EFS: an ensemble feature selection tool implemented as R-package and web-application

Crohn's Patient Serum Proteomics Reveals Response Signature for Infliximab but not Vedolizumab

BACKGROUND

Crohn's disease is a chronic inflammatory bowel disease that affects the gastrointestinal tract. Common biologic families used to treat Crohn's are tumor necrosis factor (TNF)-α blockers (infliximab and adalimumab) and immune cell adhesion blockers (vedolizumab). Given their differing mechanisms of action, the ability to monitor response and predict treatment efficacy via easy-to-obtain blood draws remains an unmet need.

METHODS

To investigate these gaps in knowledge, we leveraged 2 prospective cohorts (LOVE-CD, TAILORIX) and profiled their serum using high-dimensional isobaric-labeled proteomics before treatment and 6 weeks after treatment initiation with either vedolizumab or infliximab.

RESULTS

The proportion of patients endoscopically responding to treatment was comparable among infliximab and vedolizumab cohorts; however, the impact of vedolizumab on patient sera was negligible. In contrast, infliximab treatment induced a robust response including increased blood-gas regulatory response proteins, and concomitant decreases in inflammation-related proteins. Further analysis comparing infliximab responders and nonresponders revealed a lingering innate immune enrichments in nonresponders and a unique protease regulation signature related to clotting cascades in responders. Lastly, using samples prior to infliximab treatment, we highlight serum protein biomarkers that potentially predict a positive response to infliximab treatment.

CONCLUSIONS

These results will positively impact the determination of appropriate patient treatment and inform the selection of clinical trial outcome metrics.

Machine learning-driven development of a disease risk score for COVID-19 hospitalization and mortality: a Swedish and Norwegian register-based study

Aims

To develop a disease risk score for COVID-19-related hospitalization and mortality in Sweden and externally validate it in Norway.

Method

We employed linked data from the national health registries of Sweden and Norway to conduct our study. We focused on individuals in Sweden with confirmed SARS-CoV-2 infection through RT-PCR testing up to August 2022 as our study cohort. Within this group, we identified hospitalized cases as those who were admitted to the hospital within 14 days of testing positive for SARS-CoV-2 and matched them with five controls from the same cohort who were not hospitalized due to SARS-CoV-2. Additionally, we identified individuals who died within 30 days after being hospitalized for COVID-19. To develop our disease risk scores, we considered various factors, including demographics, infectious, somatic, and mental health conditions, recorded diagnoses, and pharmacological treatments. We also conducted age-specific analyses and assessed model performance through 5-fold cross-validation. Finally, we performed external validation using data from the Norwegian population with COVID-19 up to December 2021.

Results

During the study period, a total of 124,560 individuals in Sweden were hospitalized, and 15,877 individuals died within 30 days following COVID-19 hospitalization. Disease risk scores for both hospitalization and mortality demonstrated predictive capabilities with ROC-AUC values of 0.70 and 0.72, respectively, across the entire study period. Notably, these scores exhibited a positive correlation with the likelihood of hospitalization or death. In the external validation using data from the Norwegian COVID-19 population (consisting of 53,744 individuals), the disease risk score predicted hospitalization with an AUC of 0.47 and death with an AUC of 0.74.

Conclusion

The disease risk score showed moderately good performance to predict COVID-19-related mortality but performed poorly in predicting hospitalization when externally validated.

Surface proteome of plasma extracellular vesicles as mechanistic and clinical biomarkers for malaria

PURPOSE

Malaria is a life-threatening mosquito-borne disease caused by Plasmodium parasites, mainly in tropical and subtropical countries. Plasmodium falciparum (P. falciparum) is the most prevalent cause on the African continent and responsible for most malaria-related deaths globally. Important medical needs are biomarkers for disease severity or disease outcome. A potential source of easily accessible biomarkers are blood-borne small extracellular vesicles (sEVs).

METHODS

We performed an EV Array to find proteins on plasma sEVs that are differentially expressed in malaria patients. Plasma samples from 21 healthy subjects and 15 malaria patients were analyzed. The EV array contained 40 antibodies to capture sEVs, which were then visualized with a cocktail of biotin-conjugated CD9, CD63, and CD81 antibodies.

RESULTS

We detected significant differences in the protein decoration of sEVs between healthy subjects and malaria patients. We found CD106 to be the best discrimination marker based on receiver operating characteristic (ROC) analysis with an area under the curve of > 0.974. Additional ensemble feature selection revealed CD106, Osteopontin, CD81, major histocompatibility complex class II DR (HLA-DR), and heparin binding EGF like growth factor (HBEGF) together with thrombocytes to be a feature panel for discrimination between healthy and malaria. TNF-R-II correlated with HLA-A/B/C as well as CD9 with CD81, whereas Osteopontin negatively correlated with CD81 and CD9. Pathway analysis linked the herein identified proteins to IFN-γ signaling.

CONCLUSION

sEV-associated proteins can discriminate between healthy individuals and malaria patients and are candidates for future predictive biomarkers.

TRIAL REGISTRATION

The trial was registered in the Deutsches Register Klinischer Studien (DRKS-ID: DRKS00012518).

AI-based disease risk score for community-acquired pneumonia hospitalization

Community-acquired pneumonia (CAP) is an acute infection involving the parenchyma of the lungs, which is acquired outside of the hospital. Population-wide real-world data and artificial intelligence (AI) were used to develop a disease risk score for CAP hospitalization among older individuals. The source population included residents in Denmark aged 65 years or older in the period January 1, 1996, to July 30, 2018. 137344 individuals were hospitalized for pneumonia during the study period for which, 5 controls were matched leading to a study population of 620908 individuals. The disease risk had an average accuracy of 0.79 based on 5-fold cross-validation in predicting CAP hospitalization. The disease risk score can be useful in clinical practice to identify individuals at higher risk of CAP hospitalization and intervene to minimize their risk of being hospitalized for CAP.

A Novel Risk Calculator Incorporating Clinical Parameters, Multiparametric Magnetic Resonance Imaging, and Prostate-Specific Membrane Antigen Positron Emission Tomography for Prostate Cancer Risk Stratification Before Transperineal Prostate Biopsy

Background

Prostate-specific membrane antigen (PSMA) positron emission tomography/computed tomography (PET/CT) can detect multiparametric magnetic resonance imaging (mpMRI)-invisible prostate tumours and improve the sensitivity of detection of prostate cancer (PCa) in comparison to mpMRI alone. Numerous risk calculators have been validated as tools for stratification of men at risk of being diagnosed with clinically significant (cs)PCa.

Objective

To develop a novel risk calculator using clinical parameters and imaging parameters from mpMRI and PSMA PET/CT in a cohort of patients undergoing mpMRI and PSMA PET/CT before biopsy.

Design setting and participants

A total of 291 men from the PRIMARY prospective trial underwent mpMRI and PSMA PET/CT before transperineal prostate biopsy with sampling of systematic and targeted cores.

Outcome measurements and statistical analysis

Novel risk calculators were developed using multivariable logistic regression analysis to predict detection of overall PCa (International Society of Urological Pathology grade group [GG] ≥1) and csPCa (GG ≥2). The risk calculators were then compared with the European Randomised Study of Screening for Prostate Cancer risk calculator incorporating mpMRI (ERSPC-MRI). Resampling methods were used to evaluate the discrimination and calibration of the risk calculators and to perform decision curve analysis.

Results and limitations

Age, prostate-specific antigen, prostate volume, and mpMRI Prostate Imaging-Reporting and Data System scores were included in the MRI risk calculator, resulting in area under the receiver operating characteristic curve (AUC) values of 0.791 for overall PCa (GG ≥1) and 0.812 for csPCa (GG ≥2). Addition of the maximum standardised uptake value (SUVmax) on PSMA PET/CT for the prostate lesion, and of SUVmax for the mpMRI lesions for the MRI-PSMA risk calculator resulted in AUCs of 0.831 for overall PCa and 0.876 for csPCa (≥ISUP2).The ERSPC-MRI risk calculator had AUCs of 0.758 (p = 0.02) for overall PCa and 0.805 (p = 0.001) for csPCa. Both the MRI and MRI-PSMA risk calculators were superior to the ERSPC-MRI for both overall PCa and csPCa.

Conclusions

These novel risk calculators incorporate clinical and radiological parameters for stratification of men at risk of csPCa. The risk calculator including PSMA PET/CT data is superior to a calculator incorporating mpMRI data alone.

Patient summary

We evaluated a new risk calculator that uses clinical information and results from two types of scan to predict the risk of clinically significant prostate cancer on prostate biopsy. This risk model can guide patients and clinicians in shared decision-making and may help in avoiding unnecessary prostate biopsies.

Nomogram for Predicting Intraoperative Hemodynamic Instability in Patients With Normotensive Pheochromocytoma

CONTEXT

Intraoperative hemodynamic instability (HI) deteriorates surgical outcomes of patients with normotensive pheochromocytoma (NP).

OBJECTIVE

To characterize the hemodynamics of NP and develop and externally validate a prediction model for intraoperative HI.

METHODS

Data on 117 patients with NP (derivation cohort) and 40 patients with normotensive adrenal myelolipoma (NAM) who underwent laparoscopic adrenalectomy from January 2011 to November 2021 were retrospectively collected. Data on 22 patients with NP (independent validation cohort) were collected from another hospital during the same period. The hemodynamic characteristics of patients with NP and NAM were compared. Machine learning models were used to identify risk factors associated with HI. The final model was visualized via a nomogram.

RESULTS

Forty-eight (41%) out of 117 patients experienced HI, which was significantly more than that for NAM. A multivariate logistic regression including age, tumor size, fasting plasma glucose, and preoperative systolic blood pressure showed good discrimination measured by area under curve (0.8286; 95% CI 0.6875-0.9696 and 0.7667; 95% CI 0.5386-0.9947) for predicting HI in internal and independent validation cohorts, respectively. The sensitivities and positive predictive values were 0.6667 and 0.7692 for the internal and 0.9167 and 0.6111 for the independent validations, respectively. The final model was visualized via a nomogram and yielded net benefits across a wide range of risk thresholds in decision curve analysis.

CONCLUSION

Patients with NP experienced HI during laparoscopic adrenalectomy. The nomogram can be used for individualized prediction of intraoperative HI in patients with NP.

Nomogram for Predicting Early Mortality after Umbilical Cord Blood Transplantation in Children with Inborn Errors of Immunity

PURPOSE

Pediatric patients with inborn errors of immunity (IEI) undergoing umbilical cord blood transplantation (UCBT) are at risk of early mortality. Our aim was to develop and validate a prediction model for early mortality after UCBT in pediatric IEI patients based on pretransplant factors.

METHODS

Data from 230 pediatric IEI patients who received their first UCBT between 2014 and 2021 at a single center were analyzed retrospectively. Data from 2014-2019 and 2020-2021 were used as training and validation sets, respectively. The primary outcome of interest was early mortality. Machine learning algorithms were used to identify risk factors associated with early mortality and to build predictive models. The model with the best performance was visualized using a nomogram. Discriminative ability was measured using the area under the curve (AUC) and decision curve analysis.

RESULTS

Fifty days was determined as the cutoff for distinguishing early mortality in pediatric IEI patients undergoing UCBT. Of the 230 patients, 43 (18.7%) suffered early mortality. Multivariate logistic regression with pretransplant albumin, CD4 (absolute count), elevated C-reactive protein, and medical history of sepsis showed good discriminant AUC values of 0.7385 (95% CI, 0.5824-0.8945) and 0.827 (95% CI, 0.7409-0.9132) in predicting early mortality in the validation and training sets, respectively. The sensitivity and specificity were 0.5385 and 0.8154 for validation and 0.7667 and 0.7705 for training, respectively. The final model yielded net benefits across a reasonable range of risk thresholds.

CONCLUSION

The developed nomogram can predict early mortality in pediatric IEI patients undergoing UCBT.

Metabolite and Proteomic Profiling of Serum Reveals the Differences in Molecular Immunity between Min and Large White Pig Breeds

Pig diseases seriously threaten the health of pigs and the benefits of pig production. Previous research has indicated that Chinese native pigs, such as the Min (M) pig, has a better disease resistance ability than Large White (LW) pigs. However, the molecular mechanism of this resistance is still unclear. In our study, we used serum untargeted metabolomics and proteomics, interrogated to characterize differences in the molecular immunities between six resistant and six susceptible pigs raised in the same environment. A total of 62 metabolites were identified as being significantly exhibited in M and LW pigs. Ensemble feature selection (EFS) machine learning methods were used to predict biomarkers of metabolites and proteins, and the top 30 were selected and retained. Weighted gene co-expression network analysis (WGCNA) confirmed that four key metabolites, PC (18:1 (11 Z)/20:0), PC (14:0/P-18: 0), PC (18:3 (6 Z, 9 Z, 12 Z)/16:0), and PC (16:1 (9 Z)/22:2 (13 Z, 16 Z)), were significantly associated with phenotypes, such as cytokines, and different pig breeds. Correlation network analysis showed that 15 proteins were significantly correlated with the expression of both cytokines and unsaturated fatty acid metabolites. Quantitative trait locus (QTL) co-location analysis results showed that 13 of 15 proteins co-localized with immune or polyunsaturated fatty acid (PUFA)-related QTL. Moreover, seven of them co-localized with both immune and PUFA QTLs, including proteasome 20S subunit beta 8 (PSMB8), mannose binding lectin 1 (MBL1), and interleukin-1 receptor accessory protein (IL1RAP). These proteins may play important roles in regulating the production or metabolism of unsaturated fatty acids and immune factors. Most of the proteins could be validated with parallel reaction monitoring, which suggests that these proteins may play an essential role in producing or regulating unsaturated fatty acids and immune factors to cope with the adaptive immunity of different pig breeds. Our study provides a basis for further clarifying the disease resistance mechanism of pigs.

Circulating microparticle proteins predict pregnancies complicated by placenta accreta spectrum

Placenta accreta spectrum (PAS) is characterized by abnormal attachment of the placenta to the uterus, and attempts at placental delivery can lead to catastrophic maternal hemorrhage and death. Multidisciplinary delivery planning can significantly improve outcomes; however, current diagnostics are lacking as approximately half of pregnancies with PAS are undiagnosed prior to delivery. This is a nested case-control study of 35 cases and 70 controls with the primary objective of identifying circulating microparticle (CMP) protein panels that identify pregnancies complicated by PAS. Size exclusion chromatography and liquid chromatography with tandem mass spectrometry were used for CMP protein isolation and identification, respectively. A two-step iterative workflow was used to establish putative panels. Using plasma sampled at a median of 26 weeks' gestation, five CMP proteins distinguished PAS from controls with a mean area under the curve (AUC) of 0.83. For a separate sample taken at a median of 35 weeks' gestation, the mean AUC was 0.78. In the second trimester, canonical pathway analyses demonstrate over-representation of processes related to iron homeostasis and erythropoietin signaling. In the third trimester, these analyses revealed abnormal immune function. CMP proteins classify PAS well prior to delivery and have potential to significantly reduce maternal morbidity and mortality.

M6A-GSMS: Computational identification of N6-methyladenosine sites with GBDT and stacking learning in multiple species

N⁶-methyladenosine (m⁶A) is one of the most abundant forms of RNA methylation modifications currently known. It involves a wide range of biological processes, including degradation, stability, alternative splicing, etc. Therefore, the development of convenient and efficient m⁶A prediction technologies are urgent. In this work, a novel predictor based on GBDT and stacking learning is developed to identify m⁶A sites, which is called M6A-GSMS. To achieve accurate prediction, we explore RNA sequence information from four aspects: correlation, structure, physicochemical properties and pseudo ribonucleic acid composition. After using the GBDT algorithm for feature selection, a stacking model is constructed by combining seven basic classifiers. Compared with other state-of-the-art methods, the results show that M6A-GSMS can obtain excellent performance for identifying the m⁶A sites. The prediction accuracy of A.thaliana, D.melanogaster, M.musculus, S.cerevisiae and Human reaches 88.4%, 60.8%, 80.5%, 92.4% and 61.8%, respectively. This method provides an effective prediction for the investigation of m⁶A sites. In addition, all the datasets and codes are currently available at https://github.com/Wang-Jinyue/M6A-GSMS.Communicated by Ramaswamy H. Sarma.

Eleven quick tips for data cleaning and feature engineering

Applying computational statistics or machine learning methods to data is a key component of many scientific studies, in any field, but alone might not be sufficient to generate robust and reliable outcomes and results. Before applying any discovery method, preprocessing steps are necessary to prepare the data to the computational analysis. In this framework, data cleaning and feature engineering are key pillars of any scientific study involving data analysis and that should be adequately designed and performed since the first phases of the project. We call "feature" a variable describing a particular trait of a person or an observation, recorded usually as a column in a dataset. Even if pivotal, these data cleaning and feature engineering steps sometimes are done poorly or inefficiently, especially by beginners and unexperienced researchers. For this reason, we propose here our quick tips for data cleaning and feature engineering on how to carry out these important preprocessing steps correctly avoiding common mistakes and pitfalls. Although we designed these guidelines with bioinformatics and health informatics scenarios in mind, we believe they can more in general be applied to any scientific area. We therefore target these guidelines to any researcher or practitioners wanting to perform data cleaning or feature engineering. We believe our simple recommendations can help researchers and scholars perform better computational analyses that can lead, in turn, to more solid outcomes and more reliable discoveries.

Social Determinants Contribute to Disparities in Test Positivity, Morbidity and Mortality: Data from a Multi-Ethnic Cohort of 1094 GU Cancer Patients Undergoing Assessment for COVID-19

Background: The COVID-19 pandemic exploits existing inequalities in the social determinants of health (SDOH) that influence disease burden and access to healthcare. The role of health behaviours and socioeconomic status in genitourinary (GU) malignancy has also been highlighted. Our aim was to evaluate predictors of patient-level and neighbourhood-level factors contributing to disparities in COVID-19 outcomes in GU cancer patients. Methods: Demographic information and co-morbidities for patients screened for COVID-19 across the Mount Sinai Health System (MSHS) up to 10 June 2020 were included. Descriptive analyses and ensemble feature selection were performed to describe the relationships between these predictors and the outcomes of positive SARS-CoV-2 RT-PCR test, COVID-19-related hospitalisation, intubation and death. Results: Out of 47,379 tested individuals, 1094 had a history of GU cancer diagnosis; of these, 192 tested positive for SARS-CoV-2. Ensemble feature selection identified social determinants including zip code, race/ethnicity, age, smoking status and English as the preferred first language—being the majority of significant predictors for each of this study’s four COVID-19-related outcomes: a positive test, hospitalisation, intubation and death. Patient and neighbourhood level SDOH including zip code/ NYC borough, age, race/ethnicity, smoking status, and English as preferred language are amongst the most significant predictors of these clinically relevant outcomes for COVID-19 patients. Conclusion: Our results highlight the importance of these SDOH and the need to integrate SDOH in patient electronic medical records (EMR) with the goal to identify at-risk groups. This study’s results have implications for COVID-19 research priorities, public health goals, and policy implementations.

Vision for Improving Pregnancy Health: Innovation and the Future of Pregnancy Research

Understanding, predicting, and preventing pregnancy disorders have been a major research target. Nonetheless, the lack of progress is illustrated by research results related to preeclampsia and other hypertensive pregnancy disorders. These remain a major cause of maternal and infant mortality worldwide. There is a general consensus that the rate of progress toward understanding pregnancy disorders lags behind progress in other aspects of human health. In this presentation, we advance an explanation for this failure and suggest solutions. We propose that progress has been impeded by narrowly focused research training and limited imagination and innovation, resulting in the failure to think beyond conventional research approaches and analytical strategies. Investigations have been largely limited to hypothesis-generating approaches constrained by attempts to force poorly defined complex disorders into a single "unifying" hypothesis. Future progress could be accelerated by rethinking this approach. We advise taking advantage of innovative approaches that will generate new research strategies for investigating pregnancy abnormalities. Studies should begin before conception, assessing pregnancy longitudinally, before, during, and after pregnancy. Pregnancy disorders should be defined by pathophysiology rather than phenotype, and state of the art agnostic assessment of data should be adopted to generate new ideas. Taking advantage of new approaches mandates emphasizing innovation, inclusion of large datasets, and use of state of the art experimental and analytical techniques. A revolution in understanding pregnancy-associated disorders will depend on networks of scientists who are driven by an intense biological curiosity, a team spirit, and the tools to make new discoveries.

Evaluation of machine learning strategies for imaging confirmed prostate cancer recurrence prediction on electronic health records

BACKGROUND

The main screening parameter to monitor prostate cancer recurrence (PCR) after primary treatment is the serum concentration of prostate-specific antigen (PSA). In recent years, Ga-68-PSMA PET/CT has become an important method for additional diagnostics in patients with biochemical recurrence.

PURPOSE

While Ga-68-PSMA PET/CT performs better, it is an expensive, invasive, and time-consuming examination. Therefore, in this study, we aim to employ modern multivariate Machine Learning (ML) methods on electronic health records (EHR) of prostate cancer patients to improve the prediction of imaging confirmed PCR (IPCR).

METHODS

We retrospectively analyzed the clinical information of 272 patients, who were examined using Ga-68-PSMA PET/CT. The PSA values ranged from 0 ng/mL to 2270.38 ng/mL with a median PSA level at 1.79 ng/mL. We performed a descriptive analysis using Logistic Regression. Additionally, we evaluated the predictive performance of Logistic Regression, Support Vector Machine, Gradient Boosting, and Random Forest. Finally, we assessed the importance of all features using Ensemble Feature Selection (EFS).

RESULTS

The descriptive analysis found significant associations between IPCR and logarithmic PSA values as well as between IPCR and performed hormonal therapy. Our models were able to predict IPCR with an AUC score of 0.78 ± 0.13 (mean ± standard deviation) and a sensitivity of 0.997 ± 0.01. Features such as PSA, PSA doubling time, PSA velocity, hormonal therapy, radiation treatment, and injected activity show high importance for IPCR prediction using EFS.

CONCLUSION

This study demonstrates the potential of employing a multitude of parameters into multivariate ML models to improve identification of non-recurring patients compared to the current focus on the main screening parameter (PSA). We showed that ML models are able to predict IPCR, detectable by Ga-68-PSMA PET/CT, and thereby pave the way for optimized early imaging and treatment.

A Clinical Decision Support System for the Prediction of Quality of Life in ALS

Amyotrophic Lateral Sclerosis (ALS), also known as Motor Neuron Disease (MND), is a rare and fatal neurodegenerative disease. As ALS is currently incurable, the aim of the treatment is mainly to alleviate symptoms and improve quality of life (QoL). We designed a prototype Clinical Decision Support System (CDSS) to alert clinicians when a person with ALS is experiencing low QoL in order to inform and personalise the support they receive. Explainability is important for the success of a CDSS and its acceptance by healthcare professionals. The aim of this work isto announce our prototype (C-ALS), supported by a first short evaluation of its explainability. Given the lack of similar studies and systems, this work is a valid proof-of-concept that will lead to future work. We developed a CDSS that was evaluated by members of the team of healthcare professionals that provide care to people with ALS in the ALS/MND Multidisciplinary Clinic in Dublin, Ireland. We conducted a user study where participants were asked to review the CDSS and complete a short survey with a focus on explainability. Healthcare professionals demonstrated some uncertainty in understanding the system’s output. Based on their feedback, we altered the explanation provided in the updated version of our CDSS. C-ALS provides local explanations of its predictions in a post-hoc manner, using SHAP (SHapley Additive exPlanations). The CDSS predicts the risk of low QoL in the form of a probability, a bar plot shows the feature importance for the specific prediction, along with some verbal guidelines on how to interpret the results. Additionally, we provide the option of a global explanation of the system’s function in the form of a bar plot showing the average importance of each feature. C-ALS is available online for academic use.

POSREG: proteomic signature discovered by simultaneously optimizing its reproducibility and generalizability

Mass spectrometry-based proteomic technique has become indispensable in current exploration of complex and dynamic biological processes. Instrument development has largely ensured the effective production of proteomic data, which necessitates commensurate advances in statistical framework to discover the optimal proteomic signature. Current framework mainly emphasizes the generalizability of the identified signature in predicting the independent data but neglects the reproducibility among signatures identified from independently repeated trials on different sub-dataset. These problems seriously restricted the wide application of the proteomic technique in molecular biology and other related directions. Thus, it is crucial to enable the generalizable and reproducible discovery of the proteomic signature with the subsequent indication of phenotype association. However, no such tool has been developed and available yet. Herein, an online tool, POSREG, was therefore constructed to identify the optimal signature for a set of proteomic data. It works by (i) identifying the proteomic signature of good reproducibility and aggregating them to ensemble feature ranking by ensemble learning, (ii) assessing the generalizability of ensemble feature ranking to acquire the optimal signature and (iii) indicating the phenotype association of discovered signature. POSREG is unique in its capacity of discovering the proteomic signature by simultaneously optimizing its reproducibility and generalizability. It is now accessible free of charge without any registration or login requirement at https://idrblab.org/posreg/.

Dissecting the Genome for Drug Response Prediction

The prediction of the cancer cell lines sensitivity to a specific treatment is one of the current challenges in precision medicine. With omics and pharmacogenomics data being available for over 1000 cancer cell lines, several machine learning and deep learning algorithms have been proposed for drug sensitivity prediction. However, deciding which omics data to use and which computational methods can efficiently incorporate data from different sources is the challenge which several research groups are working on. In this review, we summarize recent advances in the representative computational methods that have been developed in the last 2 years on three public datasets: COSMIC, CCLE, NCI-60. These methods aim to improve the prediction of the cancer cell lines sensitivity to a given treatment by incorporating drug's chemical information in the input or using a priori feature selection. Finally, we discuss the latest published method which aims to improve the prediction of clinical drug response of real patients starting from cancer cell line molecular profiles.

Improving the Classification of Alzheimer's Disease Using Hybrid Gene Selection Pipeline and Deep Learning

Alzheimer’s is a progressive, irreversible, neurodegenerative brain disease. Even with prominent symptoms, it takes years to notice, decode, and reveal Alzheimer’s. However, advancements in technologies, such as imaging techniques, help in early diagnosis. Still, sometimes the results are inaccurate, which delays the treatment. Thus, the research in recent times focused on identifying the molecular biomarkers that differentiate the genotype and phenotype characteristics. However, the gene expression dataset’s generated features are huge, 1,000 or even more than 10,000. To overcome such a curse of dimensionality, feature selection techniques are introduced. We designed a gene selection pipeline combining a filter, wrapper, and unsupervised method to select the relevant genes. We combined the minimum Redundancy and maximum Relevance (mRmR), Wrapper-based Particle Swarm Optimization (WPSO), and Auto encoder to select the relevant features. We used the GSE5281 Alzheimer’s dataset from the Gene Expression Omnibus We implemented an Improved Deep Belief Network (IDBN) with simple stopping criteria after choosing the relevant genes. We used a Bayesian Optimization technique to tune the hyperparameters in the Improved Deep Belief Network. The tabulated results show that the proposed pipeline shows promising results.

Transcriptome-Wide Analysis of Human Liver Reveals Age-Related Differences in the Expression of Select Functional Gene Clusters and Evidence for a PPP1R10-Governed 'Aging Cascade'

A transcriptome-wide analysis of human liver for demonstrating differences between young and old humans has not yet been performed. However, identifying major age-related alterations in hepatic gene expression may pinpoint ontogenetic shifts with important hepatic and systemic consequences, provide novel pharmacogenetic information, offer clues to efficiently counteract symptoms of old age, and improve the overarching understanding of individual decline. Next-generation sequencing (NGS) data analyzed by the Mann-Whitney nonparametric test and Ensemble Feature Selection (EFS) bioinformatics identified 44 transcripts among 60,617 total and 19,986 protein-encoding transcripts that significantly (p = 0.0003 to 0.0464) and strikingly (EFS score > 0.3:16 transcripts; EFS score > 0.2:28 transcripts) differ between young and old livers. Most of these age-related transcripts were assigned to the categories 'regulome', 'inflammaging', 'regeneration', and 'pharmacogenes'. NGS results were confirmed by quantitative real-time polymerase chain reaction. Our results have important implications for the areas of ontogeny/aging and the age-dependent increase in major liver diseases. Finally, we present a broadly substantiated and testable hypothesis on a genetically governed 'aging cascade', wherein PPP1R10 acts as a putative ontogenetic master regulator, prominently flanked by IGFALS and DUSP1. This transcriptome-wide analysis of human liver offers potential clues towards developing safer and improved therapeutic interventions against major liver diseases and increased insights into key mechanisms underlying aging.

Prediction of antimicrobial resistance based on whole-genome sequencing and machine learning

MOTIVATION

Antimicrobial resistance (AMR) is one of the biggest global problems threatening human and animal health. Rapid and accurate AMR diagnostic methods are thus very urgently needed. However, traditional antimicrobial susceptibility testing (AST) is time-consuming, low throughput and viable only for cultivable bacteria. Machine learning methods may pave the way for automated AMR prediction based on genomic data of the bacteria. However, comparing different machine learning methods for the prediction of AMR based on different encodings and whole-genome sequencing data without previously known knowledge remains to be done.

RESULTS

In this study, we evaluated logistic regression (LR), support vector machine (SVM), random forest (RF) and convolutional neural network (CNN) for the prediction of AMR for the antibiotics ciprofloxacin, cefotaxime, ceftazidime and gentamicin. We could demonstrate that these models can effectively predict AMR with label encoding, one-hot encoding and frequency matrix chaos game representation (FCGR encoding) on whole-genome sequencing data. We trained these models on a large AMR dataset and evaluated them on an independent public dataset. Generally, RFs and CNNs perform better than LR and SVM with AUCs up to 0.96. Furthermore, we were able to identify mutations that are associated with AMR for each antibiotic.

AVAILABILITY AND IMPLEMENTATION

Source code in data preparation and model training are provided at GitHub website (https://github.com/YunxiaoRen/ML-iAMR).

SUPPLEMENTARY INFORMATION

Supplementary data are available at Bioinformatics online.

Differentiation of autism spectrum disorder and mood or anxiety disorder

Autism spectrum disorder shares many symptoms with other mental health disorders, and comorbid disorders such as mood and anxiety disorders are common, making the diagnostic process challenging. We aimed to explore the diagnostic accuracy of two standard autism spectrum disorder diagnostic instruments and to identify those behavioral items that best differentiate between autism spectrum disorder and mood and anxiety disorder in a naturalistic sample of patients utilizing autism spectrum disorder specialist services. The study included data of 847 participants (5–65 years of age, n = 586 with autism spectrum disorder, n = 261 with mood and anxiety disorder) all evaluated with the Autism Diagnostic Observation Schedule in the context of the diagnostic process. Data of the Autism Diagnostic Interview–Revised were available for 428 participants (5–51 years of age, n = 367 with autism spectrum disorder, n = 61 with mood and anxiety disorder). By means of binominal logistic regressions and an ensemble feature selection, we identified a subset of items that best differentiated between autism spectrum disorder and mood and anxiety disorder. Overall, the results indicate that a combination of communicational deficits and unusual and/or inappropriate social overtures differentiates autism spectrum disorder and mood and anxiety disorder. Aspects of social cognition are also relevant. Limitations of the current study and implications for research and practice are discussed.

Metal-Organic Framework Hybrids Aid Metabolic Profiling for Colorectal Cancer

Colorectal cancer (CRC) is the third most common fatal cancer worldwide, accounting for ≈10% of cancer-related mortality. Metabolic shift occurs from the very early stage during the development of CRC, which is of significant etiological and diagnostic importance toward precision medicine. Here, an advanced molecular tool to characterize the metabolic alterations in CRC, based on metal-organic framework (MOF) hybrids is reported. Consuming only 500 nL of plasma without any sample pretreatment, MOF hybrids yield direct metabolic fingerprints by laser desorption/ionization mass spectrometry in seconds. A diagnostic prediction model by a machine learning algorithm is constructed, to discriminate CRC patients from normal controls with an average area under the curve of 0.947 for the discovery cohort and 0.912 for the independent validation cohort. In addition, CRC-specific metabolic signature consisting of 34 potential biomarkers, based on the aforementioned diagnostic model is identified. The results advance the design of nanomaterial-based platforms for metabolic analysis and establish a new liquid biopsy tool for CRC screening compatible with the current clinical workflow in practice.

Plasmonic Alloys Reveal a Distinct Metabolic Phenotype of Early Gastric Cancer

Gastric cancer (GC) is a multifactorial process, accompanied by alterations in metabolic pathways. Non-invasive metabolic profiling facilitates GC diagnosis at early stage leading to an improved prognostic outcome. Herein, mesoporous PdPtAu alloys are designed to characterize the metabolic profiles in human blood. The elemental composition is optimized with heterogeneous surface plasmonic resonance, offering preferred charge transfer for photoinduced desorption/ionization and enhanced photothermal conversion for thermally driven desorption. The surface structure of PdPtAu is further tuned with controlled mesopores, accommodating metabolites only, rather than large interfering compounds. Consequently, the optimized PdPtAu alloy yields direct metabolic fingerprints by laser desorption/ionization mass spectrometry in seconds, consuming 500 nL of native plasma. A distinct metabolic phenotype is revealed for early GC by sparse learning, resulting in precise GC diagnosis with an area under the curve of 0.942. It is envisioned that the plasmonic alloy will open up a new era of minimally invasive blood analysis to improve the surveillance of cancer patients in the clinical setting.

Development of a novel machine learning model to predict presence of nonalcoholic steatohepatitis

Objective

To develop a computer model to predict patients with nonalcoholic steatohepatitis (NASH) using machine learning (ML).

Materials and Methods

This retrospective study utilized two databases: a) the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) nonalcoholic fatty liver disease (NAFLD) adult database (2004-2009), and b) the Optum^® de-identified Electronic Health Record dataset (2007-2018), a real-world dataset representative of common electronic health records in the United States. We developed an ML model to predict NASH, using confirmed NASH and non-NASH based on liver histology results in the NIDDK dataset to train the model.

Results

Models were trained and tested on NIDDK NAFLD data (704 patients) and the best-performing models evaluated on Optum data (~3,000,000 patients). An eXtreme Gradient Boosting model (XGBoost) consisting of 14 features exhibited high performance as measured by area under the curve (0.82), sensitivity (81%), and precision (81%) in predicting NASH. Slightly reduced performance was observed with an abbreviated feature set of 5 variables (0.79, 80%, 80%, respectively). The full model demonstrated good performance (AUC 0.76) to predict NASH in Optum data.

Discussion

The proposed model, named NASHmap, is the first ML model developed with confirmed NASH and non-NASH cases as determined through liver biopsy and validated on a large, real-world patient dataset. Both the 14 and 5-feature versions exhibit high performance.

Conclusion

The NASHmap model is a convenient and high performing tool that could be used to identify patients likely to have NASH in clinical settings, allowing better patient management and optimal allocation of clinical resources.

Auto-HMM-LMF: feature selection based method for prediction of drug response via autoencoder and hidden Markov model

BACKGROUND

Predicting the response of cancer cell lines to specific drugs is an essential problem in personalized medicine. Since drug response is closely associated with genomic information in cancer cells, some large panels of several hundred human cancer cell lines are organized with genomic and pharmacogenomic data. Although several methods have been developed to predict the drug response, there are many challenges in achieving accurate predictions. This study proposes a novel feature selection-based method, named Auto-HMM-LMF, to predict cell line-drug associations accurately. Because of the vast dimensions of the feature space for predicting the drug response, Auto-HMM-LMF focuses on the feature selection issue for exploiting a subset of inputs with a significant contribution.

RESULTS

This research introduces a novel method for feature selection of mutation data based on signature assignments and hidden Markov models. Also, we use the autoencoder models for feature selection of gene expression and copy number variation data. After selecting features, the logistic matrix factorization model is applied to predict drug response values. Besides, by comparing to one of the most powerful feature selection methods, the ensemble feature selection method (EFS), we showed that the performance of the predictive model based on selected features introduced in this paper is much better for drug response prediction. Two datasets, the Genomics of Drug Sensitivity in Cancer (GDSC) and Cancer Cell Line Encyclopedia (CCLE) are used to indicate the efficiency of the proposed method across unseen patient cell-line. Evaluation of the proposed model showed that Auto-HMM-LMF could improve the accuracy of the results of the state-of-the-art algorithms, and it can find useful features for the logistic matrix factorization method.

CONCLUSIONS

We depicted an application of Auto-HMM-LMF in exploring the new candidate drugs for head and neck cancer that showed the proposed method is useful in drug repositioning and personalized medicine. The source code of Auto-HMM-LMF method is available in https://github.com/emdadi/Auto-HMM-LMF .

A combined strategy of feature selection and machine learning to identify predictors of prediabetes

OBJECTIVE

To identify predictors of prediabetes using feature selection and machine learning on a nationally representative sample of the US population.

MATERIALS AND METHODS

We analyzed n = 6346 men and women enrolled in the National Health and Nutrition Examination Survey 2013-2014. Prediabetes was defined using American Diabetes Association guidelines. The sample was randomly partitioned to training (n = 3174) and internal validation (n = 3172) sets. Feature selection algorithms were run on training data containing 156 preselected exposure variables. Four machine learning algorithms were applied on 46 exposure variables in original and resampled training datasets built using 4 resampling methods. Predictive models were tested on internal validation data (n = 3172) and external validation data (n = 3000) prepared from National Health and Nutrition Examination Survey 2011-2012. Model performance was evaluated using area under the receiver operating characteristic curve (AUROC). Predictors were assessed by odds ratios in logistic models and variable importance in others. The Centers for Disease Control (CDC) prediabetes screening tool was the benchmark to compare model performance.

RESULTS

Prediabetes prevalence was 23.43%. The CDC prediabetes screening tool produced 64.40% AUROC. Seven optimal (≥ 70% AUROC) models identified 25 predictors including 4 potentially novel associations; 20 by both logistic and other nonlinear/ensemble models and 5 solely by the latter. All optimal models outperformed the CDC prediabetes screening tool (P < 0.05).

DISCUSSION

Combined use of feature selection and machine learning increased predictive performance outperforming the recommended screening tool. A range of predictors of prediabetes was identified.

CONCLUSION

This work demonstrated the value of combining feature selection with machine learning to identify a wide range of predictors that could enhance prediabetes prediction and clinical decision-making.

Late first trimester circulating microparticle proteins predict the risk of preeclampsia < 35 weeks and suggest phenotypic differences among affected cases

We hypothesize that first trimester circulating micro particle (CMP) proteins will define preeclampsia risk while identifying clusters of disease subtypes among cases. We performed a nested case–control analysis among women with and without preeclampsia. Cases diagnosed < 34 weeks’ gestation were matched to controls. Plasma CMPs were isolated via size exclusion chromatography and analyzed using global proteome profiling based on HRAM mass spectrometry. Logistic models then determined feature selection with best performing models determined by cross-validation. K-means clustering examined cases for phenotypic subtypes and biological pathway enrichment was examined. Our results indicated that the proteins distinguishing cases from controls were enriched in biological pathways involved in blood coagulation, hemostasis and tissue repair. A panel consisting of C1RL, GP1BA, VTNC, and ZA2G demonstrated the best distinguishing performance (AUC of 0.79). Among the cases of preeclampsia, two phenotypic sub clusters distinguished cases; one enriched for platelet degranulation and blood coagulation pathways and the other for complement and immune response-associated pathways (corrected p < 0.001). Significantly, the second of the two clusters demonstrated lower gestational age at delivery (p = 0.049), increased protein excretion (p = 0.01), more extreme laboratory derangement (p < 0.0001) and marginally increased diastolic pressure (p = 0.09). We conclude that CMP-associated proteins at 12 weeks’ gestation predict the overall risk of developing early preeclampsia and indicate distinct subtypes of pathophysiology and clinical morbidity.

Mortality Risk Profiling of Staphylococcus aureus Bacteremia by Multi-omic Serum Analysis Reveals Early Predictive and Pathogenic Signatures

Staphylococcus aureus bacteremia (SaB) causes significant disease in humans, carrying mortality rates of ∼25%. The ability to rapidly predict SaB patient responses and guide personalized treatment regimens could reduce mortality. Here, we present a resource of SaB prognostic biomarkers. Integrating proteomic and metabolomic techniques enabled the identification of >10,000 features from >200 serum samples collected upon clinical presentation. We interrogated the complexity of serum using multiple computational strategies, which provided a comprehensive view of the early host response to infection. Our biomarkers exceed the predictive capabilities of those previously reported, particularly when used in combination. Last, we validated the biological contribution of mortality-associated pathways using a murine model of SaB. Our findings represent a starting point for the development of a prognostic test for identifying high-risk patients at a time early enough to trigger intensive monitoring and interventions.

binomialRF: interpretable combinatoric efficiency of random forests to identify biomarker interactions

BACKGROUND

In this era of data science-driven bioinformatics, machine learning research has focused on feature selection as users want more interpretation and post-hoc analyses for biomarker detection. However, when there are more features (i.e., transcripts) than samples (i.e., mice or human samples) in a study, it poses major statistical challenges in biomarker detection tasks as traditional statistical techniques are underpowered in high dimension. Second and third order interactions of these features pose a substantial combinatoric dimensional challenge. In computational biology, random forest (RF) classifiers are widely used due to their flexibility, powerful performance, their ability to rank features, and their robustness to the "P > > N" high-dimensional limitation that many matrix regression algorithms face. We propose binomialRF, a feature selection technique in RFs that provides an alternative interpretation for features using a correlated binomial distribution and scales efficiently to analyze multiway interactions.

RESULTS

In both simulations and validation studies using datasets from the TCGA and UCI repositories, binomialRF showed computational gains (up to 5 to 300 times faster) while maintaining competitive variable precision and recall in identifying biomarkers' main effects and interactions. In two clinical studies, the binomialRF algorithm prioritizes previously-published relevant pathological molecular mechanisms (features) with high classification precision and recall using features alone, as well as with their statistical interactions alone.

CONCLUSION

binomialRF extends upon previous methods for identifying interpretable features in RFs and brings them together under a correlated binomial distribution to create an efficient hypothesis testing algorithm that identifies biomarkers' main effects and interactions. Preliminary results in simulations demonstrate computational gains while retaining competitive model selection and classification accuracies. Future work will extend this framework to incorporate ontologies that provide pathway-level feature selection from gene expression input data.

Integration of an interpretable machine learning algorithm to identify early life risk factors of childhood obesity among preterm infants: a prospective birth cohort

BACKGROUND

The early life risk factors of childhood obesity among preterm infants are unclear and little is known about the influence of the feeding practices. We aimed to identify early life risk factors for childhood overweight/obesity among preterm infants and to determine feeding practices that could modify the identified risk factors.

METHODS

A total of 338,413 mother-child pairs were enrolled in the Jiaxing Birth Cohort (1999 to 2013), and 2125 eligible singleton preterm born children were included for analyses. We obtained data on health examination, anthropometric measurement, lifestyle, and dietary habits of each participant at their visits to clinics. An interpretable machine learning-based analytic framework was used to identify early life predictors for childhood overweight/obesity, and Poisson regression was used to examine the associations between feeding practices and the identified leading predictor.

RESULTS

Of the eligible 2125 preterm infants (863 [40.6%] girls), 274 (12.9%) developed overweight/obesity at age 4-7 years. We summarized early life variables into 25 features and identified two most important features as predictors for childhood overweight/obesity: trajectory of infant BMI (body mass index) Z-score change during the first year of corrected age and maternal BMI at enrollment. According to the impacts of different BMI Z-score trajectories on the outcome, we classified this feature into the favored and unfavored trajectories. Compared with early introduction of solid foods (≤ 3 months of corrected age), introducing solid foods after 6 months of corrected age was significantly associated with 11% lower risk (risk ratio, 0.89; 95% CI, 0.82 to 0.97) of being in the unfavored trajectory.

CONCLUSIONS

The trajectory of BMI Z-score change within the first year of life is the most important predictor for childhood overweight/obesity among preterm infants. Introducing solid foods after 6 months of corrected age is a recommended feeding practice for mitigating the risk of being in the unfavored trajectory.

Predicting and elucidating the etiology of fatty liver disease: A machine learning modeling and validation study in the IMI DIRECT cohorts

BACKGROUND

Non-alcoholic fatty liver disease (NAFLD) is highly prevalent and causes serious health complications in individuals with and without type 2 diabetes (T2D). Early diagnosis of NAFLD is important, as this can help prevent irreversible damage to the liver and, ultimately, hepatocellular carcinomas. We sought to expand etiological understanding and develop a diagnostic tool for NAFLD using machine learning.

METHODS AND FINDINGS

We utilized the baseline data from IMI DIRECT, a multicenter prospective cohort study of 3,029 European-ancestry adults recently diagnosed with T2D (n = 795) or at high risk of developing the disease (n = 2,234). Multi-omics (genetic, transcriptomic, proteomic, and metabolomic) and clinical (liver enzymes and other serological biomarkers, anthropometry, measures of beta-cell function, insulin sensitivity, and lifestyle) data comprised the key input variables. The models were trained on MRI-image-derived liver fat content (<5% or ≥5%) available for 1,514 participants. We applied LASSO (least absolute shrinkage and selection operator) to select features from the different layers of omics data and random forest analysis to develop the models. The prediction models included clinical and omics variables separately or in combination. A model including all omics and clinical variables yielded a cross-validated receiver operating characteristic area under the curve (ROCAUC) of 0.84 (95% CI 0.82, 0.86; p < 0.001), which compared with a ROCAUC of 0.82 (95% CI 0.81, 0.83; p < 0.001) for a model including 9 clinically accessible variables. The IMI DIRECT prediction models outperformed existing noninvasive NAFLD prediction tools. One limitation is that these analyses were performed in adults of European ancestry residing in northern Europe, and it is unknown how well these findings will translate to people of other ancestries and exposed to environmental risk factors that differ from those of the present cohort. Another key limitation of this study is that the prediction was done on a binary outcome of liver fat quantity (<5% or ≥5%) rather than a continuous one.

CONCLUSIONS

In this study, we developed several models with different combinations of clinical and omics data and identified biological features that appear to be associated with liver fat accumulation. In general, the clinical variables showed better prediction ability than the complex omics variables. However, the combination of omics and clinical variables yielded the highest accuracy. We have incorporated the developed clinical models into a web interface (see: https://www.predictliverfat.org/) and made it available to the community.

TRIAL REGISTRATION

ClinicalTrials.gov NCT03814915.

FCTP-WSRC: Protein-Protein Interactions Prediction via Weighted Sparse Representation Based Classification

The task of predicting protein–protein interactions (PPIs) has been essential in the context of understanding biological processes. This paper proposes a novel computational model namely FCTP-WSRC to predict PPIs effectively. Initially, combinations of the F-vector, composition (C) and transition (T) are used to map each protein sequence onto numeric feature vectors. Afterwards, an effective feature extraction method PCA (principal component analysis) is employed to reconstruct the most discriminative feature subspaces, which is subsequently used as input in weighted sparse representation based classification (WSRC) for prediction. The FCTP-WSRC model achieves accuracies of 96.67%, 99.82%, and 98.09% for H. pylori, Human and Yeast datasets respectively. Furthermore, the FCTP-WSRC model performs well when predicting three significant PPIs networks: the single-core network (CD9), the multiple-core network (Ras-Raf-Mek-Erk-Elk-Srf pathway), and the cross-connection network (Wnt-related Network). Consequently, the promising results show that the proposed method can be a powerful tool for PPIs prediction with excellent performance and less time.

Transcriptional analysis identifies potential biomarkers and molecular regulators in pneumonia and COPD exacerbation

Lower respiratory infections, such as community-acquired pneumonia (CAP), and chronic obstructive pulmonary disease (COPD) rank among the most frequent causes of death worldwide. Improved diagnostics and profound pathophysiological insights are urgent clinical needs. In our cohort, we analysed transcriptional networks of peripheral blood mononuclear cells (PBMCs) to identify central regulators and potential biomarkers. We investigated the mRNA- and miRNA-transcriptome of PBMCs of healthy subjects and patients suffering from CAP or AECOPD by microarray and Taqman Low Density Array. Genes that correlated with PBMC composition were eliminated, and remaining differentially expressed genes were grouped into modules. One selected module (120 genes) was particularly suitable to discriminate AECOPD and CAP and most notably contained a subset of five biologically relevant mRNAs that differentiated between CAP and AECOPD with an AUC of 86.1%. Likewise, we identified several microRNAs, e.g. miR-545-3p and miR-519c-3p, which separated AECOPD and CAP. We furthermore retrieved an integrated network of differentially regulated mRNAs and microRNAs and identified HNF4A, MCC and MUC1 as central network regulators or most important discriminatory markers. In summary, transcriptional analysis retrieved potential biomarkers and central molecular features of CAP and AECOPD.

Identification of infectious disease-associated host genes using machine learning techniques

Background

With the global spread of multidrug resistance in pathogenic microbes, infectious diseases emerge as a key public health concern of the recent time. Identification of host genes associated with infectious diseases will improve our understanding about the mechanisms behind their development and help to identify novel therapeutic targets.

Results

We developed a machine learning techniques-based classification approach to identify infectious disease-associated host genes by integrating sequence and protein interaction network features. Among different methods, Deep Neural Networks (DNN) model with 16 selected features for pseudo-amino acid composition (PAAC) and network properties achieved the highest accuracy of 86.33% with sensitivity of 85.61% and specificity of 86.57%. The DNN classifier also attained an accuracy of 83.33% on a blind dataset and a sensitivity of 83.1% on an independent dataset. Furthermore, to predict unknown infectious disease-associated host genes, we applied the proposed DNN model to all reviewed proteins from the database. Seventy-six out of 100 highly-predicted infectious disease-associated genes from our study were also found in experimentally-verified human-pathogen protein-protein interactions (PPIs). Finally, we validated the highly-predicted infectious disease-associated genes by disease and gene ontology enrichment analysis and found that many of them are shared by one or more of the other diseases, such as cancer, metabolic and immune related diseases.

Conclusions

To the best of our knowledge, this is the first computational method to identify infectious disease-associated host genes. The proposed method will help large-scale prediction of host genes associated with infectious-diseases. However, our results indicated that for small datasets, advanced DNN-based method does not offer significant advantage over the simpler supervised machine learning techniques, such as Support Vector Machine (SVM) or Random Forest (RF) for the prediction of infectious disease-associated host genes. Significant overlap of infectious disease with cancer and metabolic disease on disease and gene ontology enrichment analysis suggests that these diseases perturb the functions of the same cellular signaling pathways and may be treated by drugs that tend to reverse these perturbations. Moreover, identification of novel candidate genes associated with infectious diseases would help us to explain disease pathogenesis further and develop novel therapeutics.

Surface Proteome of Plasma Extracellular Vesicles as Biomarkers for Pneumonia and Acute Exacerbation of Chronic Obstructive Pulmonary Disease

Background

Community-acquired pneumonia (CAP) and acute exacerbation of chronic obstructive pulmonary disease (AECOPD) represent a major burden of disease and death and their differential diagnosis is critical. A potential source of relevant accessible biomarkers are blood-borne small extracellular vesicles (sEVs).

Methods

We performed an extracellular vesicle array to find proteins on plasma sEVs that are differentially expressed and possibly allow the differential diagnosis between CAP and AECOPD. Plasma samples were analyzed from 21 healthy controls, 24 patients with CAP, and 10 with AECOPD . The array contained 40 antibodies to capture sEVs, which were then visualized with a cocktail of biotin-conjugated CD9, CD63, and CD81 antibodies.

Results

We detected significant differences in the protein decoration of sEVs between healthy controls and patients with CAP or AECOPD. We found CD45 and CD28 to be the best discrimination markers between CAP and AECOPD in receiver operating characteristic analyses, with an area under the curve >0.92. Additional ensemble feature selection revealed the possibility to distinguish between CAP and AECOPD even if the patient with CAP had COPD, with a panel of CD45, CD28, CTLA4 (cytotoxic T-lymphocyte-associated protein 4), tumor necrosis factor–R-II, and CD16.

Conclusion

The discrimination of sEV-associated proteins is a minimally invasive method with potential to discriminate between CAP and AECOPD.

Reproducibility of biomarker identifications from mass spectrometry proteomic data in cancer studies

Reproducibility of disease signatures and clinical biomarkers in multi-omics disease analysis has been a key challenge due to a multitude of factors. The heterogeneity of the limited sample, various biological factors such as environmental confounders, and the inherent experimental and technical noises, compounded with the inadequacy of statistical tools, can lead to the misinterpretation of results, and subsequently very different biology. In this paper, we investigate the biomarker reproducibility issues, potentially caused by differences of statistical methods with varied distribution assumptions or marker selection criteria using Mass Spectrometry proteomic ovarian tumor data. We examine the relationship between effect sizes, p values, Cauchy p values, False Discovery Rate p values, and the rank fractions of identified proteins out of thousands in the limited heterogeneous sample. We compared the markers identified from statistical single features selection approaches with machine learning wrapper methods. The results reveal marked differences when selecting the protein markers from varied methods with potential selection biases and false discoveries, which may be due to the small effects, different distribution assumptions, and p value type criteria versus prediction accuracies. The alternative solutions and other related issues are discussed in supporting the reproducibility of findings for clinical actionable outcomes.

An Ensemble Voted Feature Selection Technique for Predictive Modeling of Malwares of Android

Each Android application requires accumulations of permissions in installation time and they are considered as the features which can be utilized in permission-based identification of Android malwares. Recently, ensemble feature selection techniques have received increasing attention over conventional techniques in different applications. In this work, a cluster based voted ensemble voted feature selection technique combining five base wrapper approaches of R libraries is projected for identifying most prominent set of features in the predictive modeling of Android malwares. The proposed method preserves both the desirable features of an ensemble feature selector, accuracy and diversity. Moreover, in this work, five different data partitioning ratios are considered and the impact of those ratios on predictive model are measured using coefficient of determination (r-square) and root mean square error. The proposed strategy has created significant better outcome in term of the number of selected features and classification accuracy.

Non-invasive assessment of NAFLD as systemic disease-A machine learning perspective

Background & aims

Current non-invasive scores for the assessment of severity of non-alcoholic fatty liver disease (NAFLD) and identification of patients with non-alcoholic steatohepatitis (NASH) have insufficient performance to be included in clinical routine. In the current study, we developed a novel machine learning approach to overcome the caveats of existing approaches.

Methods

Non-invasive parameters were selected by an ensemble feature selection (EFS) from a retrospectively collected training cohort of 164 obese individuals (age: 43.5±10.3y; BMI: 54.1±10.1kg/m²) to develop a model able to predict the histological assessed NAFLD activity score (NAS). The model was evaluated in an independent validation cohort (122 patients, age: 45.2±11.75y, BMI: 50.8±8.61kg/m²).

Results

EFS identified age, γGT, HbA1c, adiponectin, and M30 as being highly associated with NAFLD. The model reached a Spearman correlation coefficient with the NAS of 0.46 in the training cohort and was able to differentiate between NAFL (NAS≤4) and NASH (NAS>4) with an AUC of 0.73. In the independent validation cohort, an AUC of 0.7 was achieved for this separation. We further analyzed the potential of the new model for disease monitoring in an obese cohort of 38 patients under lifestyle intervention for one year. While all patients lost weight under intervention, increasing scores were observed in 15 patients. Increasing scores were associated with significantly lower absolute weight loss, lower reduction of waist circumference and basal metabolic rate.

Conclusions

A newly developed model (http://CHek.heiderlab.de) can predict presence or absence of NASH with reasonable performance. The new score could be used to detect NASH and monitor disease progression or therapy response to weight loss interventions.

A Data Mining Approach Identified Salivary Biomarkers That Discriminate between Two Obesity Measures

BACKGROUND

A key mechanism of obesity involves dysregulation of metabolic and inflammatory markers. This study aimed to identify salivary biomarkers and other factors associated with obesity using an ensemble data mining approach.

METHODS

For a random cohort of over 700 subjects from 8137 Kuwait children (10.00 ± 0.67 years), four data mining methods were applied to identify important variables associated with obesity, including logistic regression by lasso regularization (Lasso), multivariate adaptive regression spline (MARS), random forests (RF), and boosting classification trees (BT). Each algorithm generated a variable importance rank list, based on an internal cross-validation procedure. An aggregated importance ranking was constructed by averaging the rank ordering of variables from individual list, weighted by the classification performance of respective models. Subsequently, the subset of top-ranking variables that were identified with at least three algorithms was evaluated by classification performance using receiver operating characteristic (ROC) analysis with bootstrap percentile resampling.

RESULTS

Obesity was defined either by the waist circumference (OBW) or by the body mass index (BMI) (OBWHO). We identified C-reactive protein (CRP), insulin, leptin, adiponectin, as salivary biomarkers associated with OBW, plus a clinical feature fitness level. A similar set of biomarkers was identified for OBWHO, but not including leptin. Tree-based clustering analysis revealed patterns that were significantly different between the OBW and OBWHO subjects.

CONCLUSION

A data mining approach based on multiple algorithms is useful for identifying factors associated with phenotypes, especially in cases where relationships are not salient, and a consensus from multiple methods can help produce a more generalizable subset of features. In this case, we have demonstrated that evaluation using the waist circumference includes association with high levels of salivary leptin, which is not seen with evaluation by BMI.

SEDE-GPS: socio-economic data enrichment based on GPS information

BACKGROUND

Microbes are essentail components of all ecosystems because they drive many biochemical processes and act as primary producers. In freshwater ecosystems, the biodiversity in and the composition of microbial communities can be used as indicators for environmental quality. Recently, some environmental features have been identified that influence microbial ecosystems. However, the impact of human action on lake microbiomes is not well understood. This is, in part, due to the fact that environmental data is, albeit theoretically accessible, not easily available.

RESULTS

In this work, we present SEDE-GPS, a tool that gathers data that are relevant to the environment of an user-provided GPS coordinate. To this end, it accesses a list of public and corporate databases and aggregates the information in a single file, which can be used for further analysis. To showcase the use of SEDE-GPS, we enriched a lake microbial ecology sequencing dataset with around 18,000 socio-economic, climate, and geographic features. The sources of SEDE-GPS are public databases such as Eurostat, the Climate Data Center, and OpenStreetMap, as well as corporate sources such as Twitter. Using machine learning and feature selection methods, we were able to identify features in the data provided by SEDE-GPS that can be used to predict lake microbiome alpha diversity.

CONCLUSION

The results presented in this study show that SEDE-GPS is a handy and easy-to-use tool for comprehensive data enrichment for studies of ecology and other processes that are affected by environmental features. Furthermore, we present lists of environmental, socio-economic, and climate features that are predictive for microbial biodiversity in lake ecosystems. These lists indicate that human action has a major impact on lake microbiomes. SEDE-GPS and its source code is available for download at http://SEDE-GPS.heiderlab.de.

Phi-Delta-Diagrams: Software Implementation of a Visual Tool for Assessing Classifier and Feature Performance

In this article, a two-tiered 2D tool is described, called ⟨φ,δ⟩ diagrams, and this tool has been devised to support the assessment of classifiers in terms of accuracy and bias. In their standard versions, these diagrams provide information, as the underlying data were in fact balanced. Their generalization, i.e., ability to account for the imbalance, will be also briefly described. In either case, the isometrics of accuracy and bias are immediately evident therein, as—according to a specific design choice—they are in fact straight lines parallel to the x-axis and y-axis, respectively. ⟨φ,δ⟩ diagrams can also be used to assess the importance of features, as highly discriminant ones are immediately evident therein. In this paper, a comprehensive introduction on how to adopt ⟨φ,δ⟩ diagrams as a standard tool for classifier and feature assessment is given. In particular, with the goal of illustrating all relevant details from a pragmatic perspective, their implementation and usage as Python and R packages will be described.

Diabetes classification model based on boosting algorithms

BACKGROUND

Diabetes mellitus is a common and complicated chronic lifelong disease. Hence, it is of high clinical significance to find the most relevant clinical indexes and to perform efficient computer-aided pre-diagnoses and diagnoses.

RESULTS

Non-parametric statistical testing is performed on hundreds of medical measurement index results between diabetic and non-diabetic populations. Two common boosting algorithms, Adaboost.M1 and LogitBoost, are selected to establish a machine model for diabetes diagnosis based on these clinical test data, involving a total of 35,669 individuals. The machine classification models built by these two algorithms have very good classification ability. Here, the LogitBoost classification model is slightly better than the Adaboost.M1 classification model. The overall accuracy of the LogitBoost classification model reached 95.30% when using 10-fold cross validation. The true positive, true negative, false positive, and false negative rates of the binary classification model were 0.921, 0.969, 0.031, and 0.079, respectively, and the area under the receiver operating characteristic curve reached 0.99.

CONCLUSIONS

The boosting algorithms show excellent performance for the diabetes classification models based on clinical medical data. The coefficient matrix of the original data is a sparse matrix, because some of the test results were missing, including some that were directly related to disease diagnosis. Therefore, the model is robust and has a degree of pre-diagnosis function. In the process of selecting the preferred test items, the most statistically significant discriminating factors between the diabetic and general populations were obtained and can be used as reference risk factors for diabetes mellitus.

diceR: an R package for class discovery using an ensemble driven approach

BACKGROUND

Given a set of features, researchers are often interested in partitioning objects into homogeneous clusters. In health research, cancer research in particular, high-throughput data is collected with the aim of segmenting patients into sub-populations to aid in disease diagnosis, prognosis or response to therapy. Cluster analysis, a class of unsupervised learning techniques, is often used for class discovery. Cluster analysis suffers from some limitations, including the need to select up-front the algorithm to be used as well as the number of clusters to generate, in addition, there may exist several groupings consistent with the data, making it very difficult to validate a final solution. Ensemble clustering is a technique used to mitigate these limitations and facilitate the generalization and reproducibility of findings in new cohorts of patients.

RESULTS

We introduce diceR (diverse cluster ensemble in R), a software package available on CRAN: https://CRAN.R-project.org/package=diceR CONCLUSIONS: diceR is designed to provide a set of tools to guide researchers through a general cluster analysis process that relies on minimizing subjective decision-making. Although developed in a biological context, the tools in diceR are data-agnostic and thus can be applied in different contexts.

A fast approach to detect gene-gene synergy

Selecting informative genes, including individually discriminant genes and synergic genes, from expression data has been useful for medical diagnosis and prognosis. Detecting synergic genes is more difficult than selecting individually discriminant genes. Several efforts have recently been made to detect gene-gene synergies, such as dendrogram-based I(X₁; X₂; Y) (mutual information), doublets (gene pairs) and MIC(X₁; X₂; Y) based on the maximal information coefficient. It is unclear whether dendrogram-based I(X₁; X₂; Y) and doublets can capture synergies efficiently. Although MIC(X₁; X₂; Y) can capture a wide range of interaction, it has a high computational cost triggered by its 3-D search. In this paper, we developed a simple and fast approach based on abs conversion type (i.e. Z = |X₁ − X₂|) and t-test, to detect interactions in simulation and real-world datasets. Our results showed that dendrogram-based I(X₁; X₂; Y) and doublets are helpless for discovering pair-wise gene interactions, our approach can discover typical pair-wise synergic genes efficiently. These synergic genes can reach comparable accuracy to the individually discriminant genes using the same number of genes. Classifier cannot learn well if synergic genes have not been converted properly. Combining individually discriminant and synergic genes can improve the prediction performance.