Establishment and validation of a multivariate logistic model for risk factors of thyroid nodules using lasso regression screening

Objective

This study aims to analyze the association between the occurrence of thyroid nodules and various factors and to establish a risk factor model for thyroid nodules.

Methods

The study population was divided into two groups: a group with thyroid nodules and a group without thyroid nodules. Regression with the least absolute shrinkage and selection operator (Lasso) was applied to the complete dataset for variable selection. Binary logistic regression was used to analyze the relationship between various influencing factors and the prevalence of thyroid nodules.

Results

Based on the screening results of Lasso regression and the subsequent establishment of the Binary Logistic Regression Model on the training dataset, it was found that advanced age (OR=1.046, 95% CI: 1.033-1.060), females (OR = 1.709, 95% CI: 1.342-2.181), overweight individuals (OR = 1.546, 95% CI: 1.165-2.058), individuals with impaired fasting glucose (OR = 1.590, 95% CI: 1.193-2.122), and those with dyslipidemia (OR = 1.588, 95% CI: 1.197-2.112) were potential risk factors for thyroid nodule disease (p<0.05). The area under the curve (AUC) of the receiver operating characteristic (ROC) curve for the Binary Logistic Regression Model is 0.68 (95% CI: 0.64-0.72).

Conclusions

advanced age, females, overweight individuals, those with impaired fasting glucose, and individuals with dyslipidemia are potential risk factors for thyroid nodule disease.

Can We Boost N-Glycopeptide Identification Confidence? Smart Collision Energy Choice Taking into Account Structure and Search Engine

High confidence and reproducibility are still challenges in bottom-up mass spectrometric N-glycopeptide identification. The collision energy used in the MS/MS measurements and the database search engine used to identify the species are perhaps the two most decisive factors. We investigated how the structural features of N-glycopeptides and the choice of the search engine influence the optimal collision energy, delivering the highest identification confidence. We carried out LC-MS/MS measurements using a series of collision energies on a large set of N-glycopeptides with both the glycan and peptide part varied and studied the behavior of Byonic, pGlyco, and GlycoQuest scores. We found that search engines show a range of behavior between peptide-centric and glycan-centric, which manifests itself already in the dependence of optimal collision energy on m/z. Using classical statistical and machine learning methods, we revealed that peptide hydrophobicity, glycan and peptide masses, and the number of mobile protons also have significant and search-engine-dependent influence, as opposed to a series of other parameters we probed. We envisioned an MS/MS workflow making a smart collision energy choice based on online available features such as the hydrophobicity (described by retention time) and glycan mass (potentially available from a scout MS/MS). Our assessment suggests that this workflow can lead to a significant gain (up to 100%) in the identification confidence, particularly for low-scoring hits close to the filtering limit, which has the potential to enhance reproducibility of N-glycopeptide analyses. Data are available via MassIVE (MSV000093110).

Identification and validation of diagnostic signature genes in non-obstructive azoospermia by machine learning

Non-obstructive azoospermia (NOA) is a common cause of male infertility, and no specific diagnostic indicators exist. In this study, we used human testis datasets GSE45885, GSE45887, and GSE108886 from GEO database as training datasets, and screened 6 signature genes (all lowly expressed in the NOA group) using Boruta algorithm and Lasso regression: C12orf54, TSSK6, OR2H1, FER1L5, C9orf153, XKR3. The diagnostic efficacy of the above genes was examined by constructing models with LightGBM algorithm: the AUC (Area Under Curve) of both ROC and Precision-Recall curves for internal validation was 1.0 (p < 0.05). For the external validation dataset GSE145467 (human testis), the AUC of its ROC curve was 0.9 and that of its Precision-Recall curve was 0.833 (p < 0.05). Next, we confirmed the cellular localization of the above genes using human testis single-cell RNA sequencing dataset GSE149512, which were all located in spermatid. Besides, the downstream regulatory mechanisms of the above genes in spermatid were inferred by GSEA algorithm: C12orf54 may be involved in the repression of E2F-related and MYC-related pathways, TSSK6 and C9orf153 may be involved in the repression of MYC-related pathways, while FER1L5 may be involved in the repression of spermatogenesis pathway. Finally, we constructed a NOA model in mice using X-ray irradiation, and quantitative Real-time PCR results showed that C12orf54, TSSK6, OR2H1, FER1L5, and C9orf153 were all lowly expressed in NOA group. In summary, we have identified novel signature genes of NOA using machine learning methods and complete experimental validation, which will be helpful for its early diagnosis.

A Machine Learning Approach Reveals Distinct Predictors of Vaping Dependence for Adolescent Daily and Non-Daily Vapers in the COVID-19 Era

Since 2016, there has been a substantial rise in e-cigarette (vaping) dependence among young people. In this prospective cohort study, we aimed to identify the different predictors of vaping dependence over 3 months among adolescents who were baseline daily and non-daily vapers. We recruited ever-vaping Canadian residents aged 16-25 years on social media platforms and asked them to complete a baseline survey in November 2020. A validated vaping dependence score (0-23) summing up their responses to nine questions was calculated at the 3-month follow-up survey. Separate lasso regression models were developed to identify predictors of higher 3-month vaping dependence score among baseline daily and non-daily vapers. Of the 1172 participants, 643 (54.9%) were daily vapers with a mean age of 19.6 ± 2.6 years and 76.4% (n = 895) of them being female. The two models achieved adequate predictive performance. Place of last vape purchase, number of days a pod lasts, and the frequency of nicotine-containing vaping were the most important predictors for dependence among daily vapers, while race, sexual orientation and reporting treatment for heart disease were the most important predictors in non-daily vapers. These findings have implications for vaping control policies that target adolescents at different stages of vape use.

CHRDL1, NEFH, TAGLN and SYNM as novel diagnostic biomarkers of benign prostatic hyperplasia and prostate cancer

BACKGROUND

Prostate cancer (PCa) and benign prostatic hyperplasia (BPH) are common male diseases whose incidence rates gradually increase with age. They seriously affect men's physical health and quality of life. This study aimed to identify new biomarkers for the diagnosis of BPH and PCa.

METHODS

Two datasets, GSE28204 and GSE134051 (including human PCa and BPH), were downloaded from the GEO database. The batch effect was removed for merging, and then differential gene expression analysis was conducted to identify BPH and PCa cases. The diagnostic biomarkers of BPH and PCa were further screened using machine learning and bioinformatics. ROC curves were drawn to evaluate the diagnostic accuracy of the selected biomarkers. An online website and qPCR were used to preliminarily explore the expression levels of PCa biomarkers. The correlations between the expression of biomarkers and the tumor microenvironment, tumor mutation load and immunotherapy drugs were evaluated.

RESULTS

We identified fifteen genes (CHRDL1, DES, FLNC, GSTP1, MYL9, TGFB3, NEFH, TAGLN, SPARCL1, SYNM, TRPM8, HPN, PLA2G7, ENTPD5 and GPR160) as critical diagnostic biomarkers. After reviewing the literature on all selected biomarkers, we found few studies on the four genes CHRDL1, NEFH, TAGLN and SYNM in BPH or PCa. We defined these four genes as new potential diagnostic biomarkers (NPDBs) of BPH and PCa. All NPDBs were downregulated in PCa patients and PCa cell lines and upregulated in BPH patients and cell lines. When the immune landscape and mutation frequencies were analyzed, the results showed that the tumor microenvironment (TME), immune landscape, tumor mutation burden, and drug response were significantly correlated with NPDB expressions.

CONCLUSIONS

We found four new diagnostic markers of BPH and PCa, which may facilitate the early diagnosis, treatment, and immunotherapeutic responses assessment and may be of major value in guiding clinical practice.

Order of Occurrence of COVID-19 Symptoms

BACKGROUND AND OBJECTIVES

COVID-19 symptoms change after onset-some show early, others later. This article examines whether the order of occurrence of symptoms can improve diagnosis of COVID-19 before test results are available.

METHODS

In total, 483 individuals who completed a COVID-19 test were recruited through Listservs. Participants then completed an online survey regarding their symptoms and test results. The order of symptoms was set according to (a) whether the participant had a "history of the symptom" due to a prior condition; and (b) whether the symptom "occurred first," or prior to, other symptoms of COVID-19. Two LASSO (Least Absolute Shrinkage and Selection Operator) regression models were developed. The first model, referred to as "time-invariant," used demographics and symptoms but not the order of symptom occurrence. The second model, referred to as "time-sensitive," used the same data set but included the order of symptom occurrence.

RESULTS

The average cross-validated area under the receiver operating characteristic (AROC) curve for the time-invariant model was 0.784. The time-sensitive model had an AROC curve of 0.799. The difference between the 2 accuracy levels was statistically significant (α < .05).

CONCLUSION

The order of symptom occurrence made a statistically significant, but small, improvement in the accuracy of the diagnosis of COVID-19.

The Role of Symptom Clusters in Triage of COVID-19 Patients

BACKGROUND AND OBJECTIVE

COVID-19 manifests with a broad range of symptoms. This study investigates whether clusters of respiratory, gastrointestinal, or neurological symptoms can be used to diagnose COVID-19.

METHODS

We surveyed symptoms of 483 subjects who had completed COVID-19 laboratory tests in the last 30 days. The survey collected data on demographic characteristics, self-reported symptoms for different types of infections within 14 days of onset of illness, and self-reported COVID-19 test results. Robust LASSO regression was used to create 3 nested models. In all 3 models, the response variable was the COVID-19 test result. In the first model, referred to as the "main effect model," the independent variables were demographic characteristics, history of chronic symptoms, and current symptoms. The second model, referred to as the "hierarchical clustering model," added clusters of variables to the list of independent variables. These clusters were established through hierarchical clustering. The third model, referred to as the "interaction-terms model," also added clusters of variables to the list of independent variables; this time clusters were established through pairwise and triple-way interaction terms. Models were constructed on a randomly selected 80% of the data and accuracy was cross-validated on the remaining 20% of the data. The process was bootstrapped 30 times. Accuracy of the 3 models was measured using the average of the cross-validated area under the receiver operating characteristic curves (AUROCs).

RESULTS

In 30 bootstrap samples, the main effect model had an AUROC of 0.78. The hierarchical clustering model had an AUROC of 0.80. The interaction-terms model had an AUROC of 0.81. Both the hierarchical cluster model and the interaction model were significantly different from the main effect model (α = .04). Patients with different races/ethnicities, genders, and ages presented with different symptom clusters.

CONCLUSIONS

Using clusters of symptoms, it is possible to more accurately diagnose COVID-19 among symptomatic patients.

Development and validation of an online nomogram for predicting the outcome of open tracheotomy decannulation: a two-center retrospective analysis

BACKGROUND

Tracheotomy decannulation is critical for patients in the intensive care unit (ICU) to recover. In this study, we developed and validated an intuitive nomogram to predict the success rate of tracheotomy decannulation.

METHODS

We collected the data of 627 ICU patients before open tracheotomy decannulation from two medical institutions, including 466 patients (135 success and 331 failure) from the First Affiliated Hospital of Anhui Medical University as a training cohort, and 161 patients (57 success and 104 failure) from the Second Affiliated Hospital of Anhui Medical University as an external validation cohort. A least absolute shrinkage and multivariate logistic regression analysis were performed to determine the independent risk factors and construct the nomogram. The area under the receiver operating characteristic curve (AUC) was used to assess discrimination and the calibration plots were used to assess consistency. The clinical application was assessed using decision curve analysis and the clinical impact curve.

RESULTS

7 independent risk factors were eventually included in the prediction model. The AUC of the training cohort, internal validation and external validation were 0.932, 0.926, and 0.915, showing good discrimination. The model performed well in terms of calibration, decision curve analysis, and clinical impact curves. The superior performance of the model was also confirmed by external validation.

CONCLUSION

This nomogram can help ICU physicians identify high-risk patients for decannulation and plan their pre-decannulation treatment accordingly.

Predictive factors and nomogram to evaluate the risk of below-ankle re-amputation in patients with diabetic foot

BACKGROUND

Diabetes mellitus, as the most common metabolic disease, is common worldwide and represents a crucial global health concern. The purpose of this research was to investigate the related risk factors and to develop a re-amputation risk nomogram in diabetic patients who have undergone an amputation.

METHODS

A observational analysis was performed on 459 patients who have underwent amputation for diabetic foot from January 2014 through December 2019 at the First Affiliated Hospital of Wenzhou Medical University. The least absolute shrinkage and selection operator regression and stepwise regression methods were implemented to determine risk selection for the re-amputation risk model, and the predictive nomogram was established with these features. Calibration curve, receiver operating characteristic curve, and decision curve analysis of this re-amputation nomogram were assessed.

RESULTS

Predictors contained in this predictive model included smoking, glycated hemoglobin A1c (HbA1c), ankle-brachial index (ABI) and C-reactive protein (CRP). Good discrimination with a C-index of 0.725 (95% CI, 0.6624-0.7876) and good calibration were displayed with this predictive model. The decision curve analysis showed that this re-amputation nomogram predicting risk adds more benefit than none strategy if the threshold probability of a patient was >6% and <59%.

CONCLUSIONS

This novel re-amputation nomogram incorporating smoking, glycated hemoglobin A1c (HbA1c), ankle-brachial index (ABI), C-reactive protein (CRP), and smoking could be easily used to predict individual re-amputation risk prediction in diabetic foot patients who have undergone an amputation. In the future, further analysis and external testing will be needed as much as possible to reconfirm that this new Nomogram can accurately predict the risk of toe re-amputation.

A Lasso and a Regression Tree Mixed-Effect Model with Random Effects for the Level, the Residual Variance, and the Autocorrelation

Research in psychology is experiencing a rapid increase in the availability of intensive longitudinal data. To use such data for predicting feelings, beliefs, and behavior, recent methodological work suggested combinations of the longitudinal mixed-effect model with Lasso regression or with regression trees. The present article adds to this literature by suggesting an extension of these models that-in addition to a random effect for the mean level-also includes a random effect for the within-subject variance and a random effect for the autocorrelation. After introducing the extended mixed-effect location scale (E-MELS), the extended mixed-effect location-scale Lasso model (Lasso E-MELS), and the extended mixed-effect location-scale tree model (E-MELS trees), we show how its parameters can be estimated using a marginal maximum likelihood approach. Using real and simulated example data, we illustrate how to use E-MELS, Lasso E-MELS, and E-MELS trees for building prediction models to forecast individuals' daily nervousness. The article is accompanied by an R package (called mels) and functions that support users in the application of the suggested models.

Brain morphology, harm avoidance, and the severity of excessive internet use

As the previous studies have mainly focused on the reward system and the corresponding brain regions, the relationship between brain morphology and excessive internet use (EIU) were not clear; the purpose of the study was to investigate if the brain regions other than the reward system were associated with EIU. Data were acquired from 131 excessive internet users. Psychological measures included internet use, life quality, personality, mental illness symptoms, impulsivity, and thought suppression. The brain was scanned with 3T magnetic resonance imaging (MRI) and six types of brain morphological indexes were calculated. Lasso regression methods were used to select the predictors. Stepwise linear regression methods were used to build the models and verify the model. The variables remaining in the model were left precentral (curve), left superior temporal (surface area), right cuneus (folding index), right rostral anterior cingulate (folding index), and harm avoidance. The independent variable was the EIU score of the worst week in the past year. The study found that the brain morphological indexes other than the reward system, including the left precentral (curve), the left superior temporal (surface area), the right cuneus (folding index), and the right rostral anterior cingulate (folding index), can predict the severity of EIU, suggesting an extensive change in the brain. In this study, a whole‐brain data analysis was conducted and it was concluded that the changes in certain brain regions were more predictive than the reward system and psychological measures or more important for EIU.

Dietary Intake and Asthma in Preschoolers: A Logistic Lasso Regression Analysis

BACKGROUND

Asthma is a common chronic disease among children, especially preschoolers. Some evidence suggests that diet may play a role in asthma, but the current findings are contradictory. The objective of our study was to determine the association between dietary intake and asthma in preschool children aged 2-5 years.

METHODS

We selected preschool children aged 2-5 years with complete data on asthma diagnosis, diet, and body mass index (BMI) from the national health and nutrition examination survey (NHANES) database. In a selected population, children with self-reported asthma were included in the final sample. In children without self-reported asthma, we further used propensity score matching (PSM) to match age and sex for sampling, maintaining a ratio of 1:4 for cases. Lasso regression was used to identify dietary factors affecting asthma in preschoolers.

RESULTS

A total of 269 children with self-reported asthma and 1,076 children without self-reported asthma were included in our study. Univariate analysis showed that there were significant differences in ethnicity and dietary zinc intake between asthmatic children and children without asthma. After adjusting for all dietary and demographic variables, the results of logistic Lasso regression analysis showed that non-Hispanic black (β = 0.65), vitamin B12 (β = 0.14), and sodium (β = 0.05) were positively associated with childhood asthma, while Vitamin K (β = -0.04) was negatively associated with childhood asthma.

CONCLUSION

In conclusion, our study confirms that non-Hispanic black and dietary sodium intake are associated with a higher risk of asthma in preschoolers. In addition, our study found that dietary vitamin B12 was positively associated with childhood asthma, while vitamin K was negatively associated with childhood asthma.

Construction and Evaluation of a High-Frequency Hearing Loss Screening Tool for Community Residents

Early screening and detection of individuals at high risk of high-frequency hearing loss and identification of risk factors are critical to reduce the prevalence at community level. However, unlike those for individuals facing occupational auditory hazards, a limited number of hearing loss screening models have been developed for community residents. Therefore, this study used lasso regression with 10-fold cross-validation for feature selection and model construction on 38 questionnaire-based variables of 4010 subjects and applied the model to training and testing cohorts to obtain a risk score. The model achieved an area under the curve (AUC) of 0.844 in the model validation stage and individuals’ risk scores were subsequently stratified into low-, medium-, and high-risk categories. A total of 92.79% (1094/1179) of subjects in the high-risk category were confirmed to have hearing loss by audiometry test, which was 3.7 times higher than that in the low-risk group (25.18%, 457/1815). Half of the key indicators were related to modifiable contexts, and they were identified as significantly associated with the incident hearing loss. These results demonstrated that the developed model would be feasible to identify residents at high risk of hearing loss via regular community-level health examinations and detecting individualized risk factors, and eventually provide precision interventions.

Specific Risk Factors for Fatal Outcome in Critically Ill COVID-19 Patients: Results from a European Multicenter Study

(1) Background: The aim of our study was to identify specific risk factors for fatal outcome in critically ill COVID-19 patients. (2) Methods: Our data set consisted of 840 patients enclosed in the LEOSS registry. Using lasso regression for variable selection, a multifactorial logistic regression model was fitted to the response variable survival. Specific risk factors and their odds ratios were derived. A nomogram was developed as a graphical representation of the model. (3) Results: 14 variables were identified as independent factors contributing to the risk of death for critically ill COVID-19 patients: age (OR 1.08, CI 1.06–1.10), cardiovascular disease (OR 1.64, CI 1.06–2.55), pulmonary disease (OR 1.87, CI 1.16–3.03), baseline Statin treatment (0.54, CI 0.33–0.87), oxygen saturation (unit = 1%, OR 0.94, CI 0.92–0.96), leukocytes (unit 1000/μL, OR 1.04, CI 1.01–1.07), lymphocytes (unit 100/μL, OR 0.96, CI 0.94–0.99), platelets (unit 100,000/μL, OR 0.70, CI 0.62–0.80), procalcitonin (unit ng/mL, OR 1.11, CI 1.05–1.18), kidney failure (OR 1.68, CI 1.05–2.70), congestive heart failure (OR 2.62, CI 1.11–6.21), severe liver failure (OR 4.93, CI 1.94–12.52), and a quick SOFA score of 3 (OR 1.78, CI 1.14–2.78). The nomogram graphically displays the importance of these 14 factors for mortality. (4) Conclusions: There are risk factors that are specific to the subpopulation of critically ill COVID-19 patients.

Chrom-Lasso: a lasso regression-based model to detect functional interactions using Hi-C data

Hi-C is a genome-wide assay based on Chromosome Conformation Capture and high-throughput sequencing to decipher 3D chromatin organization in the nucleus. However, computational methods to detect functional interactions utilizing Hi-C data face challenges including the correction for various sources of biases and the identification of functional interactions with low counts of interacting fragments. We present Chrom-Lasso, a lasso linear regression model that removes complex biases assumption-free and identifies functional interacting loci with increased power by combining information of local reads distribution surrounding the area of interest. We showed that interacting regions identified by Chrom-Lasso are more enriched for 5C validated interactions and functional GWAS hits than that of GOTHiC and Fit-Hi-C. To further demonstrate the ability of Chrom-Lasso to detect interactions of functional importance, we performed time-series Hi-C and RNA-seq during T cell activation and exhaustion. We showed that the dynamic changes in gene expression and chromatin interactions identified by Chrom-Lasso were largely concordant with each other. Finally, we experimentally confirmed Chrom-Lasso’s finding that Erbb3 was co-regulated with distinct neighboring genes at different states during T cell activation. Our results highlight Chrom-Lasso’s utility in detecting weak functional interaction between cis-regulatory elements, such as promoters and enhancers.

A novel prognostic model predicts overall survival in patients with nasopharyngeal carcinoma based on clinical features and blood biomarkers

This study aims to develop and validate a novel prognostic model to estimate overall survival (OS) in nasopharyngeal carcinoma (NPC) patients based on clinical features and blood biomarkers. We assessed the model's incremental value to the TNM staging system, clinical treatment, and Epstein‐Barr virus (EBV) DNA copy number for individual OS estimation. We retrospectively analyzed 519 consecutive patients with NPC. A prognostic model was generated using the Lasso regression model in the training cohort. Then we compared the predictive accuracy of the novel prognostic model with TNM staging, clinical treatment, and EBV DNA copy number using concordance index (C‐index), time‐dependent ROC (tdROC), and decision curve analysis (DCA). Subsequently, we built a nomogram for OS incorporating the prognostic model, TNM staging, and clinical treatment. Finally, we stratified patients into high‐risk and low‐risk groups according to the model risk score, and we analyzed the survival time of these two groups using Kaplan–Meier survival plots. All results were validated in the independent validation cohort. Using the Lasso regression, we established a prognostic model consisting of 13 variables with respect to patient prognosis. The C‐index, tdROC, and DCA showed that the prognostic model had good predictive accuracy and discriminatory power in the training cohort than did TNM staging, clinical treatment, and EBV DNA copy number. Nomogram consisting of the prognostic model, TNM staging, clinical treatment, and EBV DNA copy number showed some superior net benefit. Based on the model risk score, we split the patients into two subgroups: low‐risk (risk score ≤ −1.423) and high‐risk (risk score > −1.423). There were significant differences in OS between the two subgroups of patients. Similar results were observed in the validation cohort. The proposed novel prognostic model based on clinical features and serological markers may represent a promising tool for estimating OS in NPC patients.

Gene expression profiles for an immunoscore model in bone and soft tissue sarcoma

Background: Immune infiltration is a prognostic marker to clinical outcomes in various solid tumors. However, reports that focus on bone and soft tissue sarcoma are rare. The study aimed to analyze and identify how immune components influence prognosis and develop a novel prognostic system for sarcomas.

Methods: We retrieved the gene expression data from 3 online databases (GEO, TCGA, and TARGET). The immune fraction was estimated using the CIBERSORT algorithm. After that, we re-clustered samples by K-means and constructed immunoscore by the least absolute shrinkage and selection operator (LASSO) Cox regression model. Next, to confirm the prognostic value, nomograms were constructed.

Results: 334 samples diagnosed with 8 tumor types (including osteosarcoma) were involved in our analysis. Patients were next re-clustered into three subgroups (OS, SAR1, and SAR2) through immune composition. Survival analysis showed a significant difference between the two soft tissue groups: patients with a higher proportion of CD8+ T cells, macrophages M1, and mast cells had favorable outcomes (p=0.0018). Immunoscore models were successfully established in OS and SAR2 groups consisting of 12 and 9 cell fractions, respectively. We found immunosocre was an independent factor for overall survival time. Patients with higher immunoscore had poor prognosis (p<0.0001). Patients with metastatic lesions scored higher than those counterparts with localized tumors (p<0.05).

Conclusions: Immune fractions could be a useful tool for the classification and prognosis of bone and soft tissue sarcoma patients. This proposed immunoscore showed a promising impact on survival prediction.

Utilizing electronic health record data to understand comorbidity burden among people living with HIV: a machine learning approach

OBJECTIVES

An understanding of the predictors of comorbidity among people living with HIV (PLWH) is critical for effective HIV care management. In this study, we identified predictors of comorbidity burden among PLWH based on machine learning models with electronic health record (EHR) data.

METHODS

The study population are individuals with a HIV diagnosis between January 2005 and December 2016 in South Carolina (SC). The change of comorbidity burden, represented by the Charlson Comorbidity Index (CCI) score, was measured by the score difference between pre- and post-HIV diagnosis, and dichotomized into a binary outcome variable. Thirty-five risk predictors from multiple domains were used to predict the increase in comorbidity burden based on the logistic least absolute shrinkage and selection operator (Lasso) regression analysis using 80% data for model development and 20% data for validation.

RESULTS

Of 8253 PLWH, the mean value of the CCI score difference was 0.8 ± 1.9 (range from 0 to 21) with 2328 (28.2%) patients showing an increase in CCI score after HIV diagnosis. Top predictors for an increase in CCI score using the LASSO model included older age at HIV diagnosis, positive family history of chronic conditions, tobacco use, longer duration with retention in care, having PEBA insurance, having low recent CD4+ cell count and duration of viral suppression.

CONCLUSION

The application of machine learning methods to EHR data could identify important predictors of increased comorbidity burden among PLWH with high accuracy. Results may enhance the understanding of comorbidities and provide the evidence based data for integrated HIV and comorbidity care management of PLWH.

County-level predictors of retention in care status among people living with HIV in South Carolina from 2010 to 2016: a data-driven approach

OBJECTIVE

The aim of this study was to examine the geospatial variation of retention in care (RIC) across the counties in South Carolina (SC) from 2010 to 2016 and identify the relevant county-level predictors.

DESIGN

Aggregated data on county-level RIC among HIV patients from 2010 to 2016 were retrieved from an electronic HIV/AIDS reporting system in SC Department of Health and Environmental Control. Sociological framework of health was used to select potential county-level predictors from multiple public datasets.

METHODS

Geospatial mapping was used to display the spatial heterogeneity of county-level RIC rate in SC. Generalized linear mixed effect regression with least absolute shrinkage and selection operator (LASSO) was employed to identify county-level predictors related to the change of RIC status over time. Confusion matrix and area under the curve statistics were used to evaluate model performance.

RESULTS

More than half of the counties had their RIC rates lower than the national average. The change of county-level RIC rate from 2010 to 2016 was not significant, and spatial heterogeneity in RIC rate was identified. A total of 22 of the 31 county-level predictors were selected by LASSO for predicting county-level RIC status. Counties with lower collective efficacy, larger proportions of men and/or persons with high education were more likely to have their RIC rates lower than the national average. In contrast, numbers of accessible mental health centres were positively related to county-level RIC status.

CONCLUSION

Spatial variation in RIC could be identified, and county-level factors associated with accessible healthcare facilities and social capital significantly contributed to these variations. Structural and individual interventions targeting these factors are needed to improve the county-level RIC and reduce the spatial variation in HIV care.

Influencing Indicators and Spatial Variation of Diabetes Mellitus Prevalence in Shandong, China: A Framework for Using Data-Driven and Spatial Methods

To control and prevent the risk of diabetes, diabetes studies have identified the need to better understand and evaluate the associations between influencing indicators and the prevalence of diabetes. One constraint has been that influencing indicators have been selected mainly based on subjective judgment and tested using traditional statistical modeling methods. We proposed a framework new to diabetes studies using data-driven and spatial methods to identify the most significant influential determinants of diabetes automatically and estimated their relationships. We used data from diabetes mellitus patients' health insurance records in Shandong province, China, and collected influencing indicators of diabetes prevalence at the county level in the sociodemographic, economic, education, and geographical environment domains. We specified a framework to identify automatically the most influential determinants of diabetes, and then established the relationship between these selected influencing indicators and diabetes prevalence. Our autocorrelation results showed that the diabetes prevalence in 12 Shandong cities was significantly clustered (Moran's I = 0.328, p < 0.01). In total, 17 significant influencing indicators were selected by executing binary linear regressions and lasso regressions. The spatial error regressions in different subgroups were subject to different diabetes indicators. Some positive indicators existed significantly like per capita fruit production and other indicators correlated with diabetes prevalence negatively like the proportion of green space. Diabetes prevalence was mainly subjected to the joint effects of influencing indicators. This framework can help public health officials to inform the implementation of improved treatment and policies to attenuate diabetes diseases.

Support Vector Machine for Analyzing Contributions of Brain Regions During Task-State fMRI

The mainstream method used for the analysis of task functional Magnetic Resonance Imaging (fMRI) data, is to obtain task-related active brain regions based on generalized linear models. Machine learning as a data-driven technical method is increasingly used in fMRI data analysis. The language task data, including math task and story task, of the Human Connectome Project (HCP) was used in this work. We chose a linear support vector machine as a classifier to classify math and story tasks and compared them with the activated brain regions of a SPM statistical analysis. As a result, 13 of the 25 regions used for classification in SVM were activated regions, and 12 were non-activated regions. In particular, the right Paracentral Lobule and right Rolandic Operculum which belong to non-activated regions, contributed most to the classification. Therefore, the differences found in machine learning can provide a new understanding of the physiological mechanisms of brain regions under different tasks.

Predictive model algorithms identifying early and advanced stage ER+/HER2- breast cancer in claims data

PURPOSE

Claims databases offer large populations for research, but lack clinical details. We aimed to develop predictive models to identify estrogen receptor positive (ER+) and human epidermal growth factor negative (HER2-) early breast cancer (ESBC) and advanced stage breast cancer (ASBC) in a claims database.

METHODS

Female breast cancer cases in Anthem's Cancer Care Quality Program served as the gold standard validation sample. Predictive models were developed from clinical knowledge and empirically from claims data using logistic and lasso regression. Model performance was assessed by classification rates and c-statistics. Models were applied to the HealthCore Integrated Research Database (claims) to identify cohorts of women with ER+/HER2- ESBC and ASBC.

RESULTS

The validation sample included 3184 women with ER+/HER2- ESBC and 1436 with ER+/HER2- ASBC. Predictive models for ER+/HER2- ESBC and ASBC included 25 and 20 factors, respectively. Models had robust discrimination in identifying cases (c-stat = 0.92 for ESBC and 0.95 for ASBC). Compared with a traditional a priori algorithm developed with clinical insight alone, the ER+/HER2- ASBC-predictive model had better positive predictive value (PPV) (0.91, 95% CI, 0.90-0.93, vs 0.69, 95% CI, 0.66-0.73) and sensitivity (0.54 vs 0.35). Models were applied to the claims database to identify cohorts of 33 001 and 3198 women with ER+/HER2- ESBC and ASBC.

CONCLUSION

We conducted a validation study and developed predictive models to identify in a claims database cohorts of women with ER+/HER2- ESBC and ASBC. The models identified large cohorts in the claims data that can be used to characterize indications in the evaluation of targeted therapies.

A Novel Recovery Method of Soft X-ray Spectrum Unfolding Based on Compressive Sensing

In the experiment of inertial confinement fusion, soft X-ray spectrum unfolding can provide important information to optimize the design of the laser and target. As the laser beams increase, there are limited locations for installing detection channels to obtain measurements, and the soft X-ray spectrum can be difficult to recover. In this paper, a novel recovery method of soft X-ray spectrum unfolding based on compressive sensing is proposed, in which (1) the spectrum recovery is formulated as a problem of accurate signal recovery from very few measurements (i.e., compressive sensing), and (2) the proper basis atoms are selected adaptively over a Legendre orthogonal basis dictionary with a large size and Lasso regression in the sense of ℓ1 norm, which enables the spectrum to be accurately recovered with little measured data from the limited detection channels. Finally, the presented approach is validated with experimental data. The results show that it can still achieve comparable accuracy from only 8 spectrometer detection channels as it has previously done from 14 detection channels. This means that the presented approach is capable of recovering spectrum from the data of limited detection channels, and it can be used to save more space for other detectors.

Using Exploratory Data Mining to Identify Important Correlates of Nonsuicidal Self-Injury Frequency

OBJECTIVE

Non-suicidal self-injury (NSSI) has been linked to many adverse outcomes, with more frequent NSSI increasing the likelihood of impairment, severity, and more serious self-harming behavior (e.g., suicidality; Andover & Gibb, 2010; Darke et al., 2010). Despite the determined importance of NSSI frequency in understanding the severity of one's behavior, there is still a need to identify which constructs may be influential in predicting frequency. The current study aimed to fill this gap by identifying which correlates are most important in relation to NSSI frequency through two exploratory data mining methods.

METHOD

Seven hundred twelve undergraduate students with a history of NSSI completed self-report measures of NSSI behavior, suicidality, cognitive-affective deficits, and psychopathology symptomology.

RESULTS

Both exploratory data mining methods, lasso regression and random forests, demonstrated number of NSSI methods to be the factor with the most importance in relation to lifetime NSSI frequency. Once this variable was removed, suicide plan and depressive symptomology were significant correlates across methods.

CONCLUSIONS

The current findings support the literature between NSSI frequency and NSSI methods, but also implicate suicide plans, an often-overlooked factor, and depression in NSSI severity.

Identification of MicroRNA Targets of Capsicum spp. Using MiRTrans-a Trans-Omics Approach

The microRNA (miRNA) can regulate the transcripts that are involved in eukaryotic cell proliferation, differentiation, and metabolism. Especially for plants, our understanding of miRNA targets, is still limited. Early attempts of prediction on sequence alignments have been plagued by enormous false positives. It is helpful to improve target prediction specificity by incorporating the other data sources such as the dependency between miRNA and transcript expression or even cleaved transcripts by miRNA regulations, which are referred to as trans-omics data. In this paper, we developed MiRTrans (Prediction of MiRNA targets by Trans-omics data) to explore miRNA targets by incorporating miRNA sequencing, transcriptome sequencing, and degradome sequencing. MiRTrans consisted of three major steps. First, the target transcripts of miRNAs were predicted by scrutinizing their sequence characteristics and collected as an initial potential targets pool. Second, false positive targets were eliminated if the expression of miRNA and its targets were weakly correlated by lasso regression. Third, degradome sequencing was utilized to capture the miRNA targets by examining the cleaved transcripts that regulated by miRNAs. Finally, the predicted targets from the second and third step were combined by Fisher's combination test. MiRTrans was applied to identify the miRNA targets for Capsicum spp. (i.e., pepper). It can generate more functional miRNA targets than sequence-based predictions by evaluating functional enrichment. MiRTrans identified 58 miRNA-transcript pairs with high confidence from 18 miRNA families conserved in eudicots. Most of these targets were transcription factors; this lent support to the role of miRNA as key regulator in pepper. To our best knowledge, this work is the first attempt to investigate the miRNA targets of pepper, as well as their regulatory networks. Surprisingly, only a small proportion of miRNA-transcript pairs were shared between degradome sequencing and expression dependency predictions, suggesting that miRNA targets predicted by a single technology alone may be prone to report false negatives.

Combined registration and motion correction of longitudinal retinal OCT data

Optical coherence tomography (OCT) has become an important modality for examination of the eye. To measure layer thicknesses in the retina, automated segmentation algorithms are often used, producing accurate and reliable measurements. However, subtle changes over time are difficult to detect since the magnitude of the change can be very small. Thus, tracking disease progression over short periods of time is difficult. Additionally, unstable eye position and motion alter the consistency of these measurements, even in healthy eyes. Thus, both registration and motion correction are important for processing longitudinal data of a specific patient. In this work, we propose a method to jointly do registration and motion correction. Given two scans of the same patient, we initially extract blood vessel points from a fundus projection image generated on the OCT data and estimate point correspondences. Due to saccadic eye movements during the scan, motion is often very abrupt, producing a sparse set of large displacements between successive B-scan images. Thus, we use lasso regression to estimate the movement of each image. By iterating between this regression and a rigid point-based registration, we are able to simultaneously align and correct the data. With longitudinal data from 39 healthy control subjects, our method improves the registration accuracy by 50% compared to simple alignment to the fovea and 22% when using point-based registration only. We also show improved consistency of repeated total retina thickness measurements.

Target Detection via Network Filtering

A method of 'network filtering' has been proposed recently to detect the effects of certain external perturbations on the interacting members in a network. However, with large networks, the goal of detection seems a priori difficult to achieve, especially since the number of observations available often is much smaller than the number of variables describing the effects of the underlying network. Under the assumption that the network possesses a certain sparsity property, we provide a formal characterization of the accuracy with which the external effects can be detected, using a network filtering system that combines Lasso regression in a sparse simultaneous equation model with simple residual analysis. We explore the implications of the technical conditions underlying our characterization, in the context of various network topologies, and we illustrate our method using simulated data.