Assessing the Risk of Bias in Randomized Clinical Trials With Large Language Models.

Importance: Large language models (LLMs) may facilitate the labor-intensive process of systematic reviews. However, the exact methods and reliability remain uncertain. Objective: To explore the feasibility and reliability of using LLMs to assess risk of bias (ROB) in randomized clinical trials (RCTs). Design, Setting, and Participants: A survey study was conducted between August 10, 2023, and October 30, 2023. Thirty RCTs were selected from published systematic reviews. Main Outcomes and Measures: A structured prompt was developed to guide ChatGPT (LLM 1) and Claude (LLM 2) in assessing the ROB in these RCTs using a modified version of the Cochrane ROB tool developed by the CLARITY group at McMaster University. Each RCT was assessed twice by both models, and the results were documented. The results were compared with an assessment by 3 experts, which was considered a criterion standard. Correct assessment rates, sensitivity, specificity, and F1 scores were calculated to reflect accuracy, both overall and for each domain of the Cochrane ROB tool; consistent assessment rates and Cohen κ were calculated to gauge consistency; and assessment time was calculated to measure efficiency. Performance between the 2 models was compared using risk differences. Results: Both models demonstrated high correct assessment rates. LLM 1 reached a mean correct assessment rate of 84.5% (95% CI, 81.5%-87.3%), and LLM 2 reached a significantly higher rate of 89.5% (95% CI, 87.0%-91.8%). The risk difference between the 2 models was 0.05 (95% CI, 0.01-0.09). In most domains, domain-specific correct rates were around 80% to 90%; however, sensitivity below 0.80 was observed in domains 1 (random sequence generation), 2 (allocation concealment), and 6 (other concerns). Domains 4 (missing outcome data), 5 (selective outcome reporting), and 6 had F1 scores below 0.50. The consistent rates between the 2 assessments were 84.0% for LLM 1 and 87.3% for LLM 2. LLM 1's κ exceeded 0.80 in 7 and LLM 2's in 8 domains. The mean (SD) time needed for assessment was 77 (16) seconds for LLM 1 and 53 (12) seconds for LLM 2. Conclusions: In this survey study of applying LLMs for ROB assessment, LLM 1 and LLM 2 demonstrated substantial accuracy and consistency in evaluating RCTs, suggesting their potential as supportive tools in systematic review processes.

ABBV-319: A CD19-targeting glucocorticoid receptor modulator antibody-drug conjugate therapy for B-cell malignancies.

Glucocorticoids are key components of the current standard-of-care regimens (e.g., R-CHOP, EPOCH-R, Hyper-CVAD) for treatment of B-cell malignancy. However, systemic glucocorticoid treatment is associated with several adverse events. CD19 displays restricted expression in normal B-cells and is up-regulated in B-cell malignancies. ABBV-319 is a CD19-targeting antibody-drug conjugate (ADC) engineered to reduce glucocorticoid-associated toxicities while possessing three distinct mechanisms of action (MOA) to increase therapeutic efficacy: (1) antibody-mediated delivery of glucocorticoid receptor modulator (GRM) payload to activate apoptosis, (2) inhibition of CD19 signaling, and (3) enhanced Fc-mediated effector function via afucosylation of the antibody backbone. ABBV-319 elicited potent GRM-driven anti-tumor activity against multiple malignant B-cell lines in vitro as well as in cell line-derived xenografts (CDXs) and patient-derived xenografts (PDXs) in vivo. Remarkably, a single-dose of ABBV-319 induced sustained tumor regression and enhanced anti-tumor activity compared to repeat dosing of systemic prednisolone at the maximum tolerated dose (MTD) in mice. The unconjugated CD19 monoclonal antibody (mAb) also displayed anti-proliferative activity on a subset of B-cell lymphoma cell lines through the inhibition of PI3K signaling. Moreover, afucosylation of the CD19 mAb enhanced Fc-mediated antibody-dependent cellular cytotoxicity (ADCC), and this activity was maintained after conjugation with GRM payloads. Notably, ABBV-319 displayed superior efficacy compared to afucosylated CD19 mAb in human CD34+ PBMC-engrafted NSG-tg(Hu-IL15) transgenic mice, demonstrating enhanced anti-tumor activity when multiple MOAs are enabled. ABBV-319 also showed durable anti-tumor activity across multiple B-cell lymphoma PDX models, including non-germinal center B-cell (GCB) DLBCL and relapsed lymphoma post R-CHOP treatment. Collectively, these data support the ongoing evaluation of ABBV-319 in Phase I clinical trial (NCT05512390).

Internal multiple interactions-adsorption and external zwitterionic polymer-exclusion of restricted access materials as adsorbent for offline and online extraction of neonicotinoid pesticides in Goji samples.

A method based on novel restricted access materials (RAMs) for the determination of neonicotinoid pesticides in Goji samples using offline and online solid phase extraction (SPE) coupled with high-performance liquid chromatography (LC). RAMs were synthesized using poly(chloromethylstyrene-co-divinylbenzene) (PVBC/DVB) microspheres as substrate, styrene (St) and n-vinylpyrrolidone (NVP) were first copolymerized on the interior to construct adsorption sites, and sulfobetaine methacrylate (SBMA) was then polymerized on the exterior to form exclusion sites via two-step surface initiated-atom transfer polymerization. The prepared PVBC/DVB@poly(St-co-NVP)@poly(SBMA) RAMs could efficiently extract neonicotinoid pesticides and automatically exclude proteins. Under the optimized conditions, the developed methods of offline (magnetic SPE and SPE column) and online extraction coupled with LC both using PVBC/DVB@poly(St-co-NVP)@poly(SBMA) RAMs as the extractant, exhibit a wide linearity, low limits of detection and limit of quantification and good inter-day and intra-day precision with satisfactory recoveries. Among these methods, online extraction coupled with LC based on novel RAMs exhibits clear advantages for the determination of neonicotinoid pesticides in Goji samples has clear advantages, such as simple operation by direct injection, short extraction times, and high accuracy with less human error.

Risk of kidney and liver diseases after COVID-19 infection: A systematic review and meta-analysis.

COVID-19 is not only associated with substantial acute liver and kidney injuries, but also with an elevated risk of post-acute sequelae involving the kidney and liver system. We aimed to investigate whether COVID-19 exposure increases the long-term risk of kidney and liver disease, and what are the magnitudes of these associations. We searched PubMed, Embase, Web of Science, ClinicalTrials.gov, and the Living Overview of the Evidence COVID-19 Repository for cohort studies estimating the association between COVID-19 and kidney and liver outcomes. Random-effects meta-analyses were performed to combine the results of the included studies. We assessed the certainty of the evidence using the Grading of Recommendations Assessment, Development and Evaluation approach. Fifteen cohort studies with more than 32 million participants were included in the systematic review COVID-19 was associated with a 35% greater risk of kidney diseases (10 more per 1000 persons; low certainty evidence) and 54% greater risk of liver disease (3 more per 1000 persons; low certainty evidence). The absolute increases due to COVID-19 for acute kidney injury, chronic kidney disease, and liver test abnormality were 3, 8, and 3 per 1000 persons, respectively. Subgroup analyses found no differences between different type of kidney and liver diseases. The findings provide further evidence for the association between COVID-19 and incident kidney and liver conditions. The absolute magnitude of the effect of COVID-19 on kidney and liver outcomes was, however, relatively small.

Preparation and chromatographic evaluation of a mixed polymer brush-silica stationary phase with temperature-sensitive property.

In this study, a developed chromatographic stationary phase combines the high selectivity of mixed-mode retention with a temperature-responsive property to boost separation efficiency. Copolymer brushes were grafted onto silica gels through surface initiated-atom transfer radical polymerization by polymerizing two types of monomer, temperature-responsive vinylcaprolactam (VCl) and quinine (Qun) containing benzopyridine, a tertiary ammonium positive center, and hydroxyl groups. The obtained silica@poly(Qun-co-VCl) stationary phases were packed as a chromatographic column, and the retention behavior of hydrophobic polycyclic aromatics, highly polar nucleosides, charged organic acids and ß-agonists was studied for this column under different separation modes. The ability to separate different types of analyte shows that the silica@poly(Qun-co-VCl) column provides multiple hydrophobic, hydrophilic and electrostatic interactions toward analytes, achieving the separation of various compounds in one column. In addition, temperature-dependent resolution of polycyclic aromatics, nucleosides, organic acids and ß-agonists was investigated using modulation of the column temperature, and the column exhibited adjustable separation selectivity by simply changing the column temperature. These results demonstrate that the grafting of copolymer brushes on a silica surface, consisting of temperature-responsive poly-VCl and multifunctional groups of poly-Qun, is useful as a mixed-mode chromatographic stationary phase for thermally-modulated multiple interactions. Additionally, this column was also used for the quantitative detection of uridine and inosine from cordyceps.

Changes in Heart Rate after Pulmonary Vein Isolation in Patients with Paroxysmal Atrial Fibrillation and Sinus Bradycardia.

This study aimed to evaluate the effect of circumferential pulmonary vein isolation (CPVI) on autonomic nervous function and prognosis in patients with paroxysmal atrial fibrillation (AF) with or without sinus bradycardia.A total of 66 patients with paroxysmal AF accompanied by sinus bradycardia and who underwent CPVI were recruited as the sinus bradycardia group. A total of 91 patients with paroxysmal AF but without sinus bradycardia and who underwent catheter ablation were selected as the control group. After surgical contraindications were eliminated, CPVI was performed by three-dimensional mapping system. 24-hour dynamic electrocardiogram was used to observe the changes of heart rate before and 2 days after surgery.A total of 45 (68%) and 51 (56%) patients in the sinus bradycardia and control groups, respectively, maintained sinus rhythm. There was an increase in heart rate after CPVI in both groups. The standard deviation of normal-to-normal (NN) intervals (SDNN), standard deviation of the average NN intervals (SDANN), low frequency (LF), and LF/high frequency (HF) in the sinus bradycardia and control groups decreased after CPVI (P < 0.01). Moreover, SDANN was higher in patients with sinus bradycardia treated by successful ablation than in those with recurrence (P < 0.01), while SDNN, a standard statistical measure of heart rate variability (rMSSD), LF, and HF were significantly lower in patients with sinus bradycardia (P < 0.05).CPVI was able to produce a significant reduction effect on vagal nerve and sympathetic activity regardless of whether patients with paroxysmal AF had sinus bradycardia. Moreover, CPVI exerted a certain influence on the success rate of AF catheter ablation.

Engineering Oncolytic Coxsackievirus A21 with Small Transgenes and Enabling Cell-Mediated Virus Delivery by Integrating Viral cDNA into the Genome.

Coxsackievirus A21 (CVA21) is a naturally occurring RNA virus that, in preclinical studies and clinical trials, has demonstrated promising potential in treating a range of malignancies. Other oncolytic viruses, such as adenovirus, vesicular stomatitis virus, herpesvirus, and vaccinia virus, all can be engineered to carry one or more transgenes for various purposes, including immune modulation, virus attenuation, and induction of apoptosis of tumor cells. However, it remained unknown whether CVA21 can express therapeutic or immunomodulatory payloads due to its small size and high mutation rate. Using reverse genetics techniques, we demonstrated that a transgene encoding a truncated green fluorescent protein (GFP) of up to 141 amino acids (aa) can be successfully carried in the 5' end of the coding region. Furthermore, a chimeric virus carrying an eel fluorescent protein, UnaG (139 aa), was also made and shown to be stable, and it maintained efficient tumor cell-killing activity. Similar to other oncolytic viruses, the likelihood of delivering CVA21 by the intravenous route is low due to issues like blood absorption, neutralizing antibodies, and liver clearance. To address this problem, we designed the CVA21 cDNA under the control of a weak RNA polymerase II promoter, and subsequently, a stable cell pool in 293T cells was made by integrating the resulting CVA21 cDNA into the cell genome. We showed that the cells are viable and able to persistently generate rCVA21 de novo. The carrier cell approach described here may pave the way to designing new cell therapy strategies by arming with oncolytic viruses. IMPORTANCE As a naturally occurring virus, coxsackievirus A21 is a promising oncolytic virotherapy modality. In this study, we first used reverse genetics to determine whether A21 can stably carry transgenes and found that it could express up to 141 amino acids of foreign GFP. The chimeric virus carrying another fluorescent eel protein UnaG (139 amino acids) gene also appeared to be stable over at least 7 passages. Our results provided guidance on how to select and engineer therapeutic payloads for future A21 anticancer research. Second, the challenges of delivering oncolytic viruses by the intravenous route hamper the broader use of oncolytic viruses in the clinic. Here, we used A21 to show that cells could be engineered to stably carry and persistently release the virus by harboring the viral cDNA in the genome. The approach we presented here may pave a new way for oncolytic virus administration using cells as carriers.

Comparison of the levels of depression and anxiety in elderly aortic stenosis patients treated with surgical or transcatheter aortic valve replacement.

OBJECTIVE: Currently, only a few studies have been conducted on the mental status recovery in elderly aortic stenosis (AS) patients after treatment. How transcatheter aortic valve replacement (TAVR) and surgical aortic valve replacement (SAVR) differentially impinge on the mental status of elderly AS patients is completely unknown. The present prospective study aims to investigate this question by comparing the post-treatment levels of depression and anxiety, quality of life and frailty. METHODS: A total of 120 elderly patients (age above 70) with symptomatic AS were included, where 78 of them were treated with TAVR and 42 of them were treated with SAVR. Levels of depression and anxiety, quality of life and frailty were assessed by the Chinese version of Hospital Anxiety and Depression Scale (HADS), World Health Organization Quality of Life Instrument-Older Adults Module (WHOQOL-OLD) and clinical frailty scale, respectively. Scores were recorded and compared at admission, 1 month, 4 months and 8 months after treatment. RESULTS: Before treatment, both patient groups had similar baseline characteristics and all mental parameters. During the follow-up period, patients in the TAVR group demonstrated significant improvement in all assessed mental parameters to certain extent compared to the SAVR group. Specifically, frailty was significantly improved in the TAVR-treated patients at all three follow-up time points. Levels of depression and anxiety were significantly improved 8 months after treatment, although the remaining patient number is limited. Quality of life was only significantly improved 1 month after treatment. CONCLUSION: TAVR may provide a better mental recovery outcome in elderly AS patients.

Experimental study of methane hydrate generation characteristics in the presence of GO and Re-GO.

The industrial application of hydrate technology is greatly hindered by its slow generation rate, low gas storage rate, harsh generation conditions, and poor environmental friendliness of traditional additives. In this paper, the kinetic and thermodynamic promotion effects of graphene oxide (GO) and recovered graphene oxide (Re-GO) on methane hydrate in different systems were studied by the constant volume methods. The promotion mechanism was analyzed from the micro perspectives of molecular physical properties, interfacial reaction, and nucleation sites. It is found that GO has an excellent kinetic and thermodynamic promotion effect on CH4 hydrate generation. After the recovery process, the thermodynamic effect of Re-GO is basically unchanged, and the kinetic promotion effect is slightly reduced. Furthermore, it is verified that the GO material itself does not have a memory effect in hydrate formation. The results show that GO is an excellent accelerator of CH4 hydrate formation with high recovery value, which provides essential data and an experimental basis for the research and application of graphene oxide and hydrate technology in energy storage and cold storage.

Left Bundle Branch Area Pacing vs. Biventricular Pacing for Cardiac Resynchronization Therapy: A Meta-Analysis.

Background: Left bundle branch area pacing (LBBAP) is a recently proposed method for conduction system pacing. We performed a meta-analysis of controlled studies to compare the clinical outcome in patients who received LBBAP vs. biventricular pacing (BVP) for cardiac resynchronization therapy (CRT). Methods: PubMed, Embase, and Cochrane's Library databases were searched for relevant controlled studies. A random-effect model incorporating the potential heterogeneity was used to synthesize the results. Results: Four non-randomized controlled studies including 249 patients with heart failure (HF) for CRT were included, and the patients were followed for 6-12 months. Compared with BVP, LBBAP was associated with significantly shortened QRS duration [mean difference (MD): -29.18 ms, 95% confidence interval (CI): -33.55-24.80, I 2 = 0%, P < 0.001], improved left ventricular ejection fraction (MD: 6.93%, 95% CI: 4.69-9.17, I 2 = 0%, P < 0.001), reduced left ventricular end-diastolic dimension (MD: -2.96 mm, 95% CI: -5.48 to -0.44, I 2 = 0%, P = 0.02), and improved New York Heart Association class (MD: -0.54, 95% CI: -0.84 to -0.24, I 2 = 65%, P < 0.001). Moreover, patients who received LBBAP were more likely to achieve echocardiographic [odds ratio (OR): 5.04, 95% CI: 2.17-11.69, I 2 = 0%, P < 0.001] and clinical (OR: 7.33, 95% CI: 1.62-33.16, I 2 = 0%, P = 0.01) CRT responses. Conclusion: Current evidence from non-randomized studies suggests that LBBAP appears to be a promising method for CRT, which is associated with more remarkable improvements of symptoms and cardiac function in HF patients with indication for CRT.

Oligomerization of the FliF Domains Suggests a Coordinated Assembly of the Bacterial Flagellum MS Ring.

The bacterial flagellum is a complex, self-assembling macromolecular machine that powers bacterial motility. It plays diverse roles in bacterial virulence, including aiding in colonization and dissemination during infection. The flagellum consists of a filamentous structure protruding from the cell, and of the basal body, a large assembly that spans the cell envelope. The basal body is comprised of over 20 different proteins forming several concentric ring structures, termed the M- S- L- P- and C-rings, respectively. In particular, the MS rings are formed by a single protein FliF, which consists of two trans-membrane helices anchoring it to the inner membrane and surrounding a large periplasmic domain. Assembly of the MS ring, through oligomerization of FliF, is one of the first steps of basal body assembly. Previous computational analysis had shown that the periplasmic region of FliF consists of three structurally similar domains, termed Ring-Building Motif (RBM)1, RBM2, and RBM3. The structure of the MS-ring has been reported recently, and unexpectedly shown that these three domains adopt different symmetries, with RBM3 having a 34-mer stoichiometry, while RBM2 adopts two distinct positions in the complex, including a 23-mer ring. This observation raises some important question on the assembly of the MS ring, and the formation of this symmetry mismatch within a single protein. In this study, we analyze the oligomerization of the individual RBM domains in isolation, in the Salmonella enterica serovar Typhimurium FliF ortholog. We demonstrate that the periplasmic domain of FliF assembles into the MS ring, in the absence of the trans-membrane helices. We also report that the RBM2 and RBM3 domains oligomerize into ring structures, but not RBM1. Intriguingly, we observe that a construct encompassing RBM1 and RBM2 is monomeric, suggesting that RBM1 interacts with RBM2, and inhibits its oligomerization. However, this inhibition is lifted by the addition of RBM3. Collectively, this data suggest a mechanism for the controlled assembly of the MS ring.

Effect of the Inclination Angle on Premixed Flame Dynamics in Half-Open Ducts.

The propagation of premix methane/air flames in long half-open ducts with different inclination angles θ between the sidewall and the horizon was numerically investigated using the laminar model. The numerical result was compared with the experimental and theoretical ones to validate the numerical model. The results show that the numerical results are in good agreement with them. The investigation provides the basic understanding of the effects on changing the shape of the ducts to promote the premixed flame combustion. For methane/air, the position where the flame front begins to concave is pushed back with the increase in angle θ. The so-called "tulip" flame even disappeared, if the angle θ is bigger than one certain value. Moreover, the flame propagation speed and pressure are enhanced as the angle θ increases. In addition, the numerical simulation indicates that the burning gas creates an eddy near the tip of the flame, altering the flow field and causing the tulip flame to appear. However, with the angle θ increased, the flame propagation is restrained by the change in sidewalls, resulting in the different flow patterns to suppress the formation of tulip flames and promote flame combustion.

Modeling and analyzing neural signals with phase variability using Fisher-Rao registration.

BACKGROUND: The dynamic time warping (DTW) has recently been introduced to analyze neural signals such as EEG and fMRI where phase variability plays an important role in the data. NEW METHOD: In this study, we propose to adopt a more powerful method, referred to as the Fisher-Rao Registration (FRR), to study the phase variability. COMPARISON WITH EXISTING METHODS: We systematically compare FRR with DTW in three aspects: (1) basic framework, (2) mathematical properties, and (3) computational efficiency. RESULTS: We show that FRR has superior performance in all these aspects and the advantages are well illustrated with simulation examples. CONCLUSIONS: We then apply the FRR method to two real experimental recordings - one fMRI and one EEG data set. It is found the FRR method properly removes the phase variability in each set. Finally, we use the FRR framework to examine brain networks in these two data sets and the result demonstrates the effectiveness of the new method.

Structure-Activity Relationships of Hydroxyapatite-Binding Peptides.

Elucidating the structure-activity relationships between biomolecules and hydroxyapatite (HAP) is essential to understand bone mineralization mechanisms, develop HAP-based implants, and design drug delivery vectors. Here, four peptides identified by phage display were selected as model HAP-binding peptides (HBPs) to examine the effects of primary amino acid sequence, phosphorylation of serine, presence of charged amino acid residues, and net charge of the peptide on (1) HAP-binding affinity, (2) secondary conformation, and (3) HAP nucleation and crystal growth. Binding affinities were determined by obtaining adsorption isotherms by mass depletion, and the conformations of the peptides in solution and bound states were observed by circular dichroism. Results showed that the magnitude of the net charge primarily controlled binding affinity, with little dependence on the other HBP features. The binding affinity and conformation results were in good agreement with our previous molecular dynamics simulation results, thus providing an excellent benchmark for the simulations. Transmission electron microscopy was used to explore the effect of these HBPs on calcium phosphate (Ca-PO4) nucleation and growth. Results indicated that HBPs may inhibit nucleation of Ca-PO4 nanoparticles and their phase transition to crystalline HAP, as well as control crystal growth rates in specific crystallographic directions, thus changing the classical needle-like morphology of inorganically grown HAP crystals to a biomimetic plate-like morphology.

Toward the Understanding of Small Protein-Mediated Collagen Intrafibrillar Mineralization.

The design of improved materials for orthopedic implants and bone tissue engineering scaffolds relies on materials mimicking the properties of bone. Calcium phosphate (Ca-PO4)-mineralized collagen fibrils arranged in a characteristic hierarchical structure constitute the building blocks of mineralized vertebrate tissues and control their biomechanical and biochemical properties. Large, flexible, acidic noncollagenous proteins (ANCPs) have been shown to influence collagen mineralization but little is known about mineralization mechanisms with the aid of small proteins. Osteocalcin (OCN) is a small, highly structured biomolecule known as a multifunctional hormone in its undercarboxylated form. Here, we examined the potential mechanism of collagen intrafibrillar mineralization in vitro mediated by OCN as a model protein. Rapid and random extrafibrillar mineralization of flakey Ca-PO4 particles was observed by transmission electron microscopy mainly on the outer surfaces of collagen fibrils of a preformed collagen scaffold in the absence of the protein. In contrast, the protein stabilized hydrated, spherical nanoclusters of Ca-PO4 on the outer surface of the fibrils, thereby retarding extrafibrillar mineralization. The nanoclusters then infiltrated the fibrils resulting in intrafibrillar mineralization with HAP crystals aligned with the fibrils. This mechanism is similar to that observed for unstructured ANCPs. Results of fibrillogenesis and immunogold labeling studies showed that OCN was associated primarily with the fibrils, consistent with ex vivo studies on mineralizing turkey tendon. The present findings contribute to expanding our understanding of collagen intrafibrillar mineralization and provide insight into design synthetic macromolecular matrices for orthopedic implants and bone regeneration.

Recent Development of pH-Responsive Polymers for Cancer Nanomedicine.

Cancer remains a leading cause of death worldwide with more than 10 million new cases every year. Tumor-targeted nanomedicines have shown substantial improvements of the therapeutic index of anticancer agents, addressing the deficiencies of conventional chemotherapy, and have had a tremendous growth over past several decades. Due to the pathophysiological characteristics that almost all tumor tissues have lower pH in comparison to normal healthy tissues, among various tumor-targeted nanomaterials, pH-responsive polymeric materials have been one of the most prevalent approaches for cancer diagnosis and treatment. In this review, we summarized the types of pH-responsive polymers, describing their chemical structures and pH-response mechanisms; we illustrated the structure-property relationships of pH-responsive polymers and introduced the approaches to regulating their pH-responsive behaviors; we also highlighted the most representative applications of pH-responsive polymers in cancer imaging and therapy. This review article aims to provide general guidelines for the rational design of more effective pH-responsive nanomaterials for cancer diagnosis and treatment.

Systematically benchmarking peptide-MHC binding predictors: From synthetic to naturally processed epitopes.

A number of machine learning-based predictors have been developed for identifying immunogenic T-cell epitopes based on major histocompatibility complex (MHC) class I and II binding affinities. Rationally selecting the most appropriate tool has been complicated by the evolving training data and machine learning methods. Despite the recent advances made in generating high-quality MHC-eluted, naturally processed ligandome, the reliability of new predictors on these epitopes has yet to be evaluated. This study reports the latest benchmarking on an extensive set of MHC-binding predictors by using newly available, untested data of both synthetic and naturally processed epitopes. 32 human leukocyte antigen (HLA) class I and 24 HLA class II alleles are included in the blind test set. Artificial neural network (ANN)-based approaches demonstrated better performance than regression-based machine learning and structural modeling. Among the 18 predictors benchmarked, ANN-based mhcflurry and nn_align perform the best for MHC class I 9-mer and class II 15-mer predictions, respectively, on binding/non-binding classification (Area Under Curves = 0.911). NetMHCpan4 also demonstrated comparable predictive power. Our customization of mhcflurry to a pan-HLA predictor has achieved similar accuracy to NetMHCpan. The overall accuracy of these methods are comparable between 9-mer and 10-mer testing data. However, the top methods deliver low correlations between the predicted versus the experimental affinities for strong MHC binders. When used on naturally processed MHC-ligands, tools that have been trained on elution data (NetMHCpan4 and MixMHCpred) shows better accuracy than pure binding affinity predictor. The variability of false prediction rate is considerable among HLA types and datasets. Finally, structure-based predictor of Rosetta FlexPepDock is less optimal compared to the machine learning approaches. With our benchmarking of MHC-binding and MHC-elution predictors using a comprehensive metrics, a unbiased view for establishing best practice of T-cell epitope predictions is presented, facilitating future development of methods in immunogenomics.

Quantitatively Identifying the Roles of Interfacial Water and Solid Surface in Governing Peptide Adsorption.

Understanding the molecular mechanism of protein adsorption on solids is critical to their applications in materials synthesis and tissue engineering. Although the water phase at the surface/water interface has been recognized as three types: bulk water, intermediate water phase and surface-bound water layers, the roles of the water and surface in determining the protein adsorption are not clearly identified, particularly at the quantitative level. Herein, we provide a methodology involving the combination of microsecond strengthen sampling simulation and force integration to quantitatively characterize the water-induced contribution and the peptide-surface interactions into the adsorption free energy. Using hydroxyapatite and graphene surfaces as examples, we demonstrate how the distinct interfacial features dominate the delicate force balance between these two thermodynamics parameters, leading to surface preference/resistance to peptide adsorption. Specifically, the water layer provides sustained repelling force against peptide adsorption, as indicated by a monotonic increase in the water-induced free energy profile, whereas the contribution from the surface-peptide interactions is thermodynamically favorable to peptide adsorptions. More importantly, the revealed adsorption mechanism is critically dictated by the distribution of water phase, which plays a crucial role in establishing the force balance between the interactions of the peptide with the water layer and the surface. For the HAP surface, the charged peptide exhibits strong binding affinity to the surface, due to the controlling contribution of peptide-surface interaction in the intermediate water phase. The surface-bound water layers are observed as the origin of bioresistance of solid surfaces toward the adsorption of charge-neutral peptides. The preferred peptide adsorption on the graphene, however, is dominated by the surface-induced component at the water layers adjacent to the surface. Our results further elucidate that the intermediate water phase significantly shortens the effective range of the surface dispersion force, in contrast to the observation on the hydrophilic surface.

Osteocalcin facilitates calcium phosphate ion complex growth as revealed by free energy calculation.

The nanoscopic structural and thermodynamic basis of biomolecule-regulated assembly and crystallization of inorganic solids have a tremendous impact on the rational design of novel functional nanomaterials, but are concealed by many difficulties in molecular-level characterization. Here we demonstrate that the free energy calculation approach, enabled by combining advanced molecular simulation techniques, can unravel the structural and energetic mechanisms of protein-mediated inorganic solid nucleation. It is observed that osteocalcin (OCN), an important non-collagenous protein involved in regulating bone formation, promotes the growth of nanosized calcium phosphate (CaP) ion clusters from a supersaturated solution. Free energy calculation by umbrella sampling indicates that this effect by OCN is prominent at the scale of 1 to 3 nm ion-association complexes (IACs). The binding interactions between gamma-carboxyl glutamate and the C-terminal and, interestingly, the arginine side chains of OCN and IACs stabilize under-coordinated IACs, thus promoting their growth. The promoter effect of OCN on the enlargement and further aggregation of IACs into cluster assemblies of tens of nm are confirmed by conventional molecular dynamics simulation and dynamic light scattering experiments. To the best of our knowledge, this is the first time that the free energy landscape of the early stages of CaP nucleation is shown. The free energy change as a function of IAC size shares the feature of decreasing monotonically as shown previously for the calcium carbonate system. Therefore, the nucleation of both these major biominerals apparently involves an initial phase of liquid-like ionic aggregates. The structural and thermodynamic information regarding OCN-CaP interactions amplifies the current understanding of biomineralization mechanisms at the nanoscale, with general relevance to biomolecule-tuned fabrication of inorganic materials.

Method Evaluations for Adsorption Free Energy Calculations at the Solid/Water Interface through Metadynamics, Umbrella Sampling, and Jarzynski's Equality.

Considerable interest in characterizing protein/peptide-surface interactions has prompted extensive computational studies on calculations of adsorption free energy. However, in many cases, each individual study has focused on the application of free energy calculations to a specific system; therefore, it is difficult to combine the results into a general picture for choosing an appropriate strategy for the system of interest. Herein, three well-established computational algorithms are systemically compared and evaluated to compute the adsorption free energy of small molecules on two representative surfaces. The results clearly demonstrate that the characteristics of studied interfacial systems have crucial effects on the accuracy and efficiency of the adsorption free energy calculations. For the hydrophobic surface, steered molecular dynamics exhibits the highest efficiency, which appears to be a favorable method of choice for enhanced sampling simulations. However, for the charged surface, only the umbrella sampling method has the ability to accurately explore the adsorption free energy surface. The affinity of the water layer to the surface significantly affects the performance of free energy calculation methods, especially at the region close to the surface. Therefore, a general principle of how to discriminate between methodological and sampling issues based on the interfacial characteristics of the system under investigation is proposed.