RESUMO
Cross-validation is a common method for estimating the predictive performance of machine learning models. In a data-scarce regime, where one typically wishes to maximize the number of instances used for training the model, an approach called 'leave-one-out cross-validation' is often used. In this design, a separate model is built for predicting each data instance after training on all other instances. Since this results in a single test data point available per model trained, predictions are aggregated across the entire dataset to calculate common rank-based performance metrics such as the area under the receiver operating characteristic or precision-recall curves. In this work, we demonstrate that this approach creates a negative correlation between the average label of each training fold and the label of its corresponding test instance, a phenomenon that we term distributional bias. As machine learning models tend to regress to the mean of their training data, this distributional bias tends to negatively impact performance evaluation and hyperparameter optimization. We show that this effect generalizes to leave-P-out cross-validation and persists across a wide range of modeling and evaluation approaches, and that it can lead to a bias against stronger regularization. To address this, we propose a generalizable rebalanced cross-validation approach that corrects for distributional bias. We demonstrate that our approach improves cross-validation performance evaluation in synthetic simulations and in several published leave-one-out analyses.
RESUMO
Skin cancer mortality rates continue to rise, and survival analysis is increasingly needed to understand who is at risk and what interventions improve outcomes. However, current statistical methods are limited by inability to synthesize multiple data types, such as patient genetics, clinical history, demographics, and pathology and reveal significant multimodal relationships through predictive algorithms. Advances in computing power and data science enabled the rise of artificial intelligence (AI), which synthesizes vast amounts of data and applies algorithms that enable personalized diagnostic approaches. Here, we analyze AI methods used in skin cancer survival analysis, focusing on supervised learning, unsupervised learning, deep learning, and natural language processing. We illustrate strengths and weaknesses of these approaches with examples. Our PubMed search yielded 14 publications meeting inclusion criteria for this scoping review. Most publications focused on melanoma, particularly histopathologic interpretation with deep learning. Such concentration on a single type of skin cancer amid increasing focus on deep learning highlight growing areas for innovation; however, it also demonstrates opportunity for additional analysis that addresses other types of cutaneous malignancies and expands the scope of prognostication to combine both genetic, histopathologic, and clinical data. Moreover, researchers may leverage multiple AI methods for enhanced benefit in analyses. Expanding AI to this arena may enable improved survival analysis, targeted treatments, and outcomes.
RESUMO
Adverse pregnancy outcomes (APOs) affect a large proportion of pregnancies and represent an important cause of morbidity and mortality worldwide. Yet the pathophysiology of APOs is poorly understood, limiting our ability to prevent and treat these conditions. To search for genetic markers of maternal risk for four APOs, we performed multi-ancestry genome-wide association studies (GWAS) for pregnancy loss, gestational length, gestational diabetes, and preeclampsia. We clustered participants by their genetic ancestry and focused our analyses on three sub-cohorts with the largest sample sizes: European, African, and Admixed American. Association tests were carried out separately for each sub-cohort and then meta-analyzed together. Two novel loci were significantly associated with an increased risk of pregnancy loss: a cluster of SNPs located downstream of the TRMU gene (top SNP: rs142795512), and the SNP rs62021480 near RGMA. In the GWAS of gestational length we identified two new variants, rs2550487 and rs58548906 near WFDC1 and AC005052.1, respectively. Lastly, three new loci were significantly associated with gestational diabetes (top SNPs: rs72956265, rs10890563, rs79596863), located on or near ZBTB20, GUCY1A2, and RPL7P20, respectively. Fourteen loci previously correlated with preterm birth, gestational diabetes, and preeclampsia were found to be associated with these outcomes as well.
Assuntos
Diabetes Gestacional , Estudo de Associação Genômica Ampla , Polimorfismo de Nucleotídeo Único , Resultado da Gravidez , Humanos , Gravidez , Feminino , Resultado da Gravidez/genética , Diabetes Gestacional/genética , Adulto , Pré-Eclâmpsia/genética , Predisposição Genética para Doença , Paridade/genéticaRESUMO
Objective: Preterm birth (PTB) is a major determinant of neonatal mortality, morbidity, and childhood disability. In this article, we present a longitudinal analysis of the risk factors associated with PTB and how they have varied over the years: starting from 1968 when the CDC first started, reporting the natality data, up until 2021. Along with this article, we are also releasing an RShiny web application that will allow for easy consumption of this voluminous dataset by the research community. Further, we hope this tool can aid clinicians in the understanding and prevention of PTB. Materials and Methods: This study used the CDC Natality data from 1968 to 2021 to analyze trends in PTB outcomes across the lens of various features, including race, maternal age, education, and interval length between pregnancies. Our interactive RShiny web application, CDC NatView, allows users to explore interactions between maternal risk factors and maternal morbidity conditions and the aforementioned features. Results: Our study demonstrates how CDC data can be leveraged to conduct a longitudinal analysis of natality trends in the United States. Our key findings reveal an upward trend in late PTBs, which is concerning. Moreover, a significant disparity exists between African American and White populations in terms of PTB. These disparities persist in other areas, such as education, body-mass index, and access to prenatal care later in pregnancy. Discussion: Another notable finding is the increase in maternal age over time. Additionally, we confirm that short interpregnancy intervals (IPIs) are a risk factor for PTBs. To facilitate the exploration of pregnancy risk factors, infections, and maternal morbidity, we developed an open-source RShiny tool called CDC NatView. This software offers a user-friendly interface to interact with and visualize the CDC natality data, which constitutes an invaluable resource. Conclusion: In conclusion, our study has shed light on the rise of late PTBs and the persistent disparities in PTB rates between African American and White populations in the US. The increase in maternal age and the confirmation of a short IPI as a risk factor for PTB are noteworthy findings. Our open-source tool, CDC NatView, can be a valuable resource for further exploration of the CDC natality data to enhance our understanding of pregnancy risk factors and the interaction of PTB outcomes and maternal morbidities.
RESUMO
Preterm birth is a major cause of neonatal morbidity and mortality, but its etiology and risk factors are poorly understood. We undertook a scoping review to illustrate the breadth of risk factors for preterm birth that have been reported in the literature. We conducted a search in the PubMed database for articles published in the previous 5 years. We determined eligibility for this scoping review by screening titles and abstracts, followed by full-text review. We extracted odds ratios and other measures of association for each identified risk factor in the articles. A total of 2,509 unique articles were identified from the search, of which 314 were eligible for inclusion in our final analyses. We summarized risk factors and their relative impacts in the following categories: Activity, Psychological, Medical History, Toxicology, Genetics, and Vaginal Microbiome. Many risk factors for preterm birth have been reported. It is challenging to synthesize findings given the multitude of isolated risk factors that have been studied, inconsistent definitions of risk factors and outcomes, and use of different covariates in analyses. Novel methods of analyzing large datasets may promote a more comprehensive understanding of the etiology of preterm birth and ability to predict the outcome. KEY POINTS: · Preterm birth is difficult to predict.. · Preterm birth has many diverse risk factors.. · Holistic approaches may yield new insights..
RESUMO
Adverse pregnancy outcomes (APOs) are major risk factors for women's health during pregnancy and even in the years after pregnancy. Due to the heterogeneity of APOs, only few genetic associations have been identified. In this report, we conducted genome-wide association studies (GWASs) of 479 traits that are possibly related to APOs using a large and racially diverse study, Nulliparous Pregnancy Outcomes Study: Monitoring Mothers-to-Be (nuMoM2b). To display extensive results, we developed a web-based tool GnuMoM2b (https://gnumom2b.cumcobgyn.org/) for searching, visualizing, and sharing results from a GWAS of 479 pregnancy traits as well as phenome-wide association studies of more than 17 million single nucleotide polymorphisms. The genetic results from 3 ancestries (Europeans, Africans, and Admixed Americans) and meta-analyses are populated in GnuMoM2b. In conclusion, GnuMoM2b is a valuable resource for extraction of pregnancy-related genetic results and shows the potential to facilitate meaningful discoveries.
Assuntos
Estudo de Associação Genômica Ampla , Fenômica , Gravidez , Feminino , Humanos , Estudo de Associação Genômica Ampla/métodos , Fenótipo , Fatores de Risco , Polimorfismo de Nucleotídeo ÚnicoRESUMO
Adverse pregnancy outcomes (APOs) are major risk factors for women's health during pregnancy and even in the years after pregnancy. Due to the heterogeneity of APOs, only few genetic associations have been identified. In this report, we conducted genome-wide association studies (GWAS) of 479 traits that are possibly related to APOs using a large and racially diverse study, Nulliparous Pregnancy Outcomes Study: Monitoring Mothers-to-Be (nuMoM2b). To display the extensive results, we developed a web-based tool GnuMoM2b ( https://gnumom2b.cumcobgyn.org/ ) for searching, visualizing, and sharing results from GWAS of 479 pregnancy traits as well as phenome-wide association studies (PheWAS) of more than 17 million single nucleotide polymorphisms (SNPs). The genetic results from three ancestries (Europeans, Africans, and Admixed Americans) and meta-analyses are populated in GnuMoM2b. In conclusion, GnuMoM2b is a valuable resource for extraction of pregnancy-related genetic results and shows the potential to facilitate meaningful discoveries.
RESUMO
Preeclampsia and gestational hypertension are common pregnancy complications associated with adverse maternal and child outcomes. Current tools for prediction, prevention and treatment are limited. Here we tested the association of maternal DNA sequence variants with preeclampsia in 20,064 cases and 703,117 control individuals and with gestational hypertension in 11,027 cases and 412,788 control individuals across discovery and follow-up cohorts using multi-ancestry meta-analysis. Altogether, we identified 18 independent loci associated with preeclampsia/eclampsia and/or gestational hypertension, 12 of which are new (for example, MTHFR-CLCN6, WNT3A, NPR3, PGR and RGL3), including two loci (PLCE1 and FURIN) identified in the multitrait analysis. Identified loci highlight the role of natriuretic peptide signaling, angiogenesis, renal glomerular function, trophoblast development and immune dysregulation. We derived genome-wide polygenic risk scores that predicted preeclampsia/eclampsia and gestational hypertension in external cohorts, independent of clinical risk factors, and reclassified eligibility for low-dose aspirin to prevent preeclampsia. Collectively, these findings provide mechanistic insights into the hypertensive disorders of pregnancy and have the potential to advance pregnancy risk stratification.
Assuntos
Eclampsia , Hipertensão Induzida pela Gravidez , Hipertensão , Pré-Eclâmpsia , Gravidez , Feminino , Criança , Humanos , Hipertensão Induzida pela Gravidez/genética , Pré-Eclâmpsia/genética , Pré-Eclâmpsia/prevenção & controle , Aspirina , Fatores de RiscoRESUMO
Objective: Preeclampsia is one of the leading causes of maternal morbidity, with consequences during and after pregnancy. Because of its diverse clinical presentation, preeclampsia is an adverse pregnancy outcome that is uniquely challenging to predict and manage. In this paper, we developed machine learning models that predict the onset of preeclampsia with severe features or eclampsia at discrete time points in a nulliparous pregnant study cohort. Materials and Methods: The prospective study cohort to which we applied machine learning is the Nulliparous Pregnancy Outcomes Study: Monitoring Mothers-to-be (nuMoM2b) study, which contains information from eight clinical sites across the US. Maternal serum samples were collected for 1,857 individuals between the first and second trimesters. These patients with serum samples collected are selected as the final cohort. Results: Our prediction models achieved an AUROC of 0.72 (95% CI, 0.69-0.76), 0.75 (95% CI, 0.71-0.79), and 0.77 (95% CI, 0.74-0.80), respectively, for the three visits. Our initial models were biased toward non-Hispanic black participants with a high predictive equality ratio of 1.31. We corrected this bias and reduced this ratio to 1.14. The top features stress the importance of using several tests, particularly for biomarkers and ultrasound measurements. Placental analytes were strong predictors for screening for the early onset of preeclampsia with severe features in the first two trimesters. Conclusion: Experiments suggest that it is possible to create racial bias-free early screening models to predict the patients at risk of developing preeclampsia with severe features or eclampsia nulliparous pregnant study cohort.
RESUMO
Sequencing-based approaches for the analysis of microbial communities are susceptible to contamination, which could mask biological signals or generate artifactual ones. Methods for in silico decontamination using controls are routinely used, but do not make optimal use of information shared across samples and cannot handle taxa that only partially originate in contamination or leakage of biological material into controls. Here we present Source tracking for Contamination Removal in microBiomes (SCRuB), a probabilistic in silico decontamination method that incorporates shared information across multiple samples and controls to precisely identify and remove contamination. We validate the accuracy of SCRuB in multiple data-driven simulations and experiments, including induced contamination, and demonstrate that it outperforms state-of-the-art methods by an average of 15-20 times. We showcase the robustness of SCRuB across multiple ecosystems, data types and sequencing depths. Demonstrating its applicability to microbiome research, SCRuB facilitates improved predictions of host phenotypes, most notably the prediction of treatment response in melanoma patients using decontaminated tumor microbiome data.
Assuntos
Microbiota , Neoplasias , Humanos , Microbiota/genética , FenótipoRESUMO
Metacells are cell groupings derived from single-cell sequencing data that represent highly granular, distinct cell states. Here we present single-cell aggregation of cell states (SEACells), an algorithm for identifying metacells that overcome the sparsity of single-cell data while retaining heterogeneity obscured by traditional cell clustering. SEACells outperforms existing algorithms in identifying comprehensive, compact and well-separated metacells in both RNA and assay for transposase-accessible chromatin (ATAC) modalities across datasets with discrete cell types and continuous trajectories. We demonstrate the use of SEACells to improve gene-peak associations, compute ATAC gene scores and infer the activities of critical regulators during differentiation. Metacell-level analysis scales to large datasets and is particularly well suited for patient cohorts, where per-patient aggregation provides more robust units for data integration. We use our metacells to reveal expression dynamics and gradual reconfiguration of the chromatin landscape during hematopoietic differentiation and to uniquely identify CD4 T cell differentiation and activation states associated with disease onset and severity in a Coronavirus Disease 2019 (COVID-19) patient cohort.
Assuntos
Cromatina , Epigenômica , Humanos , Cromatina/genética , Cromatina/metabolismo , Genômica , Linfócitos T CD4-Positivos/metabolismo , Algoritmos , Análise de Célula ÚnicaRESUMO
The inference of genome rearrangement events has been extensively studied, as they play a major role in molecular evolution. However, probabilistic evolutionary models that explicitly imitate the evolutionary dynamics of such events, as well as methods to infer model parameters, are yet to be fully utilized. Here, we developed a probabilistic approach to infer genome rearrangement rate parameters using an Approximate Bayesian Computation (ABC) framework. We developed two genome rearrangement models, a basic model, which accounts for genomic changes in gene order, and a more sophisticated one which also accounts for changes in chromosome number. We characterized the ABC inference accuracy using simulations and applied our methodology to both prokaryotic and eukaryotic empirical datasets. Knowledge of genome-rearrangement rates can help elucidate their role in evolution as well as help simulate genomes with evolutionary dynamics that reflect empirical genomes.
Assuntos
Evolução Molecular , Genoma , Teorema de Bayes , Simulação por Computador , GenômicaRESUMO
Patterns of sequencing coverage along a bacterial genome-summarized by a peak-to-trough ratio (PTR)-have been shown to accurately reflect microbial growth rates, revealing a new facet of microbial dynamics and host-microbe interactions. Here, we introduce Compute PTR (CoPTR): a tool for computing PTRs from complete reference genomes and assemblies. Using simulations and data from growth experiments in simple and complex communities, we show that CoPTR is more accurate than the current state of the art while also providing more PTR estimates overall. We further develop a theory formalizing a biological interpretation for PTRs. Using a reference database of 2935 species, we applied CoPTR to a case-control study of 1304 metagenomic samples from 106 individuals with inflammatory bowel disease. We show that growth rates are personalized, are only loosely correlated with relative abundances, and are associated with disease status. We conclude by showing how PTRs can be combined with relative abundances and metabolomics to investigate their effect on the microbiome.
Assuntos
Metagenômica , Microbiota , Estudos de Casos e Controles , Genoma Bacteriano , Humanos , Metagenoma , Microbiota/genéticaRESUMO
Summary: MiSDEED (Microbial Synthetic Data Engine for Experimental Design) is a command-line tool for generating synthetic longitudinal multinode data from simulated microbial environments. It generates relative-abundance timecourses under perturbations for an arbitrary number of time points, samples, locations and data types. All simulation parameters are exposed to the user to facilitate rapid power analysis and aid in study design. Users who want additional flexibility may also use MiSDEED as a Python package. Availability and implementation: MiSDEED is written in Python and is freely available at https://github.com/pchlenski/misdeed.
RESUMO
The gut microbiome is spatially heterogeneous, with environmental niches contributing to the distribution and composition of microbial populations. A recently developed mapping technology, MaPS-seq, aims to characterize the spatial organization of the gut microbiome by providing data about local microbial populations. However, information about the global arrangement of these populations is lost by MaPS-seq. To address this, we propose a class of Gaussian mixture models (GMM) with spatial dependencies between mixture components in order to computationally recover the relative spatial arrangement of microbial communities. We demonstrate on synthetic data that our spatial models can identify global spatial dynamics, accurately cluster data, and improve parameter inference over a naive GMM. We applied our model to three MaPS-seq data sets taken from various regions of the mouse intestine. On cecal and distal colon data sets, we find our model accurately recapitulates known spatial behaviors of the gut microbiome, including compositional differences between mucus and lumen-associated populations. Our model also seems to capture the role of a pH gradient on microbial populations in the mouse ileum and proposes new behaviors as well. IMPORTANCE The spatial arrangement of the microbes in the gut microbiome is a defining characteristic of its behavior. Various experimental studies have attempted to provide glimpses into the mechanisms that contribute to microbial arrangements. However, many of these descriptions are qualitative. We developed a computational method that takes microbial spatial data and learns many of the experimentally validated spatial factors. We can then use our model to propose previously unknown spatial behaviors. Our results demonstrate that the gut microbiome, while exceptionally large, has predictable spatial patterns that can be used to help us understand its role in health and disease.
RESUMO
The identification of rare variants associated with schizophrenia has proven challenging due to genetic heterogeneity, which is reduced in founder populations. In samples from the Ashkenazi Jewish population, we report that schizophrenia cases had a greater frequency of novel missense or loss of function (MisLoF) ultra-rare variants (URVs) compared to controls, and the MisLoF URV burden was inversely correlated with polygenic risk scores in cases. Characterizing 141 "case-only" genes (MisLoF URVs in ≥3 cases with none in controls), the cadherin gene set was associated with schizophrenia. We report a recurrent case mutation in PCDHA3 that results in the formation of cytoplasmic aggregates and failure to engage in homophilic interactions on the plasma membrane in cultured cells. Modeling purifying selection, we demonstrate that deleterious URVs are greatly overrepresented in the Ashkenazi population, yielding enhanced power for association studies. Identification of the cadherin/protocadherin family as risk genes helps specify the synaptic abnormalities central to schizophrenia.
Assuntos
Caderinas/genética , Predisposição Genética para Doença/genética , Esquizofrenia/genética , Éxons/genética , Feminino , Efeito Fundador , Humanos , Judeus/genética , Masculino , MutaçãoRESUMO
Microbial communities are found across diverse environments, including within and across the human body. As many microbes are unculturable in the lab, much of what is known about a microbiome-a collection of bacteria, fungi, archaea, and viruses inhabiting an environment--is from the sequencing of DNA from within the constituent community. Here, we provide an introduction to whole-metagenome shotgun sequencing studies, a ubiquitous approach for characterizing microbial communities, by reviewing three major research areas in metagenomics: assembly, community profiling, and functional profiling. Though not exhaustive, these areas encompass a large component of the metagenomics literature. We discuss each area in depth, the challenges posed by whole-metagenome shotgun sequencing, and approaches fundamental to the solutions of each. We conclude by discussing promising areas for future research. Though our emphasis is on the human microbiome, the methods discussed are broadly applicable across study systems.