Variational Hybrid Monte Carlo for Efficient Multi-Modal Data Sampling.

The Hamiltonian Monte Carlo (HMC) sampling algorithm exploits Hamiltonian dynamics to construct efficient Markov Chain Monte Carlo (MCMC), which has become increasingly popular in machine learning and statistics. Since HMC uses the gradient information of the target distribution, it can explore the state space much more efficiently than random-walk proposals, but may suffer from high autocorrelation. In this paper, we propose Langevin Hamiltonian Monte Carlo (LHMC) to reduce the autocorrelation of the samples. Probabilistic inference involving multi-modal distributions is very difficult for dynamics-based MCMC samplers, which is easily trapped in the mode far away from other modes. To tackle this issue, we further propose a variational hybrid Monte Carlo (VHMC) which uses a variational distribution to explore the phase space and find new modes, and it is capable of sampling from multi-modal distributions effectively. A formal proof is provided that shows that the proposed method can converge to target distributions. Both synthetic and real datasets are used to evaluate its properties and performance. The experimental results verify the theory and show superior performance in multi-modal sampling.

EMR Combined with CRB-65 Superior to CURB-65 in Predicting Mortality in Patients with Community-Acquired Pneumonia.

Background: Data about eosinophil-to-lymphocyte ratio (ELR) and eosinophil-to-monocyte ratio (EMR) in patients with community-acquired pneumonia (CAP) are rare. We aimed to evaluate the role of EMR and ELR in predicting disease severity and mortality in patients with CAP. Methods: A total of 454 patients (76 with severe CAP (SCAP), 378 with non-SCAP) were enrolled from November 18, 2020, and November 21, 2021. Laboratory examination on day 1 after admission was measured. The ELR and EMR values were calculated for patients. Propensity score matching (PSM) was performed to balance potential confounding factors. Binary logistic regression model was fitted to identify the potential risk factors for disease severity and Cox proportional hazards regression model analysis for mortality in CAP. Receiver operating characteristic (ROC) analysis was performed to distinguish disease severity and mortality. Results: EMR and ELR at admission were significantly lower in SCAP patients than in non-SCAP patients (P<0.001). EMR < 0.018 ([OR] = 12.104, 95% CI: 4.970-29.479), neutrophil (NEU) ([OR]=1.098, 95% CI:1.005-1.199), and age ([OR]=1.091, 95% CI:1.054-1.130) were independent risk factors for disease severity of CAP. EMR < 0.032 ([HR] = 5.816, 95% CI: 1.704-9.848) was an independent predictor of in-hospital mortality. Combining EMR or ELR with CRB-65 improved the overall accuracy of disease severity prediction (AUC from 0.894 to 0.937), the same as CURB-65. The area under the curve of EMR (AUC=0.704; 95% CI: 0.582-0.827) to predict in-hospital mortality was higher than that of CURB-65 (AUC=0.619; 95% CI: 0.484-0.754). Otherwise, EMR combined with CRB-65 (AUC=0.721; 95% CI: 0.592-0.851) had significantly higher diagnostic accuracy for in-hospital mortality than that of CURB-65 alone. Conclusion: EMR combined with CRB-65 was superior to CURB-65 in predicting mortality in patients with CAP. This new combination was simpler and easier to obtain for physicians in clinics or admission, and it was more convenient for early recognition of patients with poor prognoses.

Predictive Value of Lysophosphatidylcholine for Determining the Disease Severity and Prognosis of Elderly Patients with Community-Acquired Pneumonia.

Purpose: To investigate the clinical value of serum lysophosphatidylcholine (LPC) as a predictive biomarker for determining disease severity and mortality risk in hospitalized elderly patients with community-acquired pneumonia (CAP). Methods: This prospective, single-center study enrolled 208 elderly patients, including 67 patients with severe CAP (SCAP) and 141 with non-SCAP between November 1st, 2020, and November 30th, 2021 at the Qingdao Municipal Hospital, Shandong Province, China. The demographic and clinical parameters were recorded for all the included patients. Serum LPC levels were measured on day 1 and 6 after admission using ELISA. Propensity score matching (PSM) was used to balance the baseline variables between SCAP and non-SCAP patient groups. Receiver operative characteristic (ROC) curve analysis was used to compare the predictive performances of LPC and other clinical parameters in discriminating between SCAP and non-SCAP patients and determining the 30-day mortality risk of the hospitalized CAP patients. Univariate and multivariate logistic regression analyses were performed to identify the independent risk factors associated with SCAP. Cox proportional hazard regression analysis was used to determine if serum LPC was an independent risk factor for the 30-day mortality of CAP patients. Results: The serum LPC levels at admission were significantly higher in the non-SCAP patients than in the SCAP patients (P = 0.011). Serum LPC level <24.36 ng/mL, and PSI score were independent risk factors for the 30-day mortality in the elderly patients with CAP. The risk of 30-day mortality in the elderly CAP patients with low serum LPC levels (< 24.36ng/mL) was >5-fold higher than in the patients with high serum LPC levels (≥ 24.36ng/mL). Conclusion: Low serum LPC levels were associated with significantly higher disease severity and 30-day mortality in the elderly patients with CAP. Therefore, serum LPC is a promising predictive biomarker for the early identification of elderly CAP patients with poor prognosis.

A multi-dimensional approach for fractionating proteins using charged membranes.

In an effort to increase selectivity among proteins with crossflow ultrafiltration, we offer and demonstrate a comprehensive approach to fractionate proteins of similar molecular weight and relatively close pI values. This multidimensional approach involves optimizing membrane charge type and density together with operating conditions such as precise control of pH, ionic strength, and transmembrane pressure for reduced membrane fouling. Each filtration experiment was performed in cross-flow configuration for ∼20 min, allowing fast screening for optimal separation as determined by maximum selectivity, Ψ, and purity, P. Using our comprehensive approach for fractionating mixtures RNase A-lysozyme and BSA-hemoglobin, we obtained values of Ψ = 9.1, P = 95.7%, and Ψ = 6.5, P = 62.1%, respectively.

Predictive Value of Annenxin A1 for Disease Severity and Prognosis in Patients with Community-Acquired Pneumonia.

This prospective, single-center study evaluated the clinical utility of annenxin (Anx)A1 level as a biomarker for determining the severity of illness and predicting the risk of death in hospitalized patients with community-acquired pneumonia (CAP). A total of 105 patients (53 with severe [S]CAP, 52 with non-SCAP) were enrolled from December 2020 to June 2021. Demographic and clinical data were recorded. Serum AnxA1 concentration on days one and six after admission was measured by enzyme-linked immunosorbent assay. AnxA1 level at admission was significantly higher in SCAP patients than in non-SCAP patients (p < 0.001) irrespective of CAP etiology and was positively correlated with Pneumonia Severity Index and Confusion, Uremia, Respiratory Rate, Blood Pressure, and Age ≥ 65 Years score. AnxA1 level was significantly lower on day six after treatment than on day one (p = 0.01). Disease severity was significantly higher in patents with AnxA1 level ≥254.13 ng/mL than in those with a level <254.13 ng/mL (p < 0.001). Kaplan-Meier analysis of 30-day mortality showed that AnxA1 level ≤670.84 ng/mL was associated with a significantly higher survival rate than a level >670.84 ng/mL. These results indicate that AnxA1 is a useful biomarker for early diagnosis and prognostic assessment of CAP.

LHX2 Is a Potential Biomarker and Associated with Immune Infiltration in Breast Cancer.

Worldwide, breast cancer is the most common malignancy. LHX2, a member of the LIM homeobox gene family and a transcription factor, plays a crucial role in numerous tumors, but the function of LHX2 in breast cancer progression remains unknown. In this study, we show that LHX2 is upregulated in breast cancer tissues and positively correlated with breast cancer progression. Meanwhile, the clinical characteristics of breast cancer and LHX2 expression showed a strong correlation. GSEA showed that a high LHX2 expression may activate the T-cell activation pathway, PI3K/AKT/mTOR signaling pathway, and apoptosis pathway. Moreover, ssGSEA showed that Th1 cells and Th2 cells had a positive correlation with LHX2 expression in breast cancer. Experiments showed that LHX2 promotes the proliferation, colony formation, migration, and invasion of breast cancer cells. Immunohistochemistry and immunofluorescence assays helped to analyze LHX2-associated immune infiltration in breast cancer. A Western blot assay proved that LHX2 activated the PI3K/AKT/mTOR pathway and the apoptosis pathway. A TUNEL assay confirmed that LHX2 inhibited apoptosis. Taken together, LHX2 plays a vital role in breast cancer's progression and prognosis and could be an immune infiltration biomarker for breast cancer, and LHX2 activates the PI3K/AKT/mTOR pathway and apoptosis pathway in breast cancer.

Effect of bisphosphonate on hip fracture in patients with osteoporosis or osteopenia according to age: a meta-analysis and systematic review.

This meta-analysis and systematic review investigated the efficacy of bisphosphonates on the incidence of hip fracture (IHF) in patients of different ages with osteoporosis or osteopenia. We searched Web of Science, Embase, the Cochrane Database, and PubMed from inception to January 10, 2021, for trials reporting the effects of bisphosphonates on the IHF. We included only randomized, double-blind, placebo-controlled clinical trials. We pooled data using a random-effects meta-analysis with risk ratios (RRs) and reported 95% CIs. We also used the Cochran Q and I² statistics to assess the heterogeneity in the results of individual studies. The primary endpoints were the total numbers of people in the bisphosphonates and placebo groups and the numbers of IHFs during the follow-up periods. Bisphosphonates reduced the IHF with an overall effect (RR: 0.66; 95% CI: 0.56 to 0.77; zoledronic acid: RR: 0.60; 95% CI: 0.46 to 0.78; risedronate: RR: 0.74; 95% CI: 0.59 to 0.94, and alendronate: RR: 0.61; 95% CI: 0.40 to 0.95). The result of the heterogeneity assessment was I²=0, p=0.97. In all age groups (all ages, ≥55 years old, ≥65 years old), bisphosphonates reduced the IHF. In the ≥55 years old and ≥65 years old age groups, the RR and 95% CI were 0.63 and 0.43 to 0.93, and 0.60 and 0.44 to 0.81, respectively. Bisphosphonate reduced the IHF in the general population and all age groups (≥55 years old and ≥65 years old). Zoledronic acid, risedronate and alendronate reduced the IHF in osteoporosis or osteopenia populations. The association between bisphosphonate and the IHF does not appear to be influenced by age.

Minocycline protects neurons against glial cells-mediated bilirubin neurotoxicity.

Unconjugated bilirubin, the end product of heme catabolism and antioxidant, induced brain damage in human neonates is a well-recognized clinical syndrome. However, the cellular and molecular mechanisms underlying bilirubin neurotoxicity remain unclear. To characterize the sequence of events leading to bilirubin-induced neurotoxicity, we investigated whether bilirubin-induced glial activation was involved in bilirubin neurotoxicity by exposing co-cultured rat glial cells and cerebellar granule neurons (CGN) to bilirubin. We found that bilirubin could markedly induce the expression of TNF-α and iNOS in glial cells, and even at low concentrations, the co-culture of glial cells with neurons significantly enhances neurotoxicity of bilirubin. Pretreatment of the co-cultured cells with minocycline protected CGN from glia-mediated bilirubin neurotoxicity and inhibited overexpression of TNF-α and iNOS in glia. Furthermore, we found that high doses of bilirubin were able to induce glial injury, and minocycline attenuated bilirubin-induced glial cell death. Our data suggest that glial cells play an important role in brain damage caused by bilirubin, and minocycline blocks bilirubin-induced encephalopathy possibly by directly and indirectly inhibiting neuronal death pathways.

Enhanced stem cell pluripotency in surface-modified electrospun fibrous matrices.

A previously screened "hit chemistry" (N-[3-(dimethylamino)propyl] methacrylamide) that supports strong attachment and long-term self-renewal of ES cells is selected and grafted to poly(ether sulfone) (PES) fibrous matrices through plasma-induced graft polymerization. The 3D modified fibers exhibit higher cell proliferation and greater expression of pluripotency markers of mouse ES cells than 2D membranes. It is the first demonstration of scaling up an optimal synthetic surface chemistry in 2D using a high throughput synthesis, screening, and selection method to 3D that strongly influences pluripotent stem cell growth.

Discovery of low mucus adhesion surfaces.

Mucus secretion from the body is ubiquitous, and finding materials that resist mucus adhesion is a major technological challenge. Here, using a high throughput platform with photo-induced graft polymerization, we first rapidly synthesized, screened and tested a library of 55 different surfaces from six functional monomer classes to discover porcine intestinal low mucus adhesion surfaces using a 1h static mucus adsorption protocol. From this preliminary screen, two chemistries, a zwitterionic ([2-(acryloyloxy)ethyl] trimethylammonium chloride) and a multiple hydroxyl (N-[tris(hydroxymethyl)methyl]acrylamide) surface, exhibited significantly low mucus adhesion from a Langmuir-type isotherm when exposed to increasing concentrations of mucus for 24 h. Apolar or hydrophobic interactions were likely the dominant attractive forces during mucus binding since many polar or hydrophilic monomers reduced mucus adhesion. Hansen solubility parameters were used to illustrate the importance of monomer polarity and hydrogen bonding in reducing mucus adsorption. For a series of polyethylene glycol (PEG) monomers with changing molecular weight from 144 g mol⁻¹ to 1100 g mol⁻¹, we observed an excellent linear correlation (R²=0.998) between relative amount adsorbed and the distance from a water point in a specialized Hansen solubility parameter plot, emphasizing the role of surface-water interactions for PEG modified surfaces.

Combinatorial synthesis with high throughput discovery of protein-resistant membrane surfaces.

Using combinatorial methods, we synthesized a series of new vinyl amide monomers and graft-polymerized them to light-sensitive poly(ether sulfone) (PES) porous films for protein resistance. To increase the discovery rate and statistical confidence, we developed high throughput surface modification methods (HTP) that allow synthesis, screening and selection of desirable monomers from a large library in a relatively short time (days). A series of amide monomers were synthesized by amidation of methacryloyl chloride with amines and grafted onto commercial poly(ether sulfone) (PES) membranes using irradiation from atmospheric pressure plasma (APP). The modified PES membrane surfaces were then tested and screened for static protein adhesion using HTP. Hydroxyl amide monomers N-(3-hydroxypropyl)methacrylamide (A3), N-(4-hydroxybutyl)methacrylamide (A4), and N-(4-hydroxybutyl)methacrylamide (A6), ethylene glycol (EG) monomer N-(3-methoxypropyl)methacrylamide (A7), and N-(2-(dimethylamino)ethyl)-N-methylmethacrylamide (A8), and N-(2-(diethylamino)ethyl)-N-methylmethacrylamide (A9) all terminated with tertiary amines and were shown to have protein resistance. The PES membranes modified with these monomers exhibited both low protein adhesion (i.e. membrane plugging or fouling) and high flux. Their performance is comparable with previously identified best performing PEG and zwitterionic monomers, i.e. the so-called gold-standard for protein resistance. Combining a Hansen solubility parameter (HSP) analysis of the amide monomers and the HTP filtration results, we conclude that monomer solubility in water correlates with protein-resistant surfaces, presumably through its effects on surface-water interactions.

High throughput atmospheric pressure plasma-induced graft polymerization for identifying protein-resistant surfaces.

Three critical aspects of searching for and understanding how to find highly resistant surfaces to protein adhesion are addressed here with specific application to synthetic membrane filtration. They include the (i) discovery of a series of previously unreported monomers from a large library of monomers with high protein resistance and subsequent low fouling characteristics for membrane ultrafiltration of protein-containing fluids, (ii) development of a new approach to investigate protein-resistant mechanisms from structure-property relationships, and (iii) adaptation of a new surface modification method, called atmospheric pressure plasma-induced graft polymerization (APP), together with a high throughput platform (HTP), for low cost vacuum-free synthesis of anti-fouling membranes. Several new high-performing chemistries comprising two polyethylene glycol (PEG), two amines and one zwitterionic monomers were identified from a library (44 commercial monomers) of five different classes of monomers as strong protein-resistant monomers. Combining our analysis here, using the Hansen solubility parameters (HSP) approach, and data from the literature, we conclude that strong interactions with water (hydrogen bonding) and surface flexibility are necessary for producing the highest protein resistance. Superior protein-resistant surfaces and subsequent anti-fouling performance was obtained with the HTP-APP as compared with our earlier HTP-photo graft-induced polymerization (PGP).