-
Statistics in Medicine Feb 2023This review condenses the knowledge on variable selection methods implemented in R and appropriate for datasets with grouped features. The focus is on regularized... (Review)
Review
This review condenses the knowledge on variable selection methods implemented in R and appropriate for datasets with grouped features. The focus is on regularized regressions identified through a systematic review of the literature, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. A total of 14 methods are discussed, most of which use penalty terms to perform group variable selection. Depending on how the methods account for the group structure, they can be classified into knowledge and data-driven approaches. The first encompass group-level and bi-level selection methods, while two-step approaches and collinearity-tolerant methods constitute the second category. The identified methods are briefly explained and their performance compared in a simulation study. This comparison demonstrated that group-level selection methods, such as the group minimax concave penalty, are superior to other methods in selecting relevant variable groups but are inferior in identifying important individual variables in scenarios where not all variables in the groups are predictive. This can be better achieved by bi-level selection methods such as group bridge. Two-step and collinearity-tolerant approaches such as elastic net and ordered homogeneity pursuit least absolute shrinkage and selection operator are inferior to knowledge-driven methods but provide results without requiring prior knowledge. Possible applications in proteomics are considered, leading to suggestions on which method to use depending on existing prior knowledge and research question.
Topics: Humans; Computer Simulation
PubMed: 36546512
DOI: 10.1002/sim.9620 -
Reproduction (Cambridge, England) Nov 2020Cow subfertility is a multi-factorial problem in many countries which is only starting to be unravelled. Molecular biology can provide a substantial source of insight...
Cow subfertility is a multi-factorial problem in many countries which is only starting to be unravelled. Molecular biology can provide a substantial source of insight into its causes and potential solutions, particularly through large scale, untargeted omics approaches. In this systematic review, we set out to compile, assess and integrate the latest proteomic and metabolomic research on cow reproduction, specifically that on the female reproductive tract and early embryo. We herein report a general improvement in technical standards throughout the temporal span examined; however, significant methodological limitations are also identified. We propose easily actionable avenues for ameliorating these shortcomings and enhancing the reach of this field. Text mining and pathway analysis corroborate the relevance of proteins and metabolites related to the triad oxidative stress-inflammation-disease on reproductive function. We envisage a breakthrough in cattle reproductive molecular research within the next few years as in vivo sample techniques are improved, omics analysis equipment becomes more affordable and widespread, and software tools for single- and multi-omics data processing are further developed. Additional investigation of the impact of local oxidative stress and inflammation on fertility, both at the local and systemic levels, is key towards realising the full potential of this field.
Topics: Animals; Cattle; Female; Fertility; Metabolome; Proteome; Reproduction
PubMed: 33030443
DOI: 10.1530/REP-20-0047 -
Frontiers in Endocrinology 2023Polycystic Ovarian Syndrome (PCOS) is the most common endocrinopathy in women of reproductive age and remains widely underdiagnosed leading to significant morbidity....
INTRODUCTION
Polycystic Ovarian Syndrome (PCOS) is the most common endocrinopathy in women of reproductive age and remains widely underdiagnosed leading to significant morbidity. Artificial intelligence (AI) and machine learning (ML) hold promise in improving diagnostics. Thus, we performed a systematic review of literature to identify the utility of AI/ML in the diagnosis or classification of PCOS.
METHODS
We applied a search strategy using the following databases MEDLINE, Embase, the Cochrane Central Register of Controlled Trials, the Web of Science, and the IEEE Xplore Digital Library using relevant keywords. Eligible studies were identified, and results were extracted for their synthesis from inception until January 1, 2022.
RESULTS
135 studies were screened and ultimately, 31 studies were included in this study. Data sources used by the AI/ML interventions included clinical data, electronic health records, and genetic and proteomic data. Ten studies (32%) employed standardized criteria (NIH, Rotterdam, or Revised International PCOS classification), while 17 (55%) used clinical information with/without imaging. The most common AI techniques employed were support vector machine (42% studies), K-nearest neighbor (26%), and regression models (23%) were the commonest AI/ML. Receiver operating curves (ROC) were employed to compare AI/ML with clinical diagnosis. Area under the ROC ranged from 73% to 100% (n=7 studies), diagnostic accuracy from 89% to 100% (n=4 studies), sensitivity from 41% to 100% (n=10 studies), specificity from 75% to 100% (n=10 studies), positive predictive value (PPV) from 68% to 95% (n=4 studies), and negative predictive value (NPV) from 94% to 99% (n=2 studies).
CONCLUSION
Artificial intelligence and machine learning provide a high diagnostic and classification performance in detecting PCOS, thereby providing an avenue for early diagnosis of this disorder. However, AI-based studies should use standardized PCOS diagnostic criteria to enhance the clinical applicability of AI/ML in PCOS and improve adherence to methodological and reporting guidelines for maximum diagnostic utility.
SYSTEMATIC REVIEW REGISTRATION
https://www.crd.york.ac.uk/prospero/, identifier CRD42022295287.
Topics: Female; Humans; Artificial Intelligence; Polycystic Ovary Syndrome; Proteomics; Machine Learning; Cluster Analysis
PubMed: 37790605
DOI: 10.3389/fendo.2023.1106625 -
Journal of Diabetes and Metabolic... Jun 2022Due to growing concerns about the obesity pandemic as a worldwide phenomenon, a global effort has been made for managing it and associated disorders. Accordingly,... (Review)
Review
PURPOSE
Due to growing concerns about the obesity pandemic as a worldwide phenomenon, a global effort has been made for managing it and associated disorders. Accordingly, metabolomics as a promising field of "OMICS" is presented for investigating different molecular pathways in obesity and related disorders through the evaluation of specific metabolites in both animal and human subjects. Herein, the aim of the present study as the first systematic review is to evaluate all available studies about different mechanisms and their biomarkers discovery using metabolomics approaches.
METHOD
The study was designed according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. Using a comprehensive search strategy we searched in databases including; Web of Science, PubMed, and Scopus using specific keywords. Based on predefined inclusion/exclusion criteria study selection has been conducted considering the type of studies, participant, and outcome measures. Quality assessment was done using CASP (Critical Appraisal Skills Programme) checklist followed by data extraction according to a predefined data extraction sheet.
RESULTS
Among the articles that resulted from electronic search, a total of 74 articles met our inclusion criteria. The most prevalent studied metabolites were amino acids and lipid derivatives and both targeted and non-targeted approaches were applied for metabolomics studies.
CONCLUSION
This systematic review summarized a wide range of studies regardless of the age, history, language, and type of the study. Further studies are needed to compare the application of emerging methods in the treatment of obesity and related disorders.
SUPPLEMENTARY INFORMATION
The online version contains supplementary material available at 10.1007/s40200-021-00917-w.
PubMed: 35673462
DOI: 10.1007/s40200-021-00917-w -
International Journal of Molecular... Apr 2022Mass spectrometry (MS)-based techniques can be a powerful tool to identify neuropsychiatric disorder biomarkers, improving prediction and diagnosis ability. Here, we... (Meta-Analysis)
Meta-Analysis Review
Mass spectrometry (MS)-based techniques can be a powerful tool to identify neuropsychiatric disorder biomarkers, improving prediction and diagnosis ability. Here, we evaluate the efficacy of MS proteomics applied to human peripheral fluids of schizophrenia (SCZ) patients to identify disease biomarkers and relevant networks of biological pathways. Following PRISMA guidelines, a search was performed for studies that used MS proteomics approaches to identify proteomic differences between SCZ patients and healthy control groups (PROSPERO database: CRD42021274183). Nineteen articles fulfilled the inclusion criteria, allowing the identification of 217 differentially expressed proteins. Gene ontology analysis identified lipid metabolism, complement and coagulation cascades, and immune response as the main enriched biological pathways. Meta-analysis results suggest the upregulation of FCN3 and downregulation of APO1, APOA2, APOC1, and APOC3 in SCZ patients. Despite the proven ability of MS proteomics to characterize SCZ, several confounding factors contribute to the heterogeneity of the findings. In the future, we encourage the scientific community to perform studies with more extensive sampling and validation cohorts, integrating omics with bioinformatics tools to provide additional comprehension of differentially expressed proteins. The produced information could harbor potential proteomic biomarkers of SCZ, contributing to individualized prognosis and stratification strategies, besides aiding in the differential diagnosis.
Topics: Biomarkers; Computational Biology; Humans; Mass Spectrometry; Proteomics; Schizophrenia
PubMed: 35563307
DOI: 10.3390/ijms23094917 -
Journal of Ethnopharmacology Aug 2014The rationale for using traditional Chinese medicine (TCM) is based on the experience that has been gained from its wide use over thousands of years. However, the... (Review)
Review
ETHNOPHARMACOLOGICAL RELEVANCE
The rationale for using traditional Chinese medicine (TCM) is based on the experience that has been gained from its wide use over thousands of years. However, the mechanisms of action of many TCM are still unclear. Proteomics, which mainly characterizes protein functions, protein-protein interactions, and protein modification in tissues or animals, can be used to investigate signaling pathway perturbations in cells or the whole body. Proteomics has improved the discovery process of effective TCM compounds, and has helped to elucidate their possible mechanisms of action. Therefore, a systematic review of the application of proteomics on TCM research is of great importance and necessity. This review strives to describe the literature on the application of proteomics to elucidate the mechanism of action of TCM on various diseases, and provide the essential discussion on the further utilization of proteomics data to accelerate TCM research.
MATERIALS AND METHODS
Literature survey was performed via electronic search on Pubmed with keywords 'Proteomics' and 'Traditional Chinese Medicine'. The papers written in English were acquired and analyzed in this review.
RESULTS
This review mainly summarizes the application of proteomics to investigate TCM remedies for neuronal disease, cancer, cardiovascular disease, diabetes, and immunology-related disease.
CONCLUSIONS
Researchers have applied proteomics to study the mechanism of action of TCM and made substantial progresses. Further studies are required to determine the protein targets of the active compounds, analyze the mechanism of actions in patients, compare the clinical effects with western medicine.
Topics: Animals; Drugs, Chinese Herbal; Ethnopharmacology; Humans; Medicine, Chinese Traditional; Proteomics
PubMed: 24862488
DOI: 10.1016/j.jep.2014.05.022 -
The Journal of Parasitology Jan 2022Cystic echinococcosis is a zoonotic disease caused by the larval stage of Echinococcus granulosus. This affliction is an endemic worldwide condition that represents a...
Cystic echinococcosis is a zoonotic disease caused by the larval stage of Echinococcus granulosus. This affliction is an endemic worldwide condition that represents a neglected parasitic disease with important socioeconomic repercussions. Proteomic characterization of larval and adult stages of E. granulosus, as well as the association between expression profiles and host interactions, is relevant for a better understanding of parasite biology, and eventually for drug design and vaccine development. This study aimed to develop a synthesis of the evidence available related to proteomics of E. granulosus. A systematic review was carried out to collect data concerning the proteomics of E. granulosus, without language or host restriction, published between 1980 and 2019. A systematic search was carried out in the Trip Database, BIREME-BVS, SciELO, Web of Science, PubMed, EMBASE, SCOPUS, EBSCO host, and LILACS, using MeSH terms, free words, and Boolean connectors, and adapting strategies to each source of information. Additionally, a manual cross-reference search was performed. Variables studied were the year of publication, geographic origin of the study, number of samples, hosts, parasitic organs, proteomic techniques, and parasite proteins verified. Nine-hundred and thirty-six related articles were identified: 17 fulfilled selection criteria, including slightly more than 188 samples. Most articles were published between 2014 and 2019 (64.7%) and were from Brazil and China (35.3% each). In reference to confirmed hosts in the primary articles, cattle (41.2%) and humans (23.5%) were the most frequently reported. Concerning proteomic techniques applied in the primary articles, LC-MS/MS was the most used (41.1%), and 890 proteins were reported by the primary articles. As the results of our search suggest, the information related to E. granulosus proteomics is scarce, heterogeneous, and scattered throughout several articles that include a diversity of tissues, samples, intermediate hosts, and proteomic techniques. Consequently, the level of evidence generated by our search is type 4.
Topics: Animals; Echinococcosis; Echinococcus granulosus; Helminth Proteins; Proteomics
PubMed: 35119469
DOI: 10.1645/20-86 -
Wiener Klinische Wochenschrift Dec 2016Osteoporosis is the most frequent bone metabolic disease. In order to improve early detection, prediction, prevention, diagnosis, and treatment of the disease, a new... (Meta-Analysis)
Meta-Analysis Review
BACKGROUND
Osteoporosis is the most frequent bone metabolic disease. In order to improve early detection, prediction, prevention, diagnosis, and treatment of the disease, a new model of P4 medicine (personalized, predictive, preventive, and participatory medicine) could be applied. The aim of this work was to systematically review the publications of four different types of "omics" studies related to osteoporosis, in order to discover novel predictive, preventive, diagnostic, and therapeutic targets for better management of the geriatric population.
METHODS
To systematically search the PubMed database, we created specific groups of criteria for four different types of "omics" information on osteoporosis: genomic, transcriptomic, proteomic, and metabolomic. We then analyzed the intersections between them in order to find correlations and common pathways or molecules with important roles in osteoporosis, and with a potential application in disease prediction, prevention, diagnosis, or treatment.
RESULTS
Altogether, 180 publications of "omics" studies in the field of osteoporosis were found and reviewed at first selection. After introducing the inclusion and exclusion criteria (the secondary selection), 46 papers were included in the systematic review.
CONCLUSIONS
The intersection of reviewed papers identified five genes (ESR1, IBSP, CTNNB1, SOX4, and IDUA) and processes like the Wnt pathway, JAK/STAT signaling, and ERK/MAPK, which should be further validated for their predictive, diagnostic, or other clinical value in osteoporosis. Such molecular insights will enable us to fit osteoporosis into the P4 strategy and could increase the effectiveness of disease prediction and prevention, with a decrease in morbidity in the geriatric population.
Topics: Aged; Aged, 80 and over; Female; Genetic Markers; Genetic Predisposition to Disease; Genetic Testing; Geriatric Assessment; Humans; Male; Osteoporosis; Precision Medicine; Prevalence; Risk Factors
PubMed: 27873024
DOI: 10.1007/s00508-016-1125-3 -
Cancer Medicine Jul 2022Salivary diagnostics and their utility as a nonaggressive approach for breast cancer diagnosis have been extensively studied in recent years. This meta-analysis assesses... (Meta-Analysis)
Meta-Analysis Review
BACKGROUND
Salivary diagnostics and their utility as a nonaggressive approach for breast cancer diagnosis have been extensively studied in recent years. This meta-analysis assesses the diagnostic value of salivary biomarkers in differentiating between patients with breast cancer and controls.
METHODS
We conducted a meta-analysis and systematic review of studies related to salivary diagnostics published in PubMed, EMBASE, Scopus, Ovid, Science Direct, Web of Science (WOS), and Google Scholar. The articles were chosen utilizing inclusion and exclusion criteria, as well as assessing their quality. Specificity and sensitivity, along with negative and positive likelihood ratios (NLR and PLR) and diagnostic odds ratio (DOR), were calculated based on random- or fixed-effects model. Area under the curve (AUC) and summary receiver-operating characteristic (SROC) were plotted and evaluated, and Fagan's Nomogram was evaluated for clinical utility.
RESULTS
Our systematic review and meta-analysis included 14 papers containing 121 study units with 8639 adult subjects (4149 breast cancer patients and 4490 controls without cancer). The pooled specificity and sensitivity were 0.727 (95% CI: 0.713-0.740) and 0.717 (95% CI: 0.703-0.730), respectively. The pooled NLR and PLR were 0.396 (95% CI: 0.364-0.432) and 2.597 (95% CI: 2.389-2.824), respectively. The pooled DOR was 7.837 (95% CI: 6.624-9.277), with the AUC equal to 0.801. The Fagan's nomogram showed post-test probabilities of 28% and 72% for negative and positive outcomes, respectively. We also conducted subgroup analyses to determine specificity, sensitivity, DOR, PLR, and NLR based on the mean age of patients (≤52 or >52 years old), saliva type (stimulated and unstimulated saliva), biomarker measurement method (mass spectrometry [MS] and non-MS measurement methods), sample size (≤55 or >55), biomarker type (proteomics, metabolomics, transcriptomics and proteomics, and reagent-free biophotonic), and nations.
CONCLUSION
Saliva, as a noninvasive biomarker, has the potential to accurately differentiate breast cancer patients from healthy controls.
Topics: Adult; Biomarkers; Breast Neoplasms; Female; Humans; Middle Aged; Odds Ratio; ROC Curve; Sensitivity and Specificity
PubMed: 35315584
DOI: 10.1002/cam4.4640 -
Human Reproduction (Oxford, England) Jan 2015Do any proteomic biomarkers previously identified for pre-eclampsia (PE) overlap with those identified in women with polycystic ovary syndrome (PCOS). (Comparative Study)
Comparative Study Review
STUDY QUESTION
Do any proteomic biomarkers previously identified for pre-eclampsia (PE) overlap with those identified in women with polycystic ovary syndrome (PCOS).
SUMMARY ANSWER
Five previously identified proteomic biomarkers were found to be common in women with PE and PCOS when compared with controls.
WHAT IS KNOWN ALREADY
Various studies have indicated an association between PCOS and PE; however, the pathophysiological mechanisms supporting this association are not known.
STUDY DESIGN, SIZE, DURATION
A systematic review and update of our PCOS proteomic biomarker database was performed, along with a parallel review of PE biomarkers. The study included papers from 1980 to December 2013.
PARTICIPANTS/MATERIALS, SETTING, METHODS
In all the studies analysed, there were a total of 1423 patients and controls. The number of proteomic biomarkers that were catalogued for PE was 192.
MAIN RESULTS AND THE ROLE OF CHANCE
Five proteomic biomarkers were shown to be differentially expressed in women with PE and PCOS when compared with controls: transferrin, fibrinogen α, β and γ chain variants, kininogen-1, annexin 2 and peroxiredoxin 2. In PE, the biomarkers were identified in serum, plasma and placenta and in PCOS, the biomarkers were identified in serum, follicular fluid, and ovarian and omental biopsies.
LIMITATIONS, REASONS FOR CAUTION
The techniques employed to detect proteomics have limited ability in identifying proteins that are of low abundance, some of which may have a diagnostic potential. The sample sizes and number of biomarkers identified from these studies do not exclude the risk of false positives, a limitation of all biomarker studies. The biomarkers common to PE and PCOS were identified from proteomic analyses of different tissues.
WIDER IMPLICATIONS OF THE FINDINGS
This data amalgamation of the proteomic studies in PE and in PCOS, for the first time, discovered a panel of five biomarkers for PE which are common to women with PCOS, including transferrin, fibrinogen α, β and γ chain variants, kininogen-1, annexin 2 and peroxiredoxin 2. If validated, these biomarkers could provide a useful framework for the knowledge infrastructure in this area. To accomplish this goal, a well co-ordinated multidisciplinary collaboration of clinicians, basic scientists and mathematicians is vital.
STUDY FUNDING/COMPETING INTERESTS
No financial support was obtained for this project. There are no conflicts of interest.
Topics: Biomarkers; Female; Humans; Polycystic Ovary Syndrome; Pre-Eclampsia; Pregnancy; Proteins; Proteomics; Retrospective Studies
PubMed: 25351721
DOI: 10.1093/humrep/deu268