Differential distribution improves gene selection stability and has competitive classification performance for patient survival.
Ontology highlight
ABSTRACT:
SUBMITTER: Strbenac D
PROVIDER: S-EPMC5291264 | biostudies-literature | 2016 Jul
REPOSITORIES: biostudies-literature
Publications
Differential distribution improves gene selection stability and has competitive classification performance for patient survival.
Strbenac Dario D Mann Graham J GJ Yang Jean Y H JY Ormerod John T JT
Nucleic acids research 20160517 13
A consistent difference in average expression level, often referred to as differential expression (DE), has long been used to identify genes useful for classification. However, recent cancer studies have shown that when transcription factors or epigenetic signals become deregulated, a change in expression variability (DV) of target genes is frequently observed. This suggests that assessing the importance of genes by either differential expression or variability alone potentially misses sets of i ...[more]
Similar Datasets
Project description:BACKGROUND: Michiels et al. (Lancet 2005; 365: 488-92) employed a resampling strategy to show that the genes identified as predictors of prognosis from resamplings of a single gene expression dataset are highly variable. The genes most frequently identified in the separate resamplings were put forward as a 'gold standard'. On a higher level, breast cancer datasets collected by different institutions can be considered as resamplings from the underlying breast cancer population. The limited overlap between published prognostic signatures confirms the trend of signature instability identified by the resampling strategy. Six breast cancer datasets, totaling 947 samples, all measured on the Affymetrix platform, are currently available. This provides a unique opportunity to employ a substantial dataset to investigate the effects of pooling datasets on classifier accuracy, signature stability and enrichment of functional categories. RESULTS: We show that the resampling strategy produces a suboptimal ranking of genes, which can not be considered to be a 'gold standard'. When pooling breast cancer datasets, we observed a synergetic effect on the classification performance in 73% of the cases. We also observe a significant positive correlation between the number of datasets that is pooled, the validation performance, the number of genes selected, and the enrichment of specific functional categories. In addition, we have evaluated the support for five explanations that have been postulated for the limited overlap of signatures. CONCLUSION: The limited overlap of current signature genes can be attributed to small sample size. Pooling datasets results in more accurate classification and a convergence of signature genes. We therefore advocate the analysis of new data within the context of a compendium, rather than analysis in isolation.
| S-EPMC2527336 | biostudies-literature
Project description:BackgroundAdvanced gastro-oesophageal cancer (GEA) treatment has been improved by the introduction of immune checkpoint inhibitors (CPIs), yet identifying predictive biomarkers remains a priority, particularly in patients with a combined positive score (CPS) < 5, where the benefit is less clear. Our study assesses certain immune microenvironment features related to sensitivity or resistance to CPIs with the aim of implementing a personalised approach across CPS < 5 GEA.DesignThrough integrative transcriptomic and clinicopathological analyses, we studied in both a retrospective and a prospective cohort, the immune tumour microenvironment features. We analysed the cell types composing the immune infiltrate highlighting their functional activity.ResultsThis integrative study allowed the identification of four different groups across our patients. Among them, we identified a cluster whose tumours expressed the most gene signatures related to immunomodulatory pathways and immunotherapy response. These tumours presented an enriched immune infiltrate showing high immune function activity that could potentially achieve the best benefit from CPIs. Finally, our findings were proven in an external CPI-exposed population, where the use of our transcriptomic results combined with CPS helped better identify those patients who could benefit from immunotherapy than using CPS alone (p = 0.043).ConclusionsThis transcriptomic classification could improve precision immunotherapy for GEA.
| S-EPMC9727124 | biostudies-literature
Project description:Light perception in birds is composed of the retina and extraretinal sites, located in the brain. Previous studies indicate that selective photostimulation of the eye decreased reproductive performance, whereas extraretinal photostimulation increases it. Differential photostimulation of the retina and extraretinal sites is based on the retina's sensitivity to green wavelengths and on the red wavelengths' ability to penetrate body tissues. We previously found that short-day exposure to green light within a long-day exposure to red light increases reproductive activity in female turkeys and broiler breeder hens. Furthermore, in a study conducted recently in our laboratory, we found that blue light repressed expression of green light receptor in the retina, which can further enhance reproduction activity in broiler breeders. Here, we examined the "brain activate/eye deactivate" hypothesis on gonadal axis activity and reproductive performance in a broiler breeder flock. Broiler breeder hens and roosters (ROSS 308) were divided into 5 light-treatment groups (controlled rooms with light-emitting diodes [LED] lamps): warm white (control), long-day (14 h) red (630 nm) and short-day (6 h) green (514 nm) (red-green), long-day green and short-day red (green-red), long-day red and short-day blue (456 nm) (red-blue), and long-day blue and short-day red (blue-red). Birds were reared from 20 to 55 wk of age. Eggs were collected daily. Weekly egg production calculated. All eggs were incubated for fertility and hatchability examination. Blood was drawn monthly for plasma analysis. At 35 wk of age (after peak production) and 55 wk of age (end of the experiment), 10 hens from each treatment group were euthanized, and selected tissues and glands were taken for gene expression trials. Providing long-day red light to extraretinal photoreceptors while maintaining retinal photoreceptors on short day with blue or green light significantly improved reproductive activities, manifested by elevated egg production and gonadal axis activity compared with Controls and primary breeder recommendations. Long-day green light reduced reproductive performances. We suggest that targeted photostimulation enhances reproductive and gonadal axis activities in broiler breeders.
| S-EPMC7597697 | biostudies-literature
Project description:BackgroundPredicting species' potential geographical range by species distribution models (SDMs) is central to understand their ecological requirements. However, the effects of using different modeling techniques need further investigation. In order to improve the prediction effect, we need to assess the predictive performance and stability of different SDMs.MethodologyWe collected the distribution data of five common tree species (Pinus massoniana, Betula platyphylla, Quercus wutaishanica, Quercus mongolica and Quercus variabilis) and simulated their potential distribution area using 13 environmental variables and six widely used SDMs: BIOCLIM, DOMAIN, MAHAL, RF, MAXENT, and SVM. Each model run was repeated 100 times (trials). We compared the predictive performance by testing the consistency between observations and simulated distributions and assessed the stability by the standard deviation, coefficient of variation, and the 99% confidence interval of Kappa and AUC values.ResultsThe mean values of AUC and Kappa from MAHAL, RF, MAXENT, and SVM trials were similar and significantly higher than those from BIOCLIM and DOMAIN trials (p<0.05), while the associated standard deviations and coefficients of variation were larger for BIOCLIM and DOMAIN trials (p<0.05), and the 99% confidence intervals for AUC and Kappa values were narrower for MAHAL, RF, MAXENT, and SVM. Compared to BIOCLIM and DOMAIN, other SDMs (MAHAL, RF, MAXENT, and SVM) had higher prediction accuracy, smaller confidence intervals, and were more stable and less affected by the random variable (randomly selected pseudo-absence points).ConclusionsAccording to the prediction performance and stability of SDMs, we can divide these six SDMs into two categories: a high performance and stability group including MAHAL, RF, MAXENT, and SVM, and a low performance and stability group consisting of BIOCLIM, and DOMAIN. We highlight that choosing appropriate SDMs to address a specific problem is an important part of the modeling process.
| S-EPMC4226630 | biostudies-literature
Project description:Feature selection is one of the main techniques used to prevent overfitting in machine learning applications. The most straightforward approach for feature selection is an exhaustive search: one can go over all possible feature combinations and pick up the model with the highest accuracy. This method together with its optimizations were actively used in biomedical research, however, publicly available implementation is missing. We present ExhauFS-the user-friendly command-line implementation of the exhaustive search approach for classification and survival regression. Aside from tool description, we included three application examples in the manuscript to comprehensively review the implemented functionality. First, we executed ExhauFS on a toy cervical cancer dataset to illustrate basic concepts. Then, multi-cohort microarray breast cancer datasets were used to construct gene signatures for 5-year recurrence classification. The vast majority of signatures constructed by ExhauFS passed 0.65 threshold of sensitivity and specificity on all datasets, including the validation one. Moreover, a number of gene signatures demonstrated reliable performance on independent RNA-seq dataset without any coefficient re-tuning, i.e., turned out to be cross-platform. Finally, Cox survival regression models were used to fit isomiR signatures for overall survival prediction for patients with colorectal cancer. Similarly to the previous example, the major part of models passed the pre-defined concordance index threshold 0.65 on all datasets. In both real-world scenarios (breast and colorectal cancer datasets), ExhauFS was benchmarked against state-of-the-art feature selection models, including L1-regularized sparse models. In case of breast cancer, we were unable to construct reliable cross-platform classifiers using alternative feature selection approaches. In case of colorectal cancer not a single model passed the same 0.65 threshold. Source codes and documentation of ExhauFS are available on GitHub: https://github.com/s-a-nersisyan/ExhauFS.
| S-EPMC8976470 | biostudies-literature
Project description:Coffee is one of the most important beverages and trade products in the world. Among the multiple research initiatives focused on coffee sustainability, plant breeding provides the best means to increase phenotypic performance and release cultivars that could meet market demands. Since coffee is well adapted to a diversity of tropical environments, an important question for those confronting the problem of evaluating phenotypic performance is the relevance of genotype-by-environment interaction. As a perennial crop with a long juvenile phase, coffee is subjected to significant temporal and spatial variations. Such facts not only hinder the selection of promising materials but also cause a majority of complaints among growers. In this study, we hypothesized that trait stability in coffee is genetically controlled and therefore is predictable using molecular information. To test it, we used genome-based methods to predict stability metrics computed with the primary goal of selecting coffee genotypes that combine high phenotypic performance and stability for target environments. Using 2 populations of Coffea canephora, evaluated across multiple years and locations, our contribution is 3-fold: (1) first, we demonstrated that the number of harvest evaluations may be reduced leading to accelerated implementation of molecular breeding; (2) we showed that stability metrics are predictable; and finally, (3) both stable and high-performance genotypes can be simultaneously predicted and selected. While this research was carried out on representative environments for coffee production with substantial crossover in genotypic ranking, we anticipate that genomic prediction can be an efficient tool to select coffee genotypes that combine high performance and stability across years and the target locations here evaluated.
| S-EPMC10234400 | biostudies-literature
Project description:We present a statistical learning framework for robust identification of differential equations from noisy spatio-temporal data. We address two issues that have so far limited the application of such methods, namely their robustness against noise and the need for manual parameter tuning, by proposing stability-based model selection to determine the level of regularization required for reproducible inference. This avoids manual parameter tuning and improves robustness against noise in the data. Our stability selection approach, termed PDE-STRIDE, can be combined with any sparsity-promoting regression method and provides an interpretable criterion for model component importance. We show that the particular combination of stability selection with the iterative hard-thresholding algorithm from compressed sensing provides a fast and robust framework for equation inference that outperforms previous approaches with respect to accuracy, amount of data required, and robustness. We illustrate the performance of PDE-STRIDE on a range of simulated benchmark problems, and we demonstrate the applicability of PDE-STRIDE on real-world data by considering purely data-driven inference of the protein interaction network for embryonic polarization in Caenorhabditis elegans. Using fluorescence microscopy images of C. elegans zygotes as input data, PDE-STRIDE is able to learn the molecular interactions of the proteins.
| S-EPMC9199075 | biostudies-literature
Project description:BACKGROUND:The performance of a new diagnostic test is typically evaluated against a comparator which is assumed to correspond closely to some true state of interest. Judgments about the new test's performance are based on the differences between the outputs of the test and comparator. It is commonly assumed that a small amount of uncertainty in the comparator's classifications will negligibly affect the measured performance of a diagnostic test. METHODS:Simulated datasets were generated to represent typical diagnostic scenarios. Comparator noise was introduced in the form of random misclassifications, and the effect on the apparent performance of the diagnostic test was determined. An actual dataset from a clinical trial on a new diagnostic test for sepsis was also analyzed. RESULTS:We demonstrate that as little as 5% misclassification of patients by the comparator can be enough to statistically invalidate performance estimates such as sensitivity, specificity and area under the receiver operating characteristic curve, if this uncertainty is not measured and taken into account. This distortion effect is found to increase non-linearly with comparator uncertainty, under some common diagnostic scenarios. For clinical populations exhibiting high degrees of classification uncertainty, failure to measure and account for this effect will introduce significant risks of drawing false conclusions. The effect of classification uncertainty is magnified further for high performing tests that would otherwise reach near-perfection in diagnostic evaluation trials. A requirement of very high diagnostic performance for clinical adoption, such as a 99% sensitivity, can be rendered nearly unachievable even for a perfect test, if the comparator diagnosis contains even small amounts of uncertainty. This paper and an accompanying online simulation tool demonstrate the effect of classification uncertainty on the apparent performance of tests across a range of typical diagnostic scenarios. Both simulated and real datasets are used to show the degradation of apparent test performance as comparator uncertainty increases. CONCLUSIONS:Overall, a 5% or greater misclassification rate by the comparator can lead to significant underestimation of true test performance. An online simulation tool allows researchers to explore this effect using their own trial parameters (https://imperfect-gold-standard.shinyapps.io/classification-noise/) and the source code is freely available (https://github.com/ksny/Imperfect-Gold-Standard).
| S-EPMC6530857 | biostudies-literature
Project description:The main objective of this research was to evaluate the efficacy of intermittent hypoxic training (IHT) on anaerobic and aerobic capacity and swimming performance in well-trained swimmers. Sixteen male swimmers were randomly divided into a hypoxia (H) group (n = 8), which trained in a normobaric hypoxia environment, and a control (C) group (n = 8), which exercised under normoxic conditions. However, one participant left the study without explanation. During the experiment group H trained on land twice per week in simulated hypoxia (FiO2 = 15.5%, corresponding to 2,500 m a.s.l); however, they conducted swim training in normoxic conditions. Group C performed the same training program under normoxic conditions. The training program included four weekly microcyles, followed by three days of recovery. During practice sessions on land, the swimmers performed 30 second sprints on an arm-ergometer, alternating with two minute high intensity intervals on a lower limb cycle ergometer. The results showed that the training on land caused a significant (p<0.05) increase in absolute maximal workload (WRmax) by 7.4% in group H and by 3.2% in group C and relative values of VO2max by 6.9% in group H and 3.7% in group C. However, absolute values of VO2max were not significantly changed. Additionally, a significant (p<0.05) increase in mean power (Pmean) during the first (11.7%) and second (11.9%) Wingate tests was only observed in group H. The delta values of lactate concentration (ΔLA) after both Wingate tests were significantly (p<0.05) higher in comparison to baseline levels by 28.8% in group H. Opposite changes were observed in delta values of blood pH (ΔpH) after both Wingate tests in group H, with a significant decrease in values of ΔpH by 33.3%. The IHT caused a significant (p<0.05) improvement in 100m and 200m swimming performance, by 2.1% and 1.8%, respectively in group H. Training in normoxia (group C), resulted in a significant (p<0.05) improvement of swimming performance at 100m and 200m, by 1.1% and 0.8%, respectively. In conclusion, the most important finding of this study includes a significant improvement in anaerobic capacity and swimming performance after high-intensity IHT. However, this training protocol had no effect on absolute values of VO2max and hematological variables.
| S-EPMC5538675 | biostudies-literature
Project description:Performance measurement at the provider group level is increasingly advocated, but different methods for selecting patients when calculating provider group performance have received little evaluation.We compared 2 currently used methods according to characteristics of the patients selected and impact on performance estimates.We analyzed Medicare claims data for fee-for-service beneficiaries with diabetes ever seen at an academic multispeciality physician group in 2003 to 2004. We examined sample size, sociodemographics, clinical characteristics, and receipt of recommended diabetes monitoring in 2004 for the groups of patients selected using 2 methods implemented in large-scale performance initiatives: the Plurality Provider Algorithm and the Diabetes Care Home method. We examined differences among discordantly assigned patients to determine evidence for differential selection regarding these measures.Fewer patients were selected under the Diabetes Care Home method (n=3558) than the Plurality Provider Algorithm (n=4859). Compared with the Plurality Provider Algorithm, the Diabetes Care Home method preferentially selected patients who were female, not entitled because of disability, older, more likely to have hypertension, and less likely to have kidney disease and peripheral vascular disease, and had lower levels of predicted utilization. Diabetes performance was higher under Diabetes Care Home method, with 67% versus 58% receiving >1 A1c tests, 70% versus 65% receiving ?1 low-density lipoprotein (LDL) test, and 38% versus 37% receiving an eye examination.The method used to select patients when calculating provider group performance may affect patient case mix and estimated performance levels, and warrants careful consideration when comparing performance estimates.
| S-EPMC3137682 | biostudies-literature