Benchmark study of feature selection strategies for multi-omics data.

BACKGROUND: In the last few years, multi-omics data, that is, datasets containing different types of high-dimensional molecular variables for the same samples, have become increasingly available. To date, several comparison studies focused on feature selection methods for omics data, but to our knowledge, none compared these methods for the special case of multi-omics data. Given that these data have specific structures that differentiate them from single-omics data, it is unclear whether different feature selection strategies may be optimal for such data. In this paper, using 15 cancer multi-omics datasets we compared four filter methods, two embedded methods, and two wrapper methods with respect to their performance in the prediction of a binary outcome in several situations that may affect the prediction results. As classifiers, we used support vector machines and random forests. The methods were compared using repeated fivefold cross-validation. The accuracy, the AUC, and the Brier score served as performance metrics. RESULTS: The results suggested that, first, the chosen number of selected features affects the predictive performance for many feature selection methods but not all. Second, whether the features were selected by data type or from all data types concurrently did not considerably affect the predictive performance, but for some methods, concurrent selection took more time. Third, regardless of which performance measure was considered, the feature selection methods mRMR, the permutation importance of random forests, and the Lasso tended to outperform the other considered methods. Here, mRMR and the permutation importance of random forests already delivered strong predictive performance when considering only a few selected features. Finally, the wrapper methods were computationally much more expensive than the filter and embedded methods. CONCLUSIONS: We recommend the permutation importance of random forests and the filter method mRMR for feature selection using multi-omics data, where, however, mRMR is considerably more computationally costly.

Synergistic Effects of Different Levels of Genomic Data for the Staging of Lung Adenocarcinoma: An Illustrative Study.

Lung adenocarcinoma (LUAD) is a common and very lethal cancer. Accurate staging is a prerequisite for its effective diagnosis and treatment. Therefore, improving the accuracy of the stage prediction of LUAD patients is of great clinical relevance. Previous works have mainly focused on single genomic data information or a small number of different omics data types concurrently for generating predictive models. A few of them have considered multi-omics data from genome to proteome. We used a publicly available dataset to illustrate the potential of multi-omics data for stage prediction in LUAD. In particular, we investigated the roles of the specific omics data types in the prediction process. We used a self-developed method, Omics-MKL, for stage prediction that combines an existing feature ranking technique Minimum Redundancy and Maximum Relevance (mRMR), which avoids redundancy among the selected features, and multiple kernel learning (MKL), applying different kernels for different omics data types. Each of the considered omics data types individually provided useful prediction results. Moreover, using multi-omics data delivered notably better results than using single-omics data. Gene expression and methylation information seem to play vital roles in the staging of LUAD. The Omics-MKL method retained 70 features after the selection process. Of these, 21 (30%) were methylation features and 34 (48.57%) were gene expression features. Moreover, 18 (25.71%) of the selected features are known to be related to LUAD, and 29 (41.43%) to lung cancer in general. Using multi-omics data from genome to proteome for predicting the stage of LUAD seems promising because each omics data type may improve the accuracy of the predictions. Here, methylation and gene expression data may play particularly important roles.

A Diagnostic Model for Kawasaki Disease Based on Immune Cell Characterization From Blood Samples.

Background: Kawasaki disease (KD) is the leading cause of acquired heart disease in children. However, distinguishing KD from febrile infections early in the disease course remains difficult. Our goal was to estimate the immune cell composition in KD patients and febrile controls (FC), and to develop a tool for KD diagnosis. Methods: We used a machine-learning algorithm, CIBERSORT, to estimate the proportions of 22 immune cell types based on blood samples from children with KD and FC. Using these immune cell compositions, a diagnostic score for predicting KD was then constructed based on LASSO regression for binary outcomes. Results: In the training set (n = 496), a model was fit which consisted of eight types of immune cells. The area under the curve (AUC) values for diagnosing KD in a held-out test set (n = 212) and an external validation set (n = 36) were 0.80 and 0.77, respectively. The most common cell types in KD blood samples were monocytes, neutrophils, CD4+-naïve and CD8+ T cells, and M0 macrophages. The diagnostic score was highly correlated to genes that had been previously reported as associated with KD, such as interleukins and chemokine receptors, and enriched in reported pathways, such as IL-6/JAK/STAT3 and TNFα signaling pathways. Conclusion: Altogether, the diagnostic score for predicting KD could potentially serve as a biomarker. Prospective studies could evaluate how incorporating the diagnostic score into a clinical algorithm would improve diagnostic accuracy further.

Survival Factors and Metabolic Pathogenesis in Elderly Patients (≥65) With COVID-19: A Multi-Center Study.

Background: Elderly patients infected with COVID-19 are reported to be facing a substantially increased risk of mortality. Clinical characteristics, treatment options, and potential survival factors remain under investigation. This study aimed to fill this gap and provide clinically relevant factors associated with survival of elderly patients with COVID-19. Methods: In this multi-center study, elderly patients (age ≥65 years old) with laboratory-confirmed COVID-19 from 4 Wuhan hospitals were included. The clinical end point was hospital discharge or deceased with last date of follow-up on Jul. 08, 2020. Clinical, demographic, and laboratory data were collected. Univariate and multivariate analysis were performed to analyze survival and risk factors. A metabolic flux analysis using a large-scale molecular model was applied to investigate the pathogenesis of SARS-CoV-2 with regard to metabolism pathways. Results: A total of 223 elderly patients infected with COVID-19 were included, 91 (40.8%) were discharged and 132 (59.2%) deceased. Acute respiratory distress syndrome (ARDS) developed in 140 (62.8%) patients, 23 (25.3%) of these patients survived. Multivariate analysis showed that potential risk factors for mortality were elevated D-Dimer (odds ratio: 1.13 [95% CI 1.04 - 1.22], p = 0.005), high immune-related metabolic index (6.42 [95% CI 2.66-15.48], p < 0.001), and increased neutrophil-to-lymphocyte ratio (1.08 [95% 1.03-1.13], p < 0.001). Elderly patients receiving interferon atmotherapy showed an increased probability of survival (0.29 [95% CI 0.17-0.51], p < 0.001). Based on these factors, an algorithm (AlgSurv) was developed to predict survival for elderly patients. The metabolic flux analysis showed that 12 metabolic pathways including phenylalanine (odds ratio: 28.27 [95% CI 10.56-75.72], p < 0.001), fatty acid (15.61 [95% CI 6.66-36.6], p < 0.001), and pyruvate (12.86 [95% CI 5.85-28.28], p < 0.001) showed a consistently lower flux in the survivors vs. the deceased subgroup. This may reflect a key pathogenic mechanism of COVID-19 infection. Conclusion: Several factors such as interferon atmotherapy and recreased activity of specific metabolic pathways were found to be associated with survival of elderly patients. Based on these findings, a survival algorithm (AlgSurv) was developed to assist the clinical stratification for elderly patients. Dysregulation of the metabolic pathways revealed in this study may aid in the drug and vaccine development against COVID-19.

Development and validation of prognostic model for predicting mortality of COVID-19 patients in Wuhan, China.

Novel coronavirus 2019 (COVID-19) infection is a global public health issue, that has now affected more than 200 countries worldwide and caused a second wave of pandemic. Severe adult respiratory syndrome-CoV-2 (SARS-CoV-2) pneumonia is associated with a high risk of mortality. However, prognostic factors predicting poor clinical outcomes of individual patients with SARS-CoV-2 pneumonia remain under intensive investigation. We conducted a retrospective, multicenter study of patients with SARS-CoV-2 who were admitted to four hospitals in Wuhan, China from December 2019 to February 2020. Mortality at the end of the follow up period was the primary outcome. Factors predicting mortality were also assessed and a prognostic model was developed, calibrated and validated. The study included 492 patients with SARS-CoV-2 who were divided into three cohorts: the training cohort (n = 237), the validation cohort 1 (n = 120), and the validation cohort 2 (n = 135). Multivariate analysis showed that five clinical parameters were predictive of mortality at the end of follow up period, including advanced age [odds ratio (OR), 1.1/years increase (p < 0.001)], increased neutrophil-to-lymphocyte ratio [(NLR) OR, 1.14/increase (p < 0.001)], elevated body temperature on admission [OR, 1.53/°C increase (p = 0.005)], increased aspartate transaminase [OR, 2.47 (p = 0.019)], and decreased total protein [OR, 1.69 (p = 0.018)]. Furthermore, the prognostic model drawn from the training cohort was validated with validation cohorts 1 and 2 with comparable area under curves (AUC) at 0.912, 0.928, and 0.883, respectively. While individual survival probabilities were assessed, the model yielded a Harrell's C index of 0.758 for the training cohort, 0.762 for the validation cohort 1, and 0.711 for the validation cohort 2, which were comparable among each other. A validated prognostic model was developed to assist in determining the clinical prognosis for SARS-CoV-2 pneumonia. Using this established model, individual patients categorized in the high risk group were associated with an increased risk of mortality, whereas patients predicted to be in the low risk group had a higher probability of survival.

Rapamycin reduced pulmonary vascular remodelling by inhibiting cell proliferation via Akt/mTOR signalling pathway down-regulation in the carotid artery-jugular vein shunt pulmonary hypertension rat model.

OBJECTIVES: Pulmonary arterial hypertension (PAH) is a common complication of congenital heart disease. However, effective treatments for PAH are rare. This study aimed to investigate the inhibitory effects of rapamycin on PAH in the carotid artery-jugular vein (CA-JV) shunt PAH rat model as well as the mechanism underlying these effects. METHODS: Twenty-four Sprague-Dawley rats were randomized into the following 3 groups: a control group, a CA-JV shunt group and a treatment group. Rapamycin (2 mg/kg/day) was administered to the treatment group, and placebo was administered to the CA-JV shunt group. Haemodynamic evaluations, pulmonary tissue samplings for morphometry and immunofluorescence and western blot analyses were performed to evaluate the effects of rapamycin on PAH. RESULTS: Rapamycin attenuated the increase of right ventricular systolic pressure (RVSP) and the right ventricular (RV) hypertrophy (RVSP: CA-JV vs CA-JV + rapamycin, P = 0.017; RV: CA-JV vs CA-JV + rapamycin, P = 0.022), as well as the intrapulmonary vessel thickening (thickness index: CA-JV vs CA-JV + rapamycin, P = 0.028; area index: CA-JV vs CA-JV + rapamycin, P = 0.014), induced by overcirculation of the pulmonary vasculature in the CA-JV shunt-induced PAH rat model. Rapamycin decreased the expression level of the indicated cell proliferation marker (α-smooth muscle actin) in the lung vessel and mechanistic target of rapamycin (mTOR) pathway components (p-mTOR: CA-JV vs CA-JV + rapamycin, P = 0.004; p-Raptor: CA-JV vs CA-JV + rapamycin, P = 0.000; p-S6K1: CA-JV vs CA-JV + rapamycin, P = 0.000; p-Akt: CA-JV vs CA-JV + rapamycin, P = 0.001; p-Rheb: CA-JV vs CA-JV + rapamycin, P = 0.000) in pulmonary tissue. CONCLUSIONS: Rapamycin reduced pulmonary vascular remodelling by inhibiting cell proliferation via Akt/mTOR signalling pathway down-regulation in the CA-JV shunt-induced PAH model in rats. Thus, rapamycin may be a novel candidate drug for the treatment of PAH.

[Study on the facial and body terminal hair growth in women in Guangdong by using modified Ferriman-Gallwey scoring system].

OBJECTIVE: To study growth of facial and body terminal hair of women in Guangdong province and its relationship with age, menstrual irregularities and polycystic ovary, and determine normative cut-off score of modified Ferriman and Gallwey (mFG). METHODS: A cross-sectional study was conducted on 2988 women at age of 20-45 years from 16 communities of two urban and two rural regions in Guangdong province from June 2008 to July 2009. Terminal body hair growth was assessed by using the modified Ferriman and Gallwey (mFG) scoring system. The normative cut-off value of mFG were calculated by using the K-means cluster analysis (K=2). Those women were classified into following groups, including 982 women at group of ages of 20- years, 765 women at group of 26- years, 597 women at group of 31- years, 384 women at group of 36- years, 260 women at group of 41-45 years. Due to absence or errors of medical records, some cases were excluded from this study. Based on menses irregularities (MI), polycystic ovaries (PCO), there were 488 cases in MI group, 2413 cases in normal menses group, 568 cases in PCO group, and 2207 cases in non-PCO group finally. The incidences of acne, MI, acanthosis nigricans, and polycystic ovaries were also analyzed in all the hirsute groups. RESULTS: (1) among 2988 women, it was observed 149 women (5%) with mFG≥7,314 women (10.5%) with ≥5,747 women with mFG≥2. (2) Cluster analysis identified an mFG score of 5 as the cut-off value that define abnormal hirsute in the total population and all the sub-groups with/without MI or PCO; (3) Based on age classification, it was found that increased age was associated with decreased trends of the percentile and cut-off value of hirsutism. The value of hirsutism of mFG were 6 in group of 20- years, 5 in group of 26- years, 4 in groups of 31- years, 36- years and 41-45 years. (4) The prevalence of acne, menstrual irregularities and POC were 45.5% (143/314), 73.6% (231/314), 25.8% (81/314) in total population, 25.1% (671/2674), 16.1% (431/2674), 19.8% (529/2674) in normal hair women, which reached statistical difference (P<0.05). The prevalence of acne, menstrual irregularities and acanthosis nigricans were 44.4% (130/293), 23.2% (68/293), 4.1% (12/293) in those age hirsute groups, 25.3% (681/2695), 16.2% (437/2695), 1.9% (51/2695) in normal hair women, which reached statistical difference (P<0.05). CONCLUSIONS: (1) among women in Guangdong province, mFG scoring showed decreased trends in women with increasing age. (2) An mFG score≥5 was cut-off value in diagnosis of hirsutism. (3) The hirsute women exhibited higher incidence of acne, menses irregularity, and acanthosis nigricans than those of women with normal hair growth.