scholarly journals Feature selection and causal analysis for microbiome studies in the presence of confounding using standardization

2021 ◽  
Vol 22 (1) ◽  
Author(s):  
Emily Goren ◽  
Chong Wang ◽  
Zhulin He ◽  
Amy M. Sheflin ◽  
Dawn Chiniquy ◽  
...  

Abstract Background Microbiome studies have uncovered associations between microbes and human, animal, and plant health outcomes. This has led to an interest in developing microbial interventions for treatment of disease and optimization of crop yields which requires identification of microbiome features that impact the outcome in the population of interest. That task is challenging because of the high dimensionality of microbiome data and the confounding that results from the complex and dynamic interactions among host, environment, and microbiome. In the presence of such confounding, variable selection and estimation procedures may have unsatisfactory performance in identifying microbial features with an effect on the outcome. Results In this manuscript, we aim to estimate population-level effects of individual microbiome features while controlling for confounding by a categorical variable. Due to the high dimensionality and confounding-induced correlation between features, we propose feature screening, selection, and estimation conditional on each stratum of the confounder followed by a standardization approach to estimation of population-level effects of individual features. Comprehensive simulation studies demonstrate the advantages of our approach in recovering relevant features. Utilizing a potential-outcomes framework, we outline assumptions required to ascribe causal, rather than associational, interpretations to the identified microbiome effects. We conducted an agricultural study of the rhizosphere microbiome of sorghum in which nitrogen fertilizer application is a confounding variable. In this study, the proposed approach identified microbial taxa that are consistent with biological understanding of potential plant-microbe interactions. Conclusions Standardization enables more accurate identification of individual microbiome features with an effect on the outcome of interest compared to other variable selection and estimation procedures when there is confounding by a categorical variable.

2020 ◽  
Author(s):  
Emily Goren ◽  
Chong Wang ◽  
Zhulin He ◽  
Amy M Sheflin ◽  
Dawn Chiniquy ◽  
...  

AbstractBackgroundMicrobiome studies have uncovered associations between microbes and human, animal, and plant health outcomes. This has led to an interest in developing microbial interventions for treatment of disease and optimization of crop yields which requires identification of microbiome features that impact the outcome in the population of interest. That task is challenging because of the high dimensionality of microbiome data and the confounding that results from the complex and dynamic interactions among host, environment, and microbiome. In the presence of such confounding, variable selection and estimation procedures may have unsatisfactory performance in identifying microbial features with an effect on the outcome.ResultsIn this manuscript, we aim to estimate population-level effects of individual microbiome features while controlling for confounding by a categorical variable. Due to the high dimensionality and confounding-induced correlation between features, we propose feature screening, selection, and estimation conditional on each stratum of the confounder followed by a standardization approach to estimation of population-level effects of individual features.Comprehensive simulation studies demonstrate the advantages of our approach in recovering relevant features. Utilizing a potential-outcomes framework, we outline assumptions required to ascribe causal, rather than associational, interpretations to the identified microbiome effects. We conducted an agricultural study of the rhizosphere microbiome of sorghum in which nitrogen fertilizer application is a confounding variable. In this study, the proposed approach identified microbial taxa that are consistent with biological understanding of potential plant-microbe interactions.ConclusionsStandardization enables more accurate identification of individual microbiome features with an effect on the outcome of interest compared to other variable selection and estimation procedures when there is confounding by a categorical variable.


Solid Earth ◽  
2016 ◽  
Vol 7 (1) ◽  
pp. 93-103 ◽  
Author(s):  
B. G. J. S. Sonneveld ◽  
M. A. Keyzer ◽  
D. Ndiaye

Abstract. Land degradation has been a persistent problem in Senegal for more than a century and by now has become a serious impediment to long-term development. In this paper, we quantify the impact of land degradation on crop yields using the results of a nationwide land degradation assessment. For this, the study needs to address two issues. First, the land degradation assessment comprises qualitative expert judgements that have to be converted into more objective, quantitative terms. We propose a land degradation index and assess its plausibility. Second, observational data on soils, land use, and rainfall do not provide sufficient information to isolate the impact of land degradation. We, therefore, design a pseudo-experiment that for sites with otherwise similar circumstances compares the yield of a site with and one without land degradation. This pairing exercise is conducted under a gradual refining of the classification of circumstances, until a more or less stable response to land degradation is obtained. In this way, we hope to have controlled sufficiently for confounding variables that will bias the estimation of the impact of land degradation on crop yields. A small number of shared characteristics reveal tendencies of "severe" land degradation levels being associated with declining yields as compared to similar sites with "low" degradation levels. However, as we zoom in at more detail some exceptions come to the fore, in particular in areas without fertilizer application. Yet, our overall conclusion is that yield reduction is associated with higher levels of land degradation, irrespective of whether fertilizer is being applied or not.


2020 ◽  
Vol 18 (1) ◽  
Author(s):  
Szilvia Harsanyi ◽  
Nandor Balogh ◽  
Laszlo Robert Kolozsvari ◽  
Laszlo Mezes ◽  
Csaba Papp ◽  
...  

Abstract Background Translating clinical guidelines into routine clinical practice is mandatory to achieve population level improvement of health. Emergence of specific therapy for acute stroke yielded the ‘time is brain’ concept introducing the need for emergency treatment, pointing to the need for increasing stroke awareness of the general population. General practitioners (GPs) manage chronic diseases and could hence catalyse stroke awareness. In our study, the knowledge of general practitioners toward accurate identification of acute stroke candidacy was investigated. Methods GPs and residents in training for family medicine participated in a survey on a voluntary basis using supervised self-administration between the 1st of February 2018 and 31st July 2018. Two clinical cases of acute stroke that differed only regarding the patient’s eligibility for intravenous thrombolysis were presented. Participants answered two open-ended questions. Text analysis was performed using NVIVO software. Results Of the 127 respondents, 69 (54.3%) were female. The median age was 49 (34–62) years. The median time spent working after graduation was 14.5 (2–22.5) years. Board-certified GPs made up 77.2% of the sample. Qualitative analysis revealed stroke as the most frequent diagnosis for both cases. Territorial localization and possible aetiology were also established. Respondents properly identified eligibility for thrombolysis. Quantitative assessment showed that having the practice closer to the stroke centre increases the likelihood of adequate diagnosis for acute stroke. Conclusions Our results show that GPs properly diagnose acute stroke and identify intravenous thrombolysis candidates. Moreover, we found that a vigorous acute stroke triage system facilitates the translation of theory into practice.


Sign in / Sign up

Export Citation Format

Share Document