A Machine Learning Application Based in Random Forest for Integrating Mass Spectrometry-Based Metabolomic Data: A Simple Screening Method for Patients With Zika Virus

Unraveling the Metabolic Alterations Induced by Zika Infection in Prostate Epithelial (PNT1a) and Adenocarcinoma (PC-3) Cell Lines

The outbreak of Zika virus infection in 2016 led to the identification of its presence in several types of biofluids, including semen. Later discoveries associated Zika infection with sexual transmission and persistent replication in cells of the male reproductive tract. Prostate epithelial and carcinoma cells are favorable to virus replication, with studies pointing to transcriptomics alterations of immune and inflammation genes upon persistence. However, metabolome alterations promoted by the Zika virus in prostate cells are unknown. Given its chronic effects and oncolytic potential, we aim to investigate the metabolic alterations induced by the Zika virus in prostate epithelial (PNT1a) and adenocarcinoma (PC-3) cells using an untargeted metabolomics approach and high-resolution mass spectrometry. PNT1a cells were viable up to 15 days post ZIKV infection, in contrast to its antiproliferative effect in the PC-3 cell lineage. Remarkable alterations in the PNT1a cell metabolism were observed upon infection, especially regarding glycerolipids, fatty acids, and acylcarnitines, which could be related to viral cellular resource exploitation, in addition to the over-time increase in oxidative stress metabolites associated with carcinogenesis. The upregulation of FA20:5 at 5 dpi in PC-3 cells corroborates the antiproliferative effect observed since this metabolite was previously reported to induce PC-3 cell death. Overall, Zika virus promotes extensive lipid alterations on both PNT1a and PC-3 cells, promoting different outcomes based on the cellular metabolic state.

Precision Medicine Approaches with Metabolomics and Artificial Intelligence

Recent technological innovations in the field of mass spectrometry have supported the use of metabolomics analysis for precision medicine. This growth has been allowed also by the application of algorithms to data analysis, including multivariate and machine learning methods, which are fundamental to managing large number of variables and samples. In the present review, we reported and discussed the application of artificial intelligence (AI) strategies for metabolomics data analysis. Particularly, we focused on widely used non-linear machine learning classifiers, such as ANN, random forest, and support vector machine (SVM) algorithms. A discussion of recent studies and research focused on disease classification, biomarker identification and early diagnosis is presented. Challenges in the implementation of metabolomics–AI systems, limitations thereof and recent tools were also discussed.

Temporal and Spatiotemporal Arboviruses Forecasting by Machine Learning: A Systematic Review

Arboviruses are a group of diseases that are transmitted by an arthropod vector. Since they are part of the Neglected Tropical Diseases that pose several public health challenges for countries around the world. The arboviruses' dynamics are governed by a combination of climatic, environmental, and human mobility factors. Arboviruses prediction models can be a support tool for decision-making by public health agents. In this study, we propose a systematic literature review to identify arboviruses prediction models, as well as models for their transmitter vector dynamics. To carry out this review, we searched reputable scientific bases such as IEE Xplore, PubMed, Science Direct, Springer Link, and Scopus. We search for studies published between the years 2015 and 2020, using a search string. A total of 429 articles were returned, however, after filtering by exclusion and inclusion criteria, 139 were included. Through this systematic review, it was possible to identify the challenges present in the construction of arboviruses prediction models, as well as the existing gap in the construction of spatiotemporal models.

Cohort profile: Colombian Cohort for the Early Prediction of Preterm Birth (COLPRET): early prediction of preterm birth based on personal medical history, clinical characteristics, vaginal microbiome, biophysical characteristics of the cervix and maternal serum biochemical markers

PURPOSE

Preterm birth (PTB) is a public health issue. Interventions to prolong the length of gestation have not achieved the expected results, as the selection of population at risk of PTB is still a challenge. Cervical length (CL) is the most accepted biomarker, however in the best scenario the CL identifies half of the patients. It is unlikely that a single measure identifies all pregnant women who will deliver before 37 weeks of gestation, considering the multiple pathways theory. We planned this cohort to study the link between the vaginal microbiome, the proteome, metabolome candidates, characteristics of the cervix and the PTB.

PARTICIPANTS

Pregnant women in the first trimester of a singleton pregnancy are invited to participate in the study. We are collecting biological samples, including vaginal fluid and blood from every patient, also performing ultrasound measurement that includes Consistency Cervical Index (CCI) and CL. The main outcome is the delivery of a neonate before 37 weeks of gestation.

FINDINGS TO DATE

We have recruited 244 pregnant women. They all have measurements of the CL and CCI. A vaginal sample for microbiome analysis has been collected in the 244 patients. Most of them agreed to blood collection, 216 (89%). By August 2021, 100 participants had already delivered. Eleven participants (11 %) had a spontaneous PTB.

FUTURE PLANS

A reference value chart for the first trimester CCI will be created. We will gather information regarding the feasibility, reproducibility and limitations of CCI. Proteomic and metabolomic analyses will be done to identify the best candidates, and we will validate their use as predictors. Finally, we plan to integrate clinical data, ultrasound measurements and biological profiles into an algorithm to obtain a multidimensional biomarker to identify the individual risk for PTB.

Identifying peripheral arterial disease in the elderly patients using machine-learning algorithms

BACKGROUND

Peripheral artery disease (PAD) is a common syndrome in elderly people. Recently, artificial intelligence (AI) algorithms, in particular machine-learning algorithms, have been increasingly used in disease diagnosis.

AIM

In this study, we designed an effective diagnostic model of PAD in the elderly patients using artificial intelligence.

METHODS

The study was performed with 539 participants, all over 80 years in age, who underwent the measurements of Doppler ultrasonography and ankle-brachial pressure index (ABI). Blood samples were collected. ABI and two machine-learning algorithms (MLAs)-logistic regression and a random forest (RF) model-were established to diagnose PAD. The sensitivity and specificity of the models were analyzed. An additional RF model was designed based on the most significant features of the original RF model and a prospective study was conducted to demonstrate its external validity.

RESULTS

Thirteen of the 28 features introduced to the MLAs differed significantly between PAD and non-PAD participants. The respective sensitivities and specificities of logistic regression, RF, and ABI were as follows: logistic regression (81.5%, 83.8%), RF (89.3%, 91.6%) and ABI (85.1%, 84.5%). In the prospective study, the newly designed RF model based on the most significant seven features exhibited an acceptable performance rate for the diagnosis of PAD with 100.0% sensitivity and 90.3% specificity.

CONCLUSIONS

An RF model was a more effective method than the logistic regression and ABI for the diagnosis of PAD in an elderly cohort.

Assessment of the Risk of Severe Dengue Using Intrahost Viral Population in Dengue Virus Serotype 2 Patients via Machine Learning

Dengue virus, a positive-sense single-stranded RNA virus, continuously threatens human health. Although several criteria for evaluation of severe dengue have been recently established, the ability to prognose the risk of severe outcomes for dengue patients remains limited. Mutant spectra of RNA viruses, including single nucleotide variants (SNVs) and defective virus genomes (DVGs), contribute to viral virulence and growth. Here, we determine the potency of intrahost viral population in dengue patients with primary infection that progresses into severe dengue. A total of 65 dengue virus serotype 2 infected patients in primary infection including 17 severe cases were enrolled. We utilized deep sequencing to directly define the frequency of SNVs and detection times of DVGs in sera of dengue patients and analyzed their associations with severe dengue. Among the detected SNVs and DVGs, the frequencies of 9 SNVs and the detection time of 1 DVG exhibited statistically significant differences between patients with dengue fever and those with severe dengue. By utilizing the detected frequencies/times of the selected SNVs/DVG as features, the machine learning model showed high average with a value of area under the receiver operating characteristic curve (AUROC, 0.966 ± 0.064). The elevation of the frequency of SNVs at E (nucleotide position 995 and 2216), NS2A (nucleotide position 4105), NS3 (nucleotide position 4536, 4606), and NS5 protein (nucleotide position 7643 and 10067) and the detection times of the selected DVG that had a deletion junction in the E protein region (nucleotide positions of the junction: between 969 and 1022) increased the possibility of dengue patients for severe dengue. In summary, we demonstrated the detected frequencies/times of SNVs/DVG in dengue patients associated with severe disease and successfully utilized them to discriminate severe patients using machine learning algorithm. The identified SNVs and DVGs that are associated with severe dengue will expand our understanding of intrahost viral population in dengue pathogenesis.

Untargeted Metabolomics Insights into Newborns with Congenital Zika Infection

Zika virus (ZIKV), an emerging virus belonging to the Flaviviridae family, causes severe neurological clinical complications and has been associated with Guillain-Barré syndrome, fetal abnormalities known collectively as congenital Zika syndrome, and microcephaly. Studies have shown that ZIKV infection can alter cellular metabolism, directly affecting neural development. Brain growth requires controlled cellular metabolism, which is essential for cell proliferation and maturation. However, little is known regarding the metabolic profile of ZIKV-infected newborns and possible associations related to microcephaly. Furthering the understanding surrounding underlying mechanisms is essential to developing personalized treatments for affected individuals. Thus, metabolomics, the study of the metabolites produced by or modified in an organism, constitutes a valuable approach in the study of complex diseases. Here, 26 serum samples from ZIKV-positive newborns with or without microcephaly, as well as controls, were analyzed using an untargeted metabolomics approach involving gas chromatography-mass spectrometry (GC-MS). Significant alterations in essential and non-essential amino acids, as well as carbohydrates (including aldohexoses, such as glucose or mannose) and their derivatives (urea and pyruvic acid), were observed in the metabolic profiles analyzed. Our results provide insight into relevant metabolic processes in patients with ZIKV and microcephaly.

Metabolomics applied in the study of emerging arboviruses caused by Aedes aegypti mosquitoes: A review

Arboviruses, such as chikungunya, dengue, yellow fever, and zika, caused by the bite of the Aedes aegypti mosquito, have been a frequent public health problem, with a high incidence of outbreaks in tropical and subtropical countries. These diseases are easily confused with a flu-like illness and present very similar symptoms, difficult to distinguish, and treat appropriately. The effects that these infections cause in the organism are fundamentally derived from complex metabolic processes. A prominent area of science that investigates the changes in the metabolism of complex organisms is the metabolomics. Metabolomics measures the metabolites produced or altered in biological organisms, through the use of robust analytical platforms, such as separation techniques hyphenated with mass spectrometry, combined with bioinformatics. This review article presents an overview of the basic concepts of metabolomics workflow and advances in this field, and compiles research articles that use this omic approach to study these arboviruses. In this context, the metabolomics is applied to search new therapies, understand the viral replication mechanisms, and access the host-virus interactions.

Does leukotriene F4 play a major role in the infection mechanism of Candida sp.?

Candidiasis is the most common fungal infection affecting hospitalized patients, especially immunocompromised and critical patients. Limitations regarding the assertive diagnosis of both Candidemia and Candidiasis not only impairs the introduction of effective treatments but also lays a heavy financial burden over the health system. Furthermore, it is still challenging to ascertain whether diagnostic methods are accurate and whether treatment is effective for patients with Candidemia. These constraints come from the uncertainty of the pathophysiological mechanism by which the pathogen establishes the opportunistic infection. Additionally, it is the reason why some patients present positive blood culture results, and others do not, and why it is very difficult during clinical routines to prove Candidemia or invasive candidiasis. Taking into account the current situation, this contribution proposes two markers that may help to understand the mechanisms of infection by the pathogen: Leukotriene F4 and 5,6-dihydroxy-eicosatetraenoic. These two lipids putatively modulate the host's immune response, and the initial data presented in this contribution suggest that these lipids allow the opportunistic infection to be installed. The study was carried out using an omics-based platform using direct-infusion high-resolution mass spectrometry and allied with bioinformatics tools to provide accurate and reliable results for biomarker candidates screening.

Metabolomics and Machine Learning Approaches Combined in Pursuit for More Accurate Paracoccidioidomycosis Diagnoses

Brazil and many other Latin American countries are areas of endemicity for different neglected diseases, and the fungal infection paracoccidioidomycosis (PCM) is one of them. Among the clinical manifestations, pneumopathy associated with skin and mucosal lesions is the most frequent. PCM definitive diagnosis depends on yeast microscopic visualization and immunological tests, but both present ambiguous results and difficulty in differentiating PCM from other fungal infections. This research has employed metabolomics analysis through high-resolution mass spectrometry to identify PCM biomarkers in serum samples in order to improve diagnosis for this debilitating disease. To upgrade the biomarker selection, machine learning approaches, using Random Forest classifiers, were combined with metabolomics data analysis. The proposed combination of these two analytical methods resulted in the identification of a set of 19 PCM biomarkers that show accuracy of 97.1%, specificity of 100%, and sensitivity of 94.1%. The obtained results are promising and present great potential to improve PCM definitive diagnosis and adequate pharmacological treatment, reducing the incidence of PCM sequelae and resulting in a better quality of life.IMPORTANCE Paracoccidioidomycosis (PCM) is a fungal infection typically found in Latin American countries, especially in Brazil. The identification of this disease is based on techniques that may fail sometimes. Intending to improve PCM detection in patient samples, this study used the combination of two of the newest technologies, artificial intelligence and metabolomics. This combination allowed PCM detection, independently of disease form, through identification of a set of molecules present in patients' blood. The great difference in this research was the ability to detect disease with better confidence than the routine methods employed today. Another important point is that among the molecules, it was possible to identify some indicators of contamination and other infection that might worsen patients' condition. Thus, the present work shows a great potential to improve PCM diagnosis and even disease management, considering the possibility to identify concomitant harmful factors.

MOTA: Network-Based Multi-Omic Data Integration for Biomarker Discovery

The recent advancement of omic technologies provides researchers with the possibility to search for disease-associated biomarkers at the system level. The integrative analysis of data from a large number of molecules involved at various layers of the biological system offers a great opportunity to rank disease biomarker candidates. In this paper, we propose MOTA, a network-based method that uses data acquired at multiple layers to rank candidate disease biomarkers. The networks constructed by MOTA allow users to investigate the biological significance of the top-ranked biomarker candidates. We evaluated the performance of MOTA in ranking disease-associated molecules from three sets of multi-omic data representing three cohorts of hepatocellular carcinoma (HCC) cases and controls with liver cirrhosis. The results demonstrate that MOTA allows the identification of more top-ranked metabolite biomarker candidates that are shared by two different cohorts compared to traditional statistical methods. Moreover, the mRNA candidates top-ranked by MOTA comprise more cancer driver genes compared to those ranked by traditional differential expression methods.

Reducing False-Positive Results in Newborn Screening Using Machine Learning

Newborn screening (NBS) for inborn metabolic disorders is a highly successful public health program that by design is accompanied by false-positive results. Here we trained a Random Forest machine learning classifier on screening data to improve prediction of true and false positives. Data included 39 metabolic analytes detected by tandem mass spectrometry and clinical variables such as gestational age and birth weight. Analytical performance was evaluated for a cohort of 2777 screen positives reported by the California NBS program, which consisted of 235 confirmed cases and 2542 false positives for one of four disorders: glutaric acidemia type 1 (GA-1), methylmalonic acidemia (MMA), ornithine transcarbamylase deficiency (OTCD), and very long-chain acyl-CoA dehydrogenase deficiency (VLCADD). Without changing the sensitivity to detect these disorders in screening, Random Forest-based analysis of all metabolites reduced the number of false positives for GA-1 by 89%, for MMA by 45%, for OTCD by 98%, and for VLCADD by 2%. All primary disease markers and previously reported analytes such as methionine for MMA and OTCD were among the top-ranked analytes. Random Forest's ability to classify GA-1 false positives was found similar to results obtained using Clinical Laboratory Integrated Reports (CLIR). We developed an online Random Forest tool for interpretive analysis of increasingly complex data from newborn screening.

Severity Index for Suspected Arbovirus (SISA): Machine learning for accurate prediction of hospitalization in subjects suspected of arboviral infection

BACKGROUND

Dengue, chikungunya, and Zika are arboviruses of major global health concern. Decisions regarding the clinical management of suspected arboviral infection are challenging in resource-limited settings, particularly when deciding on patient hospitalization. The objective of this study was to determine if hospitalization of individuals with suspected arboviral infections could be predicted using subject intake data.

METHODOLOGY/PRINCIPAL FINDINGS

Two prediction models were developed using data from a surveillance study in Machala, a city in southern coastal Ecuador with a high burden of arboviral infections. Data were obtained from subjects who presented at sentinel medical centers with suspected arboviral infection (November 2013 to September 2017). The first prediction model-called the Severity Index for Suspected Arbovirus (SISA)-used only demographic and symptom data. The second prediction model-called the Severity Index for Suspected Arbovirus with Laboratory (SISAL)-incorporated laboratory data. These models were selected by comparing the prediction ability of seven machine learning algorithms; the area under the receiver operating characteristic curve from the prediction of a test dataset was used to select the final algorithm for each model. After eliminating those with missing data, the SISA dataset had 534 subjects, and the SISAL dataset had 98 subjects. For SISA, the best prediction algorithm was the generalized boosting model, with an AUC of 0.91. For SISAL, the best prediction algorithm was the elastic net with an AUC of 0.94. A sensitivity analysis revealed that SISA and SISAL are not directly comparable to one another.

CONCLUSIONS/SIGNIFICANCE

Both SISA and SISAL were able to predict arbovirus hospitalization with a high degree of accuracy in our dataset. These algorithms will need to be tested and validated on new data from future patients. Machine learning is a powerful prediction tool and provides an excellent option for new management tools and clinical assessment of arboviral infection.

Combining Machine Learning and Metabolomics to Identify Weight Gain Biomarkers

Weight gain is a metabolic disorder that often culminates in the development of obesity and other comorbidities such as diabetes. Obesity is characterized by the development of a chronic, subclinical systemic inflammation, and is regarded as a remarkably important factor that contributes to the development of such comorbidities. Therefore, laboratory methods that allow the identification of subjects at higher risk for severe weight-associated morbidity are of utter importance, considering the health, and safety of populations. This contribution analyzed the plasma of 180 Brazilian individuals, equally divided into a eutrophic control group and case group, to assess the presence of biomarkers related to weight gain, aiming at characterizing the phenotype of this population. Samples were analyzed by mass spectrometry and most discriminant features were determined by a machine learning approach using Random Forest algorithm. Five biomarkers related to the pathogenesis and chronicity of inflammation in weight gain were identified. Two metabolites of arachidonic acid were upregulated in the case group, indicating the presence of inflammation, as well as two other molecules related to dysfunctions in the cycle of nitric oxide (NO) and increase in superoxide production. Finally, a fifth case group marker observed in this study may indicate the trigger for diabetes in overweight and obesity individuals. The use of mass spectrometry combined with machine learning analyses to prospect and characterize biomarkers associated with weight gain will pave the way for elucidating potential therapeutic and prognostic targets.