we demonstrated that CPS, rather than the pLVPK-like virulence plasmid, has a greater effect on mucoid phenotypic changes. CPS could be used as a surrogate marker of virulence in ST11-KL64 CRKP strains. CONCLUSIONS: ST11-KL64 CRKP strains sacrifice certain advantages to develop pathogenicity by changing CPS with two opposite in vivo evolution strategies.

Catecholate siderophore receptor CirA impacts cefiderocol susceptibility in Klebsiella pneumoniae.

Cefiderocol is a cephalosporin antibiotic presenting expanded antimicrobial activity. CirA is a gateway for cefiderocol to enter bacterial cells. We found that cirA1 and cirA198 were primary alleles in Klebsiella pneumoniae. CirA1 exhibited higher iron-transporting ability than CirA198 in iron-limited conditions. The cefiderocol minimum inhibitory concentration (MIC) increased from 0.5 mg/L to 2 mg/L when cirA1 was mutated to cirA198. Consistently, the MIC showed a 4-fold decrease when cirA198 was mutated to cirA1. Therefore, CirA1 has higher capacity to transport siderophores, contributing to increased cefiderocol susceptibility.

Metalloporphyrin hemin modified carbon nanotube decorated titanium carbide with redox catalytic ability for electrochemical determination of hydrogen peroxide and uric acid.

Herein, a functional hybrid named metalloporphyrin hemin modified carbon nanotube decorated titanium carbide (Hemin/CNT/Ti3C2Tx) with redox catalytic ability is studied and applied to electrochemical sensing. The morphology, crystal structure and chemical composition of the hybrid are investigated through field emission scanning electron microscope (FESEM), transmission electron microscope (TEM), X-ray diffractometer (XRD), X-ray photoelectron spectroscopy (XPS), thermogravimetric analysis (TGA), and Brunauer-Emmett-Teller (BET) analysis. Benefiting from the unique multilayer structure, strong enrichment ability, excellent electrochemical performance of CNT/Ti3C2Tx and outstanding catalytic property of hemin, the developed Hemin/CNT/Ti3C2Tx hybrid presents desirable performance for electrochemical sensing. Two independent electrochemical sensing techniques including chronoamperometry (i-t) and differential pulse voltammetry (DPV) are adopted to realize hydrogen peroxide (H2O2) and uric acid (UA) determination, demonstrating the excellent redox catalytic ability of Hemin/CNT/Ti3C2Tx. More importantly, it is successfully applied to real-time monitoring of H2O2 released from living cells, and accurate determination of UA in urine samples is also realized, demonstrating a good practicability.

Comparing core-genome MLST with PFGE and MLST for cluster analysis of carbapenem-resistant Acinetobacter baumannii.

OBJECTIVES: Carbapenem-resistant Acinetobacter baumannii (CRAB) is a prevalent pathogen contributing to hospital infections. Pulsed-field gel electrophoresis (PFGE), multilocus sequence typing (MLST) and core-genome MLST (cgMLST) are frequently used methods to illuminate the nosocomial transmission of CRAB. In this study, we compared the discriminatory power of the three typing methods. METHODS: Antimicrobial susceptibility tests were performed by the broth microdilution and Vitek2 methods. PFGE, MLST and cgMLST were conducted to determine the clonality and phylogenetic relationship of the strains. Whole-genome sequence data were acquired by an Illumina HiSeq 2000, and cgMLST was analysed by the Ridom SeqSphere+ v.7.2.3 software. RESULTS: A total of 149 carbapenem-resistant A. baumannii isolates had 15 different PFGE profiles (A-O type), and 73 of the isolates had related subtypes (A1 and A2), accounting for the majority of type A isolates. The maximum-likelihood phylogenetic analysis based on the cgMLST genes grouped the same PFGE clonal pattern A into nine different clusters. ST_Pasteur grouped all the strains into ST2, whereas ST_Oxford grouped the PFGE clonal pattern A isolates into six STs. In addition, the gdhB allele in the ST_Oxford scheme had two copies in five strains, which complicated the ST_Oxford typing. CONCLUSIONS: cgMLST was more discriminant than PFGE and MLST. CgMLST is the most suitable and comprehensive method for genotyping A. baumannii in surveillance and epidemiological research.

Staphylococcal cassette chromosome mec amplification as a mechanism for ceftobiprole resistance in clinical methicillin-resistant Staphylococcus aureus isolates.

OBJECTIVES: In this study, we evaluated the ceftobiprole (BPR) susceptibilities of 472 methicillin-resistant Staphylococcus aureus (MRSA) isolates, and investigated the mechanisms underlying BPR resistance. METHODS: For all MRSA isolates, BPR MIC was determined by agar dilution. We sequenced the BPR-resistant isolates through Illumina short- and MinION long-read sequencing. We also selected MRSA isolates of ST5, ST59, and ST239, and exposed them to increasing BRP concentrations. The isolated mutants developing BPR resistance were sequenced. RESULTS: A total of 471 MRSA isolates were susceptible to BPR, with MICs ranging from 0.25 to 2 mg/L. Compared with HA-MRSA isolates (MIC50 = 2 mg/L; MIC90 = 2 mg/L), CA-MRSA isolates (MIC50 = 0.5; MIC90 = 2 mg/L) were more susceptible to BPR (p < 0.001). Compared with isolates with staphylococcal cassette chromosome mec (SCCmec) type II or III (MIC50 = 2 mg/L; MIC90 = 2 mg/L), isolates with SCCmec type IV (MIC50 = 1 mg/L; MIC90 = 1 mg/L) or V (MIC50 = 0.5 mg/L; MIC90 = 1 mg/L) were more susceptible to BPR (p < 0.001). Nanopore sequencing revealed two copies of SCCmec repeats in the BPR-resistant MRSA isolate. In addition, SCCmec amplification could be induced by BPR exposure in ST239 MRSA isolates; however, no amplification was observed in the other lineages. The induced BPR-resistant MRSA isolates also acquired mutations in mecA and other genes, such as guaA, guaB, relA, rpoA, and oatA, which were speculated as factors contributing to BPR-resistance development. DISCUSSION: BPR showed significant antibacterial activity against MRSA isolates in China; however, the emergence of a BPR-resistant isolate before its launch was a cause for concern. Multiple genes and pathways are potentially involved in the development of BPR resistance in MRSA, and our data demonstrated the role of nanopore-sequencing in revealing the tandem repeat-mediated resistance mechanism in MRSA.

Emergence of High-Level Cefiderocol Resistance in Carbapenem-Resistant Klebsiella pneumoniae from Bloodstream Infections in Patients with Hematologic Malignancies in China.

Cefiderocol is a novel siderophore cephalosporin exhibiting potent antimicrobial activities. Although cefiderocol has not been approved in China, resistance is emerging. A multicenter study was performed to evaluate the cefiderocol resistance in carbapenem-resistant Klebsiella pneumoniae (CRKP) strains from bloodstream infections in patients with hematologic malignancies in China. Clinical data analysis and whole-genome sequencing were conducted for collected cefiderocol-resistant CRKP strains. CRISPR-Cas9 system was employed to construct site-specific mutagenesis for gene cirA. Plasmid curing and cloning were performed to assess the effect of ß-lactamases on cefiderocol resistance. Total 86 CRKP strains were collected. The MICs of cefiderocol ranged from 0.06 to >256 mg/L. Among four cefiderocol-nonsusceptible strains (4/86, 4.7%), two cefiderocol-resistant strains AR8538 (MIC = 32 mg/L) and AR8416 (MIC > 256 mg/L) were isolated from two patients with acute lymphocytic leukemia (frequency of resistance, 2/86, 2.3%). Metallo- and serine-ß-lactamase inhibitors addition would decrease the MIC of cefiderocol from 32 to 1 mg/L in AR8538, which harbors blaSHV-12, blaDHA-1, and two copies of blaNDM-1 in different plasmids. Avibactam did not impact cefiderocol susceptibility of AR8416, which produces NDM-5. However, we found a deficient CirA in AR8416. Using the same K serotype strain D3, we proved CirA deficiency or carrying NDM individually reduced cefiderocol susceptibility, but their simultaneously existence rendered a high-level cefiderocol resistance. In summary, the resistance of CRKP against cefiderocol is mediated by multiple factors, including the deficiency of CirA, metallo- or serine-ß-lactamases, while a high-level cefiderocol resistance could be rendered by the combined effect of NDM expression and CirA deficiency. IMPORTANCE Cefiderocol-resistant CRKP strains are emerging in bloodstream infections in Chinese patients with hematologic malignancies, although cefiderocol has not been approved for clinical use in China. Our study proved that the resistance of CRKP against cefiderocol is mediated by multiple factors, including the deficiency of CirA, metallo- or serine-ß-lactamases, while a high-level cefiderocol resistance could be rendered by the combined effect of NDM expression and CirA deficiency. As NDM production is one of the most critical mechanisms resulting in carbapenem resistance, it would pose great challenges on the clinical efficacy of cefiderocol in future.

Risk factors for infection and mortality caused by carbapenem-resistant Klebsiella pneumoniae: A large multicentre case-control and cohort study.

OBJECTIVES: To elucidate the predictors of carbapenem-resistant Klebsiella pneumoniae (CRKP) infection and help clinicians better identify CRKP infection at an early stage. METHODS: We conducted a multicentre case-control study of 422 patients with CRKP infection and 948 with carbapenem-susceptible K. pneumoniae (CSKP) infection from March to July 2017. Binary logistic regression was used to identify risk factors for CRKP infection. The subgroups of CRKP respiratory infection, intra-abdominal infection, and bloodstream infection were also evaluated. Patients were followed up for 28 days. Independent risk factors for 28-day crude mortality of CRKP infection were analysed using Cox proportional hazards regression models. RESULTS: Longer stay of hospitalization, stay in the intensive care unit (ICU), previous exposure to antibacterial agents (especially carbapenems, quinolones, aminoglycosides, and tigecycline), invasive procedures, intravascular catheter use, tracheotomy, and admission to ICU in the preceding 90 days were risk factors for CRKP infection. Carbapenem exposure was the only common predictor of different types of CRKP infection. The 28-day crude mortality of CRKP infection was 24.2% and was independently associated with sex, admitted unit, and type of infection. CONCLUSIONS: Strict policies for antibiotic use, cautious decisions regarding the implementation of invasive procedures, and careful management of patients with catheters, especially intravascular catheters, are necessary to handle CRKP infection.

Epidemiology, evolution and cryptic susceptibility of methicillin-resistant Staphylococcus aureus in China: a whole-genome-based survey.

OBJECTIVES: The aim of this study was to investigate the genomic epidemiology of methicillin-resistant Staphylococcus aureus (MRSA) in China to identify predominant lineages and their associations with clinical data and antimicrobial resistance profiles. METHODS: We performed a national prevalence study of patients with S. aureus infections in 22 tertiary hospitals in China from 2015 to 2017. Clinical data from patients and the antimicrobial phenotypes were collected for each isolate. Genome sequencing was performed on a proportion of isolates and a phylogenetic analysis was undertaken. Genotypic and phenotypic ß-lactam susceptibilities were compared. RESULTS: A total of 1900 patients with S. aureus infections were included, of which 40% involved MRSA. Community-associated MRSA (CA-MRSA) infections were 24% of the total isolates. Genomic data showed that more than three-quarters of the MRSA were from three dominant lineages CC239 (25%, 116/471), CC5 (21%, 96/471) and CC59 (33%, 154/471) with CC59 accounting for more than half of the CA-MRSA isolates. Penicillin susceptibility genomic features were observed in 53% (251/470) of MRSA, including almost all of the CC59 (152/154) lineage, and 96% (242/251) of these isolates demonstrated in vitro susceptibility to penicillin or amoxicillin combined with clavulanic acid. Phylogenetic analysis indicated that the CC59 lineage can be divided into six lineages with all Asian CC59 isolates likely arising from an ancestral Mainland China lineage. CONCLUSIONS: This study showed a high prevalence of CA-MRSA in China, largely due to the widespread presence of CC59. As almost all isolates in this lineage possess genetic variants leading to increased ß-lactam susceptibility, we suggest that to improve antibiotic stewardship combinations of penicillins and ß-lactamase inhibitors should be included in the antibiotic susceptibility testing panels used to inform treatment decisions and research undertaken on this combination therapy.

Secondary bloodstream infection in critically ill patients with COVID-19.

OBJECTIVE: Secondary infection, especially bloodstream infection, is an important cause of death in critically ill patients with COVID-19. We aimed to describe secondary bloodstream infection (SBI) in critically ill adults with COVID-19 in the intensive care unit (ICU) and to explore risk factors related to SBI. METHODS: We reviewed all SBI cases among critically ill patients with COVID-19 from 12 February 2020 to 24 March 2020 in the COVID-19 ICU of Jingmen First People's Hospital. We compared risk factors associated with bloodstream infection in this study. All SBIs were confirmed by blood culture. RESULTS: We identified five cases of SBI among the 32 patients: three with Enterococcus faecium, one mixed septicemia (E. faecium and Candida albicans), and one C. parapsilosis. There were no significant differences between the SBI group and non-SBI group. Significant risk factors for SBI were extracorporeal membrane oxygenation, central venous catheter, indwelling urethral catheter, and nasogastric tube. CONCLUSIONS: Our findings confirmed that the incidence of secondary infection, particularly SBI, and mortality are high among critically ill patients with COVID-19. We showed that long-term hospitalization and invasive procedures such as tracheotomy, central venous catheter, indwelling urethral catheter, and nasogastric tube are risk factors for SBI and other complications.

Prevalence and Characteristics of Ceftriaxone-Resistant Salmonella in Children's Hospital in Hangzhou, China.

The non-Typhi Salmonella (NTS) infection is critical to children's health, and the ceftriaxone is the important empirical treatment choice. With the increase resistance rate of ceftriaxone in Salmonella, the molecular epidemiology and resistance mechanism of ceftriaxone-resistant Salmonella needs to be studied. From July 2019 to July 2020, a total of 205 NTS isolates were collected, 195 of which (95.1%) were cultured from stool, but 10 isolates were isolated from an extraintestinal site. Serogroup B accounted for the vast majority (137/205) among the isolates. Fifty-three isolates were resistant to ceftriaxone, and 50 were isolated from children younger than 4years of age. The resistance rates for ceftriaxone, ciprofloxacin, and levofloxacin were significantly higher in younger children than the older children. The resistance genes in the ceftriaxone-susceptible isolates were detected by PCR, and ceftriaxone-resistant Salmonella were selected for further whole-genome sequencing. Whole-genome analysis showed that serotype Typhimurium and its monophasic variant was the most prevalent in ceftriaxone-resistant isolates (37/53), which comprised ST34 (33/53), ST19 (2/53), and ST99 (2/53), and they were close related in the phylogenetic tree. However, the other isolates were diverse, which included one Enteritidis (ST11), one Indiana (ST17), one Derby (ST40), four Kentucky (ST198), two Goldcoast (ST2529, ST358), one Muenster (ST321), one Virchow (ST359), one Rissen (ST469), one Kedougou (ST1543), two Uganda (ST684), and one Kottbus (ST8839). Moreover, CTX-M-55 ESBLs production (33/53) was found to be mainly responsible for ceftriaxone resistance, followed by bla CTX-M-65 (12/53), bla CTX-M-14 (4/53), bla CTX-M-9 (2/53), bla CTX-M-64 (1/53), bla CTX-M-130 (1/53), and bla CMY-2 (1/53). ISEcp1, IS903B, IS Kpn26, IS1F, and IS26 were connected to antimicrobial resistance genes transfer. In conclusion, the dissemination of ESBL-producing Salmonella isolates resulted in an increased prevalence of ceftriaxone resistance in young children. The high rate of multidrug resistance should be given additional attention.

Wave-packet behaviors of the defocusing nonlinear Schrödinger equation based on the modified physics-informed neural networks.

Investigated in this paper is the defocusing nonlinear Schrödinger (NLS) equation, which is used for describing the wave-packet dynamics in certain weakly nonlinear media. With the physics-informed neural networks (PINNs), we modify the corresponding loss function in the existing literature and obtain two types of dark solitons, type-I and type-II solitons. It is demonstrated that the modified loss function presents higher-precision wave-packet behaviors based on fewer initial and boundary data. Taking type-I solitons into consideration, we find that when only a small fraction of initial and boundary data are given, the prediction accuracy of the wave packets will be increased one or two orders of magnitude at least if the modification term of the loss function is introduced. Furthermore, for the inverse problem, the modified loss function provides a better estimate of the nonlinear coefficient of the NLS equation based on fewer observed data of the wave packets. For type-II solitons, we compare the required data and predicted results of the PINNs with those of the conventional time-splitting finite difference (TSFD) method and reveal that achieving the same precision of the wave-packet behavior, the PINNs with the modified loss functions require only one tenth of the amount of the initial and boundary data of the TSFD method. Besides, both unmodified and modified loss functions are exploited for predicting the behaviors of Gaussian wave packets, and it is observed that the predicted result of the modified loss function agrees with the high-precision solution of the time-splitting Fourier pseudospectral method, whereas the unmodified loss function fails.

Emergence of Ceftazidime/Avibactam and Tigecycline Resistance in Carbapenem-Resistant Klebsiella pneumoniae Due to In-Host Microevolution.

Klebsiella pneumoniae can cause both hospital- and community-acquired clinical infections. Last-line antibiotics against carbapenem-resistant K. pneumoniae (CRKP), such as ceftazidime/avibactam (CZA) and tigecycline (TGC), remain limited as treatment choices. This study aimed to investigate the mechanisms by which CRKP acquires CZA and TGC resistance in vivo under ß-lactam antibiotic and TGC exposure. Three CRKP strains (XDX16, XDX31 and XDX51) were consecutively isolated from an inpatient with a urinary tract infection in two months. PFGE and MLST showed that these strains were closely related and belonged to sequence type (ST) 4496, which is a novel ST closely related to ST11. Compared to XDX16 and XDX31, XDX51 developed CZA and TGC resistance. Sequencing showed that double copies of blaKPC-2 were located on a 108 kb IncFII plasmid, increasing blaKPC-2 expression in XDX51. In addition, ramR was interrupted by Insertion sequence (IS) Kpn14 in XDX51, with this strain exhibiting upregulation of ramA, acrA and acrB expression compared with XDX16 and XDX31. Furthermore, LPS analysis suggested that the O-antigen in XDX51 was defective due to ISKpn26 insertion in the rhamnosyl transferase gene wbbL, which slightly reduced TGC susceptibility. In brief, CZA resistance was caused mainly by blaKPC-2 duplication, and TGC resistance was caused by ramR inactivation with additional LPS changes due to IS element insertion in wbbL. Notably, CRKP developed TGC and CZA resistance within one month under TGC and ß-lactam treatment without exposure to CZA. The CRKP clone ST4496 has the ability to evolve CZA and TGC resistance rapidly, posing a potential threat to inpatients during antibiotic treatment.

SIRVD-DL: A COVID-19 deep learning prediction model based on time-dependent SIRVD.

COVID-19 is one of the biggest challenges that human beings have faced recently. Many researchers have proposed different prediction methods for establishing a virus transmission model and predicting the trend of COVID-19. Among them, the methods based on artificial intelligence are currently the most interesting and widely used. However, only using artificial intelligence methods for prediction cannot capture the time change pattern of the transmission of infectious diseases. To solve this problem, this paper proposes a COVID-19 prediction model based on time-dependent SIRVD by using deep learning. This model combines deep learning technology with the mathematical model of infectious diseases, and forecasts the parameters in the mathematical model of infectious diseases by fusing deep learning models such as LSTM and other time prediction methods. In the current situation of mass vaccination, we analyzed COVID-19 data from January 15, 2021, to May 27, 2021 in seven countries - India, Argentina, Brazil, South Korea, Russia, the United Kingdom, France, Germany, and Italy. The experimental results show that the prediction model not only has a 50% improvement in single-day predictions compared to pure deep learning methods, but also can be adapted to short- and medium-term predictions, which makes the overall prediction more interpretable and robust.

A random forest model based on core genome allelic profiles of MRSA for penicillin plus potassium clavulanate susceptibility prediction.

Treatment failure of methicillin-resistant Staphylococcus aureus (MRSA) infections remains problematic in clinical practice because therapeutic options are limited. Penicillin plus potassium clavulanate combination (PENC) was shown to have potential for treating some MRSA infections. We investigated the susceptibility of MRSA isolates and constructed a drug susceptibility prediction model for the phenotype of the PENC. We determined the minimum inhibitory concentration of PENC for MRSA (n=284) in a teaching hospital (SRRSH-MRSA). PENC susceptibility genotypes were analysed using a published genotyping scheme based on the mecA sequence. mecA expression in MRSA isolates was analysed by qPCR. We established a random forest model for predicting PENC-susceptible phenotypes using core genome allelic profiles from cgMLST analysis. We identified S2-R isolates with susceptible mecA genotypes but PENC-resistant phenotypes; these isolates expressed mecA at higher levels than did S2 MRSA (2.61 vs 0.98, P<0.05), indicating the limitation of using a single factor for predicting drug susceptibility. Using the data of selected UK-sourced MRSA (n=74) and MRSA collected in a previous national survey (NA-MRSA, n=471) as a training set, we built a model with accuracies of 0.94 and 0.93 for SRRSH-MRSA and UK-sourced MRSA (n=287, NAM-MRSA) validation sets. The AUROC of this model for SRRSH-MRSA and NAM-MRSA was 0.96 and 0.97. Although the source of the training set data affects the scope of application of the prediction model, our data demonstrated the power of the machine learning approach in predicting susceptibility from cgMLST results.