Biology subjects

Mi, X.

Publications and source records attributed to Mi, X..

23 records · Page 2Linked to original sources

bioRxiv2020

Species packing and the latitudinal gradient in local beta-diversity

The decline in species richness at higher latitudes is among the most fundamental patterns in ecology. Whether changes in species composition across space (beta-diversity) contribute to this gradient of overall species richness (gamma-diversity) remains hotly debated. Previous studies that failed to resolve the issue suffered from a well-known tendency for small samples in areas with high gamma-diversity to have inflated measures of beta-diversity. Here, we provide here a novel analytical test, using beta-diversity metrics that correct the gamma-diversity and sampling biases, to compare beta-diversity and species packing across a latitudinal gradient in tree species richness of 21 large forest plots along a large environmental gradient in East Asia. We demonstrate that after accounting for topography and correcting the gamma-diversity bias, tropical forests still have higher beta-diversity than temperate analogs. This suggests that beta-diversity contributes to the latitudinal species richness gradient as a component of gamma-diversity. Moreover, both niche specialization and niche marginality (a measure of niche spacing along an environmental gradient) also increase towards the equator, after controlling for the effect of topographic heterogeneity. This supports the joint importance of tighter species packing and larger niche space in tropical forests while also demonstrating the importance of local processes in controlling beta-diversity.

ecology↗

bioRxiv2020

A Deep Learning Semiparametric Regression for Adjusting Complex Confounding Structures

Deep Treatment Learning (deepTL), a robust yet efficient deep learning-based semiparametric regression approach, is proposed to adjust the complex confounding structures in comparative effectiveness analysis of observational data, e.g. electronic health record (EHR) data, in which complex confounding structures are often embedded. Specifically, we develop a deep learning neural network with a score-based ensembling scheme for flexible function approximation. An improved semiparametric procedure is further developed to enhance the performance of the proposed method under finite sample settings. Comprehensive numerical studies have demonstrated the superior performance of the proposed methods as compared with existing methods, with a remarkably reduced bias and mean squared error in parameter estimates. The proposed research is motivated by a post-surgery pain study, which is also used to illustrate the practical application of deepTL. Finally, an R package, "deepTL", is developed to implement the proposed method.

neuroscience↗

bioRxiv2020

Permutation-based Identification of Important Biomarkers for Complex Diseases via Black-box Models

Study of human disease remains challenging due to convoluted disease etiologies and complex molecular mechanisms at genetic, genomic, and proteomic levels. Many machine learning-based methods, including deep learning and random forest, have been developed and widely used to alleviate some analytic challenges in complex human disease studies. While enjoying the modeling flexibility and robustness, these model frameworks suffer from non-transparency and difficulty in interpreting the role of each individual feature due to their intrinsic black-box natures. However, identifying important biomarkers associated with complex human diseases is a critical pursuit towards assisting researchers to establish novel hypotheses regarding prevention, diagnosis and treatment of complex human diseases. Herein, we propose a Permutation-based Feature Importance Test (PermFIT) for estimating and testing the feature importance, and for assisting interpretation of individual feature in various black-box frameworks, including deep neural networks, random forests, and support vector machines. PermFIT (available at https://github.com/SkadiEye/deepTL) is implemented in a computationally efficient manner, without model refitting for each permuted data. We conduct extensive numerical studies under various scenarios, and show that PermFIT not only yields valid statistical inference, but also helps to improve the prediction accuracy of black-box models with top selected features. With the application to the Cancer Genome Atlas (TCGA) kidney tumor data and the HITChip atlas BMI data, PermFIT clearly demonstrates its practical usage in identifying important biomarkers and boosting performance of black-box predictive models.

bioinformatics↗

bioRxiv2020

Foundation species across a latitudinal gradient in China

AbstractFoundation species play important roles in structuring forest communities and ecosystems. Foundation species are difficult to identify without long-term observations or experiments and their foundational roles rarely are identified before they are declining or threatened. We used new statistical criteria based on size-frequency distributions, species diversity, and spatial codispersion among woody plants to identify potential ("candidate") foundation species in 12 large forest dynamics plots spanning 26 degrees of latitude in China. We used these data to identify a suite of candidate foundation species in Chinese forests; test the hypothesis that foundation woody plant species are more frequent in the temperate zone than in the tropics; and compare these results with comparable data from the Americas to suggest candidate foundation genera in Northern Hemisphere forests. We identified more candidate foundation species in temperate plots than in subtropical or tropical plots, and this relationship was independent of the latitudinal gradient in overall species richness. Two species of Acer, the canopy tree Acer ukurunduense and the shrubby treelet Acer barbinerve were the only two species that met both criteria in full to be considered as candidate foundation species. When we relaxed the diversity criteria, Acer, Tilia, and Juglans spp., and Corlyus mandshurica were frequently identified as candidate foundation species. In tropical plots, the tree Mezzettiopsis creaghii and the shrubs or treelets Aporusa yunnanensis and Ficus hispida had some characteristics associated with foundation species. Species diversity of co-occurring woody species was negatively associated with basal area of candidate foundation species more frequently at 5- and 10-m spatial grains (scale) than at a 20-m grain. Conversely, Bray-Curtis dissimilarity was positively associated with basal area of candidate foundation species more frequently at 5-m than at 10- or 20-m grains. Our data support the hypothesis that foundation species should be more common in temperate than in tropical or boreal forests, and suggest that in the Northern Hemisphere that Acer be investigated further as a foundation tree genus.

ecology↗

bioRxiv2018

An event-based paradigm for analyzing fluorescent astrocyte activity uncovers novel single-cell and population-level physiology

Recent work examining astrocytic physiology centers on fluorescence imaging approaches, due to development of sensitive fluorescent indicators and observation of spatiotemporally complex calcium and glutamate activity. However, the field remains hindered in fully characterizing these dynamics, both within single cells and at the population-level, because of the insufficiency of current region-of-interest-based approaches to describe activity that is often spatially unfixed, size-varying, and propagative. Here, we present a paradigm-shifting analytical framework that releases astrocyte biologists from ROI-based tools. Astrocyte Quantitative Analysis (AQuA) software enables users to take an event-based approach to accurately capture and quantify the irregular activity observed in astrocyte imaging datasets. We apply AQuA to a range of ex vivo and in vivo imaging data, and uncover previously undescribed physiological phenomena in each. Since AQuA is data-driven and based on machine learning principles, it can be applied across model organisms, fluorescent indicators, experimental modes, and imaging resolutions and speeds, enabling researchers to elucidate fundamental astrocyte physiology.

neuroscience↗