Genomic structure predicts metabolite dynamics in microbial communities.

The metabolic activities of microbial communities play a defining role in the evolution and persistence of life on Earth, driving redox reactions that give rise to global biogeochemical cycles. Community metabolism emerges from a hierarchy of processes, including gene expression, ecological interactions, and environmental factors. In wild communities, gene content is correlated with environmental context, but predicting metabolite dynamics from genomes remains elusive. Here, we show, for the process of denitrification, that metabolite dynamics of a community are predictable from the genes each member of the community possesses. A simple linear regression reveals a sparse and generalizable mapping from gene content to metabolite dynamics for genomically diverse bacteria. A consumer-resource model correctly predicts community metabolite dynamics from single-strain phenotypes. Our results demonstrate that the conserved impacts of metabolic genes can predict community metabolite dynamics, enabling the prediction of metabolite dynamics from metagenomes, designing denitrifying communities, and discovering how genome evolution impacts metabolism.

Transcriptional and chromatin-based partitioning mechanisms uncouple protein scaling from cell size.

Biosynthesis scales with cell size such that protein concentrations generally remain constant as cells grow. As an exception, synthesis of the cell-cycle inhibitor Whi5 "sub-scales" with cell size so that its concentration is lower in larger cells to promote cell-cycle entry. Here, we find that transcriptional control uncouples Whi5 synthesis from cell size, and we identify histones as the major class of sub-scaling transcripts besides WHI5 by screening for similar genes. Histone synthesis is thereby matched to genome content rather than cell size. Such sub-scaling proteins are challenged by asymmetric cell division because proteins are typically partitioned in proportion to newborn cell volume. To avoid this fate, Whi5 uses chromatin-binding to partition similar protein amounts to each newborn cell regardless of cell size. Disrupting both Whi5 synthesis and chromatin-based partitioning weakens G1 size control. Thus, specific transcriptional and partitioning mechanisms determine protein sub-scaling to control cell size.

Megastudies improve the impact of applied behavioural science.

Policy-makers are increasingly turning to behavioural science for insights about how to improve citizens' decisions and outcomes1. Typically, different scientists test different intervention ideas in different samples using different outcomes over different time intervals2. The lack of comparability of such individual investigations limits their potential to inform policy. Here, to address this limitation and accelerate the pace of discovery, we introduce the megastudy-a massive field experiment in which the effects of many different interventions are compared in the same population on the same objectively measured outcome for the same duration. In a megastudy targeting physical exercise among 61,293 members of an American fitness chain, 30 scientists from 15 different US universities worked in small independent teams to design a total of 54 different four-week digital programmes (or interventions) encouraging exercise. We show that 45% of these interventions significantly increased weekly gym visits by 9% to 27%; the top-performing intervention offered microrewards for returning to the gym after a missed workout. Only 8% of interventions induced behaviour change that was significant and measurable after the four-week intervention. Conditioning on the 45% of interventions that increased exercise during the intervention, we detected carry-over effects that were proportionally similar to those measured in previous research3-6. Forecasts by impartial judges failed to predict which interventions would be most effective, underscoring the value of testing many ideas at once and, therefore, the potential for megastudies to improve the evidentiary value of behavioural science.

Correlates of protection against SARS-CoV-2 in rhesus macaques.

Recent studies have reported the protective efficacy of both natural1 and vaccine-induced2-7 immunity against challenge with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in rhesus macaques. However, the importance of humoral and cellular immunity for protection against infection with SARS-CoV-2 remains to be determined. Here we show that the adoptive transfer of purified IgG from convalescent rhesus macaques (Macaca mulatta) protects naive recipient macaques against challenge with SARS-CoV-2 in a dose-dependent fashion. Depletion of CD8+ T cells in convalescent macaques partially abrogated the protective efficacy of natural immunity against rechallenge with SARS-CoV-2, which suggests a role for cellular immunity in the context of waning or subprotective antibody titres. These data demonstrate that relatively low antibody titres are sufficient for protection against SARS-CoV-2 in rhesus macaques, and that cellular immune responses may contribute to protection if antibody responses are suboptimal. We also show that higher antibody titres are required for treatment of SARS-CoV-2 infection in macaques. These findings have implications for the development of SARS-CoV-2 vaccines and immune-based therapeutic agents.

150 million years of sustained increase in pterosaur flight efficiency.

The long-term accumulation of biodiversity has been punctuated by remarkable evolutionary transitions that allowed organisms to exploit new ecological opportunities. Mesozoic flying reptiles (the pterosaurs), which dominated the skies for more than 150 million years, were the product of one such transition. The ancestors of pterosaurs were small and probably bipedal early archosaurs1, which were certainly well-adapted to terrestrial locomotion. Pterosaurs diverged from dinosaur ancestors in the Early Triassic epoch (around 245 million years ago); however, the first fossils of pterosaurs are dated to 25 million years later, in the Late Triassic epoch. Therefore, in the absence of proto-pterosaur fossils, it is difficult to study how flight first evolved in this group. Here we describe the evolutionary dynamics of the adaptation of pterosaurs to a new method of locomotion. The earliest known pterosaurs took flight and subsequently appear to have become capable and efficient flyers. However, it seems clear that transitioning between forms of locomotion2,3-from terrestrial to volant-challenged early pterosaurs by imposing a high energetic burden, thus requiring flight to provide some offsetting fitness benefits. Using phylogenetic statistical methods and biophysical models combined with information from the fossil record, we detect an evolutionary signal of natural selection that acted to increase flight efficiency over millions of years. Our results show that there was still considerable room for improvement in terms of efficiency after the appearance of flight. However, in the Azhdarchoidea4, a clade that exhibits gigantism, we test the hypothesis that there was a decreased reliance on flight5-7 and find evidence for reduced selection on flight efficiency in this clade. Our approach offers a blueprint to objectively study functional and energetic changes through geological time at a more nuanced level than has previously been possible.

Data-driven learning of structure augments quantitative prediction of biological responses.

Multi-factor screenings are commonly used in diverse applications in medicine and bioengineering, including optimizing combination drug treatments and microbiome engineering. Despite the advances in high-throughput technologies, large-scale experiments typically remain prohibitively expensive. Here we introduce a machine learning platform, structure-augmented regression (SAR), that exploits the intrinsic structure of each biological system to learn a high-accuracy model with minimal data requirement. Under different environmental perturbations, each biological system exhibits a unique, structured phenotypic response. This structure can be learned based on limited data and once learned, can constrain subsequent quantitative predictions. We demonstrate that SAR requires significantly fewer data comparing to other existing machine-learning methods to achieve a high prediction accuracy, first on simulated data, then on experimental data of various systems and input dimensions. We then show how a learned structure can guide effective design of new experiments. Our approach has implications for predictive control of biological systems and an integration of machine learning prediction and experimental design.

A systematic analysis of regression models for protein engineering.

To optimize proteins for particular traits holds great promise for industrial and pharmaceutical purposes. Machine Learning is increasingly applied in this field to predict properties of proteins, thereby guiding the experimental optimization process. A natural question is: How much progress are we making with such predictions, and how important is the choice of regressor and representation? In this paper, we demonstrate that different assessment criteria for regressor performance can lead to dramatically different conclusions, depending on the choice of metric, and how one defines generalization. We highlight the fundamental issues of sample bias in typical regression scenarios and how this can lead to misleading conclusions about regressor performance. Finally, we make the case for the importance of calibrated uncertainty in this domain.

m6Aexpress-enet: Predicting the regulatory expression m6A sites by an enet-regularization negative binomial regression model.

As the most abundant mRNA modification, m6A controls and influences many aspects of mRNA metabolism including the mRNA stability and degradation. However, the role of specific m6A sites in regulating gene expression still remains unclear. In additional, the multicollinearity problem caused by the correlation of methylation level of multiple m6A sites in each gene could influence the prediction performance. To address the above challenges, we propose an elastic-net regularized negative binomial regression model (called m6Aexpress-enet) to predict which m6A site could potentially regulate its gene expression. Comprehensive evaluations on simulated datasets demonstrate that m6Aexpress-enet could achieve the top prediction performance. Applying m6Aexpress-enet on real MeRIP-seq data from human lymphoblastoid cell lines, we have uncovered the complex regulatory pattern of predicted m6A sites and their unique enrichment pathway of the constructed co-methylation modules. m6Aexpress-enet proves itself as a powerful tool to enable biologists to discover the mechanism of m6A regulatory gene expression. Furthermore, the source code and the step-by-step implementation of m6Aexpress-enet is freely accessed at https://github.com/tengzhangs/m6Aexpress-enet.

Model-free prediction test with application to genomics data.

Testing the significance of predictors in a regression model is one of the most important topics in statistics. This problem is especially difficult without any parametric assumptions on the data. This paper aims to test the null hypothesis that given confounding variables Z, X does not significantly contribute to the prediction of Y under the model-free setting, where X and Z are possibly high dimensional. We propose a general framework that first fits nonparametric machine learning regression algorithms on [Formula: see text] and [Formula: see text], then compares the prediction power of the two models. The proposed method allows us to leverage the strength of the most powerful regression algorithms developed in the modern machine learning community. The P value for the test can be easily obtained by permutation. In simulations, we find that the proposed method is more powerful compared to existing methods. The proposed method allows us to draw biologically meaningful conclusions from two gene expression data analyses without strong distributional assumptions: 1) testing the prediction power of sequencing RNA for the proteins in cellular indexing of transcriptomes and epitopes by sequencing data and 2) identification of spatially variable genes in spatially resolved transcriptomics data.

A practical problem with Egger regression in Mendelian randomization.

Mendelian randomization (MR) is an instrumental variable (IV) method using genetic variants such as single nucleotide polymorphisms (SNPs) as IVs to disentangle the causal relationship between an exposure and an outcome. Since any causal conclusion critically depends on the three valid IV assumptions, which will likely be violated in practice, MR methods robust to the IV assumptions are greatly needed. As such a method, Egger regression stands out as one of the most widely used due to its easy use and perceived robustness. Although Egger regression is claimed to be robust to directional pleiotropy under the instrument strength independent of direct effect (InSIDE) assumption, it is known to be dependent on the orientations/coding schemes of SNPs (i.e. which allele of an SNP is selected as the reference group). The current practice, as recommended as the default setting in some popular MR software packages, is to orientate the SNPs to be all positively associated with the exposure, which however, to our knowledge, has not been fully studied to assess its robustness and potential impact. We use both numerical examples (with both real data and simulated data) and analytical results to demonstrate the practical problem of Egger regression with respect to its heavy dependence on the SNP orientations. Under the assumption that InSIDE holds for some specific (and unknown) coding scheme of the SNPs, we analytically show that other coding schemes would in general lead to the violation of InSIDE. Other related MR and IV regression methods may suffer from the same problem. Cautions should be taken when applying Egger regression (and related MR and IV regression methods) in practice.

Differential network connectivity analysis for microbiome data adjusted for clinical covariates using jackknife pseudo-values.

BACKGROUND: A recent breakthrough in differential network (DN) analysis of microbiome data has been realized with the advent of next-generation sequencing technologies. The DN analysis disentangles the microbial co-abundance among taxa by comparing the network properties between two or more graphs under different biological conditions. However, the existing methods to the DN analysis for microbiome data do not adjust for other clinical differences between subjects. RESULTS: We propose a Statistical Approach via Pseudo-value Information and Estimation for Differential Network Analysis (SOHPIE-DNA) that incorporates additional covariates such as continuous age and categorical BMI. SOHPIE-DNA is a regression technique adopting jackknife pseudo-values that can be implemented readily for the analysis. We demonstrate through simulations that SOHPIE-DNA consistently reaches higher recall and F1-score, while maintaining similar precision and accuracy to existing methods (NetCoMi and MDiNE). Lastly, we apply SOHPIE-DNA on two real datasets from the American Gut Project and the Diet Exchange Study to showcase the utility. The analysis of the Diet Exchange Study is to showcase that SOHPIE-DNA can also be used to incorporate the temporal change of connectivity of taxa with the inclusion of additional covariates. As a result, our method has found taxa that are related to the prevention of intestinal inflammation and severity of fatigue in advanced metastatic cancer patients. CONCLUSION: SOHPIE-DNA is the first attempt of introducing the regression framework for the DN analysis in microbiome data. This enables the prediction of characteristics of a connectivity of a network with the presence of additional covariate information in the regression. The R package with a vignette of our methodology is available through the CRAN repository ( https://CRAN.R-project.org/package=SOHPIE ), named SOHPIE (pronounced as Sofie). The source code and user manual can be found at https://github.com/sjahnn/SOHPIE-DNA .

Climate dependence of the macrofaunal effect on litter decomposition-A global meta-regression analysis.

Litter decomposition by microorganisms and animals is influenced by climate and has been found to be higher in warm and wet than in cold and dry biomes. We, however, hypothesized that the macrofaunal effect on decomposition should increase with temperature and aridity since larger animals are more tolerant to aridity than smaller organisms. This hypothesis was supported by our global analysis of macrofauna exclusion studies. Macrofauna increased litter mass loss on average by 40%, twofold higher than the highest previous estimation of macrofaunal effect on decomposition. The strongest effect was found in subtropical deserts where faunal decomposition had not been considered important. Our results highlight the need to consider animal size when exploring climate dependence of faunal decomposition, and the disproportionately large role of macrofauna in regulating litter decomposition in warm drylands. This new realization is critical for understanding element cycling in the face of global warming and aridification.

Recent Methodological Trends in Epidemiology: No Need for Data-Driven Variable Selection?

Variable selection in regression models is a particularly important issue in epidemiology, where one usually encounters observational studies. In contrast to randomized trials or experiments, confounding is often not controlled by the study design, but has to be accounted for by suitable statistical methods. For instance, when risk factors should be identified with unconfounded effect estimates, multivariable regression techniques can help to adjust for confounders. We investigated the current practice of variable selection in 4 major epidemiologic journals in 2019 and found that the majority of articles used subject-matter knowledge to determine a priori the set of included variables. In comparison with previous reviews from 2008 and 2015, fewer articles applied data-driven variable selection. Furthermore, for most articles the main aim of analysis was hypothesis-driven effect estimation in rather low-dimensional data situations (i.e., large sample size compared with the number of variables). Based on our results, we discuss the role of data-driven variable selection in epidemiology.

Welch-weighted Egger regression reduces false positives due to correlated pleiotropy in Mendelian randomization.

Modern population-scale biobanks contain simultaneous measurements of many phenotypes, providing unprecedented opportunity to study the relationship between biomarkers and disease. However, inferring causal effects from observational data is notoriously challenging. Mendelian randomization (MR) has recently received increased attention as a class of methods for estimating causal effects using genetic associations. However, standard methods result in pervasive false positives when two traits share a heritable, unobserved common cause. This is the problem of correlated pleiotropy. Here, we introduce a flexible framework for simulating traits with a common genetic confounder that generalizes recently proposed models, as well as a simple approach we call Welch-weighted Egger regression (WWER) for estimating causal effects. We show in comprehensive simulations that our method substantially reduces false positives due to correlated pleiotropy while being fast enough to apply to hundreds of phenotypes. We apply our method first to a subset of the UK Biobank consisting of blood traits and inflammatory disease, and then to a broader set of 411 heritable phenotypes. We detect many effects with strong literature support, as well as numerous behavioral effects that appear to stem from physician advice given to people at high risk for disease. We conclude that WWER is a powerful tool for exploratory data analysis in ever-growing databases of genotypes and phenotypes.

Centralization of Pancreaticoduodenectomy: A Systematic Review and Spline Regression Analysis to Recommend Minimum Volume for a Specialist Pancreas Service.

OBJECTIVE: Through a systematic review and spline curve analysis, to better define the minimum volume threshold for hospitals to perform (pancreaticoduodenectomy) and the high-volume center. BACKGROUND: The pancreaticoduodenectomy (PD) is a resource-intensive procedure, with high morbidity and long hospital stays resulting in centralization towards high-volume hospitals; the published definition of high volume remains variable. MATERIALS AND METHODS: Following a systematic review of studies comparing PD outcomes across volume groups, semiparametric regression modeling of morbidity (%), mortality (%), length of stay (days), lymph node harvest (number of nodes), and cost ($USD) as continuous variables were performed and fitted as a smoothed function of splines. If this showed a nonlinear association, then a "zero-crossing" technique was used, which produced "first and second derivatives" to identify volume thresholds. RESULTS: Our analysis of 33 cohort studies (198,377 patients) showed 55 PDs/year and 43 PDs/year were the threshold value required to achieve the lowest morbidity and highest lymph node harvest, with model estimated df 5.154 ( P <0.001) and 8.254 ( P <0.001), respectively. The threshold value for mortality was ~45 PDs/year (model 9.219 ( P <0.001)), with the lowest mortality value (the optimum value) at ~70 PDs/year (ie, a high-volume center). No significant association was observed for cost ( edf =2, P =0.989) and length of stay ( edf =2.04, P =0.099). CONCLUSIONS: There is a significant benefit from the centralization of PD, with 55 PDs/year and 43 PDs/year as the threshold value required to achieve the lowest morbidity and highest lymph node harvest, respectively. To achieve mortality benefit, the minimum procedure threshold is 45 PDs/year, with the lowest and optimum mortality value (ie, a high-volume center) at approximately 70 PDs/year.

Neuropathic Pain After Burn Injury: A Severe and Common Problem in Recovery.

OBJECTIVE: The aim of this study was to address the limited understanding of neuropathic pain (NP) among burn survivors by comprehensively examining its prevalence and related factors on a national scale using the Burn Model System (BMS) National Database. BACKGROUND: NP is a common but underexplored complaint among burn survivors, greatly affecting their quality of life and functionality well beyond the initial injury. Existing data on NP and its consequences in burn survivors are limited to select single-institution studies, lacking a comprehensive national perspective. METHODS: The BMS National Database was queried to identify burn patients responding to NP-related questions at enrollment, 6 months, 12 months, 2 years, and 5 years postinjury. Descriptive statistics and regression analyses were used to explore associations between demographic/clinical characteristics and self-reported NP at different time points. RESULTS: There were 915 patients included for analysis. At discharge, 66.5% of patients experienced NP in their burn scars. Those with NP had significantly higher Patient-Reported Outcomes Measurement Information System 29 (PROMIS-29) pain inference, itch, anxiety, depression, and sleep disturbance scores and were less able to partake in social roles. Multiple logistic regression revealed male sex, % total body surface area, and moderate-to-severe pain as predictors of NP at 6 months. At 12 months, % total body surface area and moderate-to-severe pain remained significant predictors, while ethnicity and employment status emerged as significant predictors at 24 months. CONCLUSIONS: This study highlights the significant prevalence of NP in burn patients and its adverse impacts on their physical, psychological, and social well-being. The findings underscore the necessity of a comprehensive approach to NP treatment, addressing both physical symptoms and psychosocial factors.

Year 1 of the Bundled Payments for Care Improvement-Advanced Model.

BACKGROUND: The Center for Medicare and Medicaid Innovation launched the Medicare Bundled Payments for Care Improvement-Advanced (BPCI-A) program for hospitals in October 2018. Information is needed about the effects of the program on health care utilization and Medicare payments. METHODS: We conducted a modified segmented regression analysis using Medicare claims and including patients with discharge dates from January 2017 through September 2019 to assess differences between BPCI-A participants and two control groups: hospitals that never joined the BPCI-A program (nonjoining hospitals) and hospitals that joined the BPCI-A program in January 2020, after the conclusion of the intervention period (late-joining hospitals). The primary outcomes were the differences in changes in quarterly trends in 90-day per-episode Medicare payments and the percentage of patients with readmission within 90 days after discharge. Secondary outcomes were mortality, volume, and case mix. RESULTS: A total of 826 BPCI-A participant hospitals were compared with 2016 nonjoining hospitals and 334 late-joining hospitals. Among BPCI-A hospitals, the mean baseline 90-day per-episode Medicare payment was $27,315; the change in the quarterly trends in the intervention period as compared with baseline was -$78 per quarter. Among nonjoining hospitals, the mean baseline 90-day per-episode Medicare payment was $25,994; the change in quarterly trends as compared with baseline was -$26 per quarter (difference between nonjoining hospitals and BPCI-A hospitals, $52 [95% confidence interval {CI}, 34 to 70] per quarter; P<0.001; 0.2% of the baseline payment). Among late-joining hospitals, the mean baseline 90-day per-episode Medicare payment was $26,807; the change in the quarterly trends as compared with baseline was $4 per quarter (difference between late-joining hospitals and BPCI-A hospitals, $82 [95% CI, 41 to 122] per quarter; P<0.001; 0.3% of the baseline payment). There were no meaningful differences in the changes with regard to readmission, mortality, volume, or case mix. CONCLUSIONS: The BPCI-A program was associated with small reductions in Medicare payments among participating hospitals as compared with control hospitals. (Funded by the National Heart, Lung, and Blood Institute.).

Waning Immunity after the BNT162b2 Vaccine in Israel.

BACKGROUND: In December 2020, Israel began a mass vaccination campaign against coronavirus disease 2019 (Covid-19) by administering the BNT162b2 vaccine, which led to a sharp curtailing of the outbreak. After a period with almost no cases of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, a resurgent Covid-19 outbreak began in mid-June 2021. Possible reasons for the resurgence were reduced vaccine effectiveness against the delta (B.1.617.2) variant and waning immunity. The extent of waning immunity of the vaccine against the delta variant in Israel is unclear. METHODS: We used data on confirmed infection and severe disease collected from an Israeli national database for the period of July 11 to 31, 2021, for all Israeli residents who had been fully vaccinated before June 2021. We used a Poisson regression model to compare rates of confirmed SARS-CoV-2 infection and severe Covid-19 among persons vaccinated during different time periods, with stratification according to age group and with adjustment for possible confounding factors. RESULTS: Among persons 60 years of age or older, the rate of infection in the July 11-31 period was higher among persons who became fully vaccinated in January 2021 (when they were first eligible) than among those fully vaccinated 2 months later, in March (rate ratio, 1.6; 95% confidence interval [CI], 1.3 to 2.0). Among persons 40 to 59 years of age, the rate ratio for infection among those fully vaccinated in February (when they were first eligible), as compared with 2 months later, in April, was 1.7 (95% CI, 1.4 to 2.1). Among persons 16 to 39 years of age, the rate ratio for infection among those fully vaccinated in March (when they were first eligible), as compared with 2 months later, in May, was 1.6 (95% CI, 1.3 to 2.0). The rate ratio for severe disease among persons fully vaccinated in the month when they were first eligible, as compared with those fully vaccinated in March, was 1.8 (95% CI, 1.1 to 2.9) among persons 60 years of age or older and 2.2 (95% CI, 0.6 to 7.7) among those 40 to 59 years of age; owing to small numbers, the rate ratio could not be calculated among persons 16 to 39 years of age. CONCLUSIONS: These findings indicate that immunity against the delta variant of SARS-CoV-2 waned in all age groups a few months after receipt of the second dose of vaccine.

Treatment of Multisystem Inflammatory Syndrome in Children.

BACKGROUND: Evidence is urgently needed to support treatment decisions for children with multisystem inflammatory syndrome (MIS-C) associated with severe acute respiratory syndrome coronavirus 2. METHODS: We performed an international observational cohort study of clinical and outcome data regarding suspected MIS-C that had been uploaded by physicians onto a Web-based database. We used inverse-probability weighting and generalized linear models to evaluate intravenous immune globulin (IVIG) as a reference, as compared with IVIG plus glucocorticoids and glucocorticoids alone. There were two primary outcomes: the first was a composite of inotropic support or mechanical ventilation by day 2 or later or death; the second was a reduction in disease severity on an ordinal scale by day 2. Secondary outcomes included treatment escalation and the time until a reduction in organ failure and inflammation. RESULTS: Data were available regarding the course of treatment for 614 children from 32 countries from June 2020 through February 2021; 490 met the World Health Organization criteria for MIS-C. Of the 614 children with suspected MIS-C, 246 received primary treatment with IVIG alone, 208 with IVIG plus glucocorticoids, and 99 with glucocorticoids alone; 22 children received other treatment combinations, including biologic agents, and 39 received no immunomodulatory therapy. Receipt of inotropic or ventilatory support or death occurred in 56 patients who received IVIG plus glucocorticoids (adjusted odds ratio for the comparison with IVIG alone, 0.77; 95% confidence interval [CI], 0.33 to 1.82) and in 17 patients who received glucocorticoids alone (adjusted odds ratio, 0.54; 95% CI, 0.22 to 1.33). The adjusted odds ratios for a reduction in disease severity were similar in the two groups, as compared with IVIG alone (0.90 for IVIG plus glucocorticoids and 0.93 for glucocorticoids alone). The time until a reduction in disease severity was similar in the three groups. CONCLUSIONS: We found no evidence that recovery from MIS-C differed after primary treatment with IVIG alone, IVIG plus glucocorticoids, or glucocorticoids alone, although significant differences may emerge as more data accrue. (Funded by the European Union's Horizon 2020 Program and others; BATS ISRCTN number, ISRCTN69546370.).

Feature selection for support vector regression using a genetic algorithm.

Support vector regression (SVR) is particularly beneficial when the outcome and predictors are nonlinearly related. However, when many covariates are available, the method's flexibility can lead to overfitting and an overall loss in predictive accuracy. To overcome this drawback, we develop a feature selection method for SVR based on a genetic algorithm that iteratively searches across potential subsets of covariates to find those that yield the best performance according to a user-defined fitness function. We evaluate the performance of our feature selection method for SVR, comparing it to alternate methods including LASSO and random forest, in a simulation study. We find that our method yields higher predictive accuracy than SVR without feature selection. Our method outperforms LASSO when the relationship between covariates and outcome is nonlinear. Random forest performs equivalently to our method in some scenarios, but more poorly when covariates are correlated. We apply our method to predict donor kidney function 1 year after transplant using data from the United Network for Organ Sharing national registry.