Current experimental and computational methods have limitations in accurately and efficiently classi-fying ion channels within vast protein spaces.Here we have developed a deep learning algorithm,GPT2 Ion Channel Classifier(GPT2-ICC),which effectively distinguishing ion channels from a test set con-taining approximately 239 times more non-ion-channel proteins.GPT2-ICC integrates representation learning with a large language model(LLM)-based classifier,enabling highly accurate identification of potential ion channels.Several potential ion channels were predicated from the unannotated human proteome,further demonstrating GPT2-ICC's generalization ability.This study marks a significant advancement in artificial-intelligence-driven ion channel research,highlighting the adaptability and effectiveness of combining representation learning with LLMs to address the challenges of imbalanced protein sequence data.Moreover,it provides a valuable computational tool for uncovering previously uncharacterized ion channels.

Objective To investigate the relationships of serum amyloid A(SAA),eotaxin and pentraxin 3(PTX3)levels with Mycoplasma pneumoniae DNA(MP-DNA)level,lung function and prognosis in children with Mycoplasma pneumoniae pneumonia(MPP).Methods A total of 163 chil-dren with MPP were enrolled in MPP group and divided into good prognosis group(n=109)and poor prognosis group(n=54)based on prognosis;another 124 healthy children with physical examinations in the same period were included as control group.Levels of serum SAA,eotaxin and PTX3 were measured in the children.Multivariable Logistic regression analysis was conducted to explore risk fac-tors for poor prognosis in children with MPP.Results Compared with the control group,the MPP group had significantly increased levels of SAA,eotaxin,PTX3 and MP-DNA positivity(P＜0.01);the MPP group also had significantly lower values of peak expiratory flow(PEF),forced expiratory volume in the first second(FEV1),forced vital capacity(FVC),tidal expiratory flow time to total expiratory time(tPTEF/TE),and tidal volume compared with the control group(P＜0.01).Compared with the good prognosis group,the poor prognosis group had significantly increased MP-DNA level,SAA,eotaxin and PTX3 levels,as well as decreased values of PEF,FEV1,FVC,tPTEF/TE and tidal volume(P＜0.01).Multivariable Logistic regression analysis showed that high MP-DNA lev-el,high SAA,eotaxin and PTX3 levels,low values of PEF,FEV1,FVC,tPTEF/TE and tidal vol-ume were risk factors for poor prognosis in children with MPP(P＜0.01).The SAA,PTX3 and eotaxin levels in children with MPP were positively correlated with their MP-DNA level and negative-ly correlated with lung function indicators(P＜0.05).Conclusion SAA,PTX3 and eotaxin lev-els are positively correlated with MP-DNA level and negatively correlated with lung function indica-tors in children with MPP.SAA,PTX3 and eotaxin levels can be used to assess the prognosis of children with MPP.

Current experimental and computational methods have limitations in accurately and efficiently classifying ion channels within vast protein spaces. Here we have developed a deep learning algorithm, GPT2 Ion Channel Classifier (GPT2-ICC), which effectively distinguishing ion channels from a test set containing approximately 239 times more non-ion-channel proteins. GPT2-ICC integrates representation learning with a large language model (LLM)-based classifier, enabling highly accurate identification of potential ion channels. Several potential ion channels were predicated from the unannotated human proteome, further demonstrating GPT2-ICC's generalization ability. This study marks a significant advancement in artificial-intelligence-driven ion channel research, highlighting the adaptability and effectiveness of combining representation learning with LLMs to address the challenges of imbalanced protein sequence data. Moreover, it provides a valuable computational tool for uncovering previously uncharacterized ion channels.

Objective:To analyze the real-world practices of resecting sessile colorectal polyps of varying long diameters using cold forcep polypectomy (CFP), cold snare polypectomy (CSP), or endoscopic mucosal resection (EMR).Methods:A total of 12 290 nonpedunculated colorectal polyps of long diameter ≤19 mm (from 10 295 patients) were retrospectively enrolled from January 2022 to December 2023. Polypectomy was conducted by 30 endoscopists. The polyps were categorized into three groups based on long diameter: 1-5 mm, >5-10 mm and >10-19 mm, and the differences of polypectomy methods were compared in three groups. The usage of hemostatic clips in CSP among >5-10 mm polyps and the changes in resection methods between 2022 and 2023 were analyzed.Results:CFP (6 769 polyps, 81.7%) was the predominant method for resecting 1-5 mm sessile polyps (8 289 polyps). For sessile polyps sized >5-10 mm (2 455 polyps), CSP was used most (1 372, 55.9%), although its utilization varied significantly among physicians with the median usage rate of 52.9% (40.3%, 60.0%). EMR (1 349 poolyps, 87.3%) was the main method for >10-19 mm sessile polyps. The usage rate of CSP in sessile polypectomy for polyps >5-10 mm significantly increased from 45.7% (503/1 101) in 2022 to 64.2% (869/1 354) in 2023. The overall frequency of using clip in CSP for >5-10 mm sessile polyps was 40.1% (550/1 372), demonstrating notable variability among different endoscopists with median usage rate of 48.3% (29.8%, 67.9%).Conclusion:Varied resection methods are observed among endoscopists for sessile polyps measuring ≤19 mm. CFP is primarily utilized for polyps of 1-5 mm, while CSP is favored for polyps >5-10 mm, with an increasing annual usage rate. EMR is the main approach for the polyps >10-19 mm. Additionally, notable variations in the use of metal clips during CSP are observed among different physicians.

OBJECTIVE To analyze the distribution and drug resistance of pathogenic bacteria in hospital-associat-ed infections for patients after renal transplantation,and to provide a reference for the rational selection of antibac-terial drugs for anti-infective treatment in such patients.METHODS Clinical data were collected from 89 hospital-ized patients who underwent renal transplantation and developed hospital-associated infection at Beidaihe Rehabili-tation and Recuperation Center(formerly known as 281 Hospital)from 2017 to 2021.The pathogenic bacteria de-tected in different types of hospital-associated infections and the drug resistance of the main pathogenic bacteria were analyzed.RESULTS A total of 89 pathogenic bacterial strains were isolated from the 89 patients with hospi-tal-associated infection,including 74 gram-negative bacterial strains(83.15％)and 15 gram-positive bacterial strains(16.85％).The predominant pathogenic bacterial strains were Klebsiella pneumoniae,Escherichia coli and Pseudomonas aeruginosa.Respiratory tract infection was the most common.E.coli showed severe drug re-sistance,with high drug resistance rates to various antibacterial drugs(complete drug resistance to piperacillin and ampicillin).K.pneumoniae was generally drug resistant to ampicillin.Enterobacter cloacae showed high sensitivi-ty to carbapenems,amikacin and enzyme inhibitor combinations.P.aeruginosa had low drug resistance rates to imipenem and meropenem.Staphylococcus aureus was completely drug resistant to antibacterial drugs such as penicillin,erythromycin and gentamicin,and no drug-resistant strains to vancomycin and linezolid were found.CONCLUSIONS The gram-negative bacteria are dominant among the bacteria isolated from the renal trans-planted patients with postoperative hospital-associated infections,with a high detection rate of E.coli and a severe drug resistance situation.Therefore,it is necessary to strengthen the monitoring of drug resistance in renal trans-plant recipients and make rational choices of antibacterial drugs.

OBJECTIVE To assess the current status of hospital-acquired infections and their economic burden in or-thopedic patients based on diagnosis-related groups(DRG).METHOD Based on the National Health Insurance dis-ease diagnosis-related groups,32 413 orthopedic patients from a tertiary care hospital in Beijing in 2021 were grouped,hospital-acquired infections were retrospectively analyzed,and the direct and indirect economic burdens of different DRG groups were assess using indictors such as hospitalization time and cost,bed turnover loss,and labor time loss.RESULTS A total of 32 413 patients were included,the incidence of hospital-acquired infection was 0.47％(153/32 413),the site of infection was predominantly the surgical site(57.99％),and hospital-acquired infections in the hematologic system had a greater impact on cost-consumption indices and time-consumption indi-ces.The infection cases were concentrated in 19.58％of the DRGs groups.The IF23 group(lower limb bone sur-gery with complications and comorbidities)had the highest direct economic burden(24 010 yuan/case)due to hos-pital-acquired infections,and the increase in the cost of consumables and medication was the main factor causing the direct economic burden.At both the hospital level and family-society level,the top three DRG groups in terms of indirect economic burden due to hospital-acquired infections were IB15,IB13 and IF23.CONCLUSION Hospital-acquired infections in orthopedic patients have a tendency to be concentrated,quantitatively assessment of their e-conomic burden based on DRGs not only illustrates the importance of hospital-acquired infection prevention and control,but also accurately identifies the disease groups that require focused management,providing an evidence-based basis for precise prevention and control of hospital-acquired infections.

OBJECTIVE To analyze the distribution and drug resistance of pathogenic bacteria in hospital-associat-ed infections for patients after renal transplantation,and to provide a reference for the rational selection of antibac-terial drugs for anti-infective treatment in such patients.METHODS Clinical data were collected from 89 hospital-ized patients who underwent renal transplantation and developed hospital-associated infection at Beidaihe Rehabili-tation and Recuperation Center(formerly known as 281 Hospital)from 2017 to 2021.The pathogenic bacteria de-tected in different types of hospital-associated infections and the drug resistance of the main pathogenic bacteria were analyzed.RESULTS A total of 89 pathogenic bacterial strains were isolated from the 89 patients with hospi-tal-associated infection,including 74 gram-negative bacterial strains(83.15％)and 15 gram-positive bacterial strains(16.85％).The predominant pathogenic bacterial strains were Klebsiella pneumoniae,Escherichia coli and Pseudomonas aeruginosa.Respiratory tract infection was the most common.E.coli showed severe drug re-sistance,with high drug resistance rates to various antibacterial drugs(complete drug resistance to piperacillin and ampicillin).K.pneumoniae was generally drug resistant to ampicillin.Enterobacter cloacae showed high sensitivi-ty to carbapenems,amikacin and enzyme inhibitor combinations.P.aeruginosa had low drug resistance rates to imipenem and meropenem.Staphylococcus aureus was completely drug resistant to antibacterial drugs such as penicillin,erythromycin and gentamicin,and no drug-resistant strains to vancomycin and linezolid were found.CONCLUSIONS The gram-negative bacteria are dominant among the bacteria isolated from the renal trans-planted patients with postoperative hospital-associated infections,with a high detection rate of E.coli and a severe drug resistance situation.Therefore,it is necessary to strengthen the monitoring of drug resistance in renal trans-plant recipients and make rational choices of antibacterial drugs.

OBJECTIVE To assess the current status of hospital-acquired infections and their economic burden in or-thopedic patients based on diagnosis-related groups(DRG).METHOD Based on the National Health Insurance dis-ease diagnosis-related groups,32 413 orthopedic patients from a tertiary care hospital in Beijing in 2021 were grouped,hospital-acquired infections were retrospectively analyzed,and the direct and indirect economic burdens of different DRG groups were assess using indictors such as hospitalization time and cost,bed turnover loss,and labor time loss.RESULTS A total of 32 413 patients were included,the incidence of hospital-acquired infection was 0.47％(153/32 413),the site of infection was predominantly the surgical site(57.99％),and hospital-acquired infections in the hematologic system had a greater impact on cost-consumption indices and time-consumption indi-ces.The infection cases were concentrated in 19.58％of the DRGs groups.The IF23 group(lower limb bone sur-gery with complications and comorbidities)had the highest direct economic burden(24 010 yuan/case)due to hos-pital-acquired infections,and the increase in the cost of consumables and medication was the main factor causing the direct economic burden.At both the hospital level and family-society level,the top three DRG groups in terms of indirect economic burden due to hospital-acquired infections were IB15,IB13 and IF23.CONCLUSION Hospital-acquired infections in orthopedic patients have a tendency to be concentrated,quantitatively assessment of their e-conomic burden based on DRGs not only illustrates the importance of hospital-acquired infection prevention and control,but also accurately identifies the disease groups that require focused management,providing an evidence-based basis for precise prevention and control of hospital-acquired infections.

Objective:To analyze the real-world practices of resecting sessile colorectal polyps of varying long diameters using cold forcep polypectomy (CFP), cold snare polypectomy (CSP), or endoscopic mucosal resection (EMR).Methods:A total of 12 290 nonpedunculated colorectal polyps of long diameter ≤19 mm (from 10 295 patients) were retrospectively enrolled from January 2022 to December 2023. Polypectomy was conducted by 30 endoscopists. The polyps were categorized into three groups based on long diameter: 1-5 mm, >5-10 mm and >10-19 mm, and the differences of polypectomy methods were compared in three groups. The usage of hemostatic clips in CSP among >5-10 mm polyps and the changes in resection methods between 2022 and 2023 were analyzed.Results:CFP (6 769 polyps, 81.7%) was the predominant method for resecting 1-5 mm sessile polyps (8 289 polyps). For sessile polyps sized >5-10 mm (2 455 polyps), CSP was used most (1 372, 55.9%), although its utilization varied significantly among physicians with the median usage rate of 52.9% (40.3%, 60.0%). EMR (1 349 poolyps, 87.3%) was the main method for >10-19 mm sessile polyps. The usage rate of CSP in sessile polypectomy for polyps >5-10 mm significantly increased from 45.7% (503/1 101) in 2022 to 64.2% (869/1 354) in 2023. The overall frequency of using clip in CSP for >5-10 mm sessile polyps was 40.1% (550/1 372), demonstrating notable variability among different endoscopists with median usage rate of 48.3% (29.8%, 67.9%).Conclusion:Varied resection methods are observed among endoscopists for sessile polyps measuring ≤19 mm. CFP is primarily utilized for polyps of 1-5 mm, while CSP is favored for polyps >5-10 mm, with an increasing annual usage rate. EMR is the main approach for the polyps >10-19 mm. Additionally, notable variations in the use of metal clips during CSP are observed among different physicians.

Objective:To provide references for the prevention and treatment of hyperuricemia (HUA) in aircrews by investigating the prevalence of HUA in aircrews and analyzing the correlation between HUA and other biochemical indicators.Methods:The physical examination data of male aircrews who participated in physical examination in Beidaihe Rehabilitation and Recuperation Center of PLA from 2020 to 2022 were retrospectively analyzed. According to the year, they were divided into 2020, 2021 and 2022 groups. According to the age, they were divided into 20-29, 30-39, 40-49 and 50-59 years groups. The prevalence of HUA among different years and age groups were compared and analyzed. The subjects were divided into HUA group and non-HUA group according to whether they had HUA. The differences of physical examination indexes were compared between 2 groups, and the potential effects of other biochemical indexes on HUA were analyzed by multivariate Logistic regression.Results:A total of 2 365 aircrews were enrolled. Eight hundred and sixty-six aircrews were diagnosed as HUA, accounting for 36.62%. The prevalence rate increased year by year, and the difference was statistically significant ( χ2=6.70, P=0.010). There was no significant difference in prevalence rate among different age groups ( P>0.05). There were significant differences in creatinine ( t=9.40, P<0.001), serum uric acid ( t=59.66, P<0.001), urea nitrogen ( t=4.60, P<0.001), direct bilirubin ( Z=3.40, P=0.001), alanine aminotransferase ( Z=7.59, P<0.001), low density lipoprotein cholesterol ( t=5.89, P<0.001), triglyceride ( Z=8.02, P<0.001) and total cholesterol ( t=6.79, P<0.001) between HUA group and non-HUA group. Multivariate Logistic regression analysis showed that increased alanine aminotransferase ( OR=1.014, P<0.001), urea nitrogen ( OR=1.091, P<0.001), creatinine ( OR=1.038, P<0.001), and triglyceride ( OR=1.164, P=0.001) level were the risk factors for HUA. Conclusions:The prevalence of HUA in aircrew is increasing year by year. Alanine aminotransferase, urea nitrogen, creatinine and triglyceride are the risk factors for HUA. It is suggested that attention should be paid to the increasing trend of HUA in aircrews year by year, to the impact of abnormal liver and kidney function on metabolism and blood lipid levels, and early intervention measures should be taken for related risk factors.