Biology subjects

Silverman, J. D.

Publications and source records attributed to Silverman, J. D..

At least 19 recordsLinked to original sources

bioRxiv2025

Not every gene is special: one simple rule to control the false discovery rate when analysing high-throughput sequencing data

1Differential expression and differential abundance analyses are commonplace in studies employing high-throughput sequencing approaches; however different tools often fail to return comparable results when applied to the same dataset. Most tools employ various normalisations to attempt to correct for technical variation in the count data. Previously, we demonstrated that these normalisations are often inappropriate due to incorrect assumptions regarding the overall scale (i.e. size) of the biological system in question. In this study, we conducted 100 permutation analyses of 10 RNA-seq datasets to show that scale misspecification results in poor control of the false discovery rate by several commonly-used analysis tools. Moreover, we demonstrate that this can be ameliorated by using a scale model in ALDEx2 or ALDEx3 to account for uncertainty around the size of the system scale. We show that there is an inherent trade-off between satisfactory control of false-discovery rates and sensitivity and that no tool offers both. We also quantify how increasing scale uncertainty affects the difference between groups required for a feature to be reported as differentially expressed. Finally, we provide universal guidance on choosing an appropriate amount of scale uncertainty for any type of analysis. Overall, our work highlights the strengths and pitfalls of commonly used tools for differential expression analyses and highlights the choice between sensitivity and false-discovery rate control that all researchers are making when analysing sequencing data.

bioinformatics↗

bioRxiv2025

Uncertainty Modeling Outperforms Machine Learning for Microbiome Data Analysis

Microbiome sequencing measures relative rather than absolute abundances, providing no direct information about total microbial load. Normalization methods attempt to compensate, but rely on strong, often untestable assumptions that can bias inference. Experimental measurements of load (e.g., qPCR, flow cytometry) offer a solution, but remain costly and uncommon. A recent high-profile study proposed that machine learning could bypass this limitation by predicting microbial load from sequencing data alone. To evaluate this claim, we assembled mutt, the largest public database of paired sequencing and load measurements, spanning 35 studies and over 15,000 samples. Using mutt, we show that published machine learning models fail to generalize: on average they perform worse than a naive baseline that always predicted the training set mean. These failures stem from covariate shift-limited shared taxa between studies, differences in community composition, and differences in preprocessing pipelines-that silently derail model inputs. In contrast, Bayesian partially identified models do not attempt to impute microbial load, but instead propagate scale uncertainty through downstream analyses. Across 30 benchmark datasets, Bayesian partially identified models consistently outperformed normalization and machine learning approaches, providing a principled and reproducible foundation for microbiome inference.

bioinformatics↗

bioRxiv2025

Scale Reliant Mixed Effects Models Enhance Microbiome Data Analysis

Linear models, including those used for differential abundance analyses, are frequently used in microbiome research to assess how experimental conditions (e.g., disease state or age) affect microbial abundance. Linear mixed-effects models (MEMs) extend linear models to accommodate complex designs, such as longitudinal sampling or hierarchical study structures. However, when applied to microbiome data, existing MEM approaches suffer from high false positive and false negative rates because sequence counts are compositional - they reflect relative rather than absolute abundances. Current methods attempt to overcome this limitation through normalization, but these approaches imply on strong, often unrealistic assumptions about the unmeasured biological scale (e.g., total microbial load). Here we introduce scale-reliant mixed-effects models (SR-MEM), which extend our earlier scale-reliant inference framework by explicitly modeling uncertainty in the unmeasured scale via user-defined probability distributions. By treating scale as a latent variable rather than fixing it through normalization, SR-MEM yields robust inference for complex experimental designs. Across simulations and real datasets, SR-MEM is the only method that consistently controls the false discovery rate while achieving higher power than both normalization and bias-correction methods. SR-MEM can also incorporate external scale measurements (e.g., flow cytometry, qPCR) or leverage independent studies to further improve inference. An accessible implementation is provided in the ALDEx3 R package, enabling more rigorous and reproducible analysis of microbial communities.

bioinformatics↗

bioRxiv2025

PCR Bias Impacts Microbiome Ecological Analyses

Polymerase Chain Reaction (PCR) is a critical step in amplicon-based microbial community profiling, allowing the selective amplification of marker genes such as 16S rRNA from environmental or host-associated samples. Despite its widespread use, PCR is known to introduce amplification bias, where some DNA sequences are preferentially amplified over others due to factors such as primer-template mismatches, sequence GC content, and secondary structures. Although these biases are known to affect transcript abundance, their implications for ecological metrics remain poorly understood. In this study, we conduct a comprehensive evaluation of how PCR-bias influences both within-samples (-diversity) and between-sample ({beta}-diversity) analyses. We show that perturbation-invariant diversity measures remain unaffected by PCR bias, but widely used metrics such as Shannon diversity and Weighted-Unifrac are sensitive, with their values varying according to the true community composition. To address this, we provide theoretical and empirical insight into how PCR-induced bias varies across ecological analyses and community structures, and we offer practical guidance on when bias-correction methods should be applied. Our findings highlight the importance of selecting appropriate diversity metrics for PCR-based microbial ecology workflows and offer guidance for improving the reliability of diversity analyses.

microbiology↗

bioRxiv2024

Replacing Normalizations with Interval AssumptionsImproves the Rigor and Robustness of DifferentialExpression and Differential Abundance Analyses

Standard methods for differential expression and differential abundance analysis rely on normalization to address sample-to-sample variation in sequencing depth. However, normalizations imply strict, unrealistic assumptions about the unmeasured scale of biological systems (e.g., microbial load or total cellular transcription). This introduces bias that can lead to false positives and false negatives. To overcome these limitations, we suggest replacing normalizations with interval assumptions. This approach allows researchers to explicitly define plausible lower and upper bounds on the unmeasured biological systems scale, making these assumptions more realistic, transparent, and flexible than those imposed by traditional normalizations. Compared to recent alternatives like scale models and sensitivity analyses, interval assumptions are easier to use, resulting in potentially reduced false positives and false negatives, and have stronger guarantees of Type-I error control. We make interval assumptions accessible by introducing a modified version of ALDEx2 as a publicly available software package. Through simulations and real data studies, we show these methods can reduce false positives and false negatives compared to normalization-based tools.

bioinformatics↗

bioRxiv2024

Beyond Normalization: Incorporating Scale Uncertainty in Microbiome and Gene Expression Analysis

Though statistical normalizations are often used in differential abundance or differential expression analysis to address sample-to-sample variation in sequencing depth, we offer a better alternative. These normalizations often make strong, implicit assumptions about the scale of biological systems (e.g., microbial load). Thus, analyses are susceptible to even slight errors in these assumptions, leading to elevated rates of false positives and false negatives. We introduce scale models as a generalization of normalizations so researchers can model potential errors in assumptions about scale. By incorporating scale models into the popular ALDEx2 software, we enhance the reproducibility of analyses while often drastically decreasing false positive and false negative rates. We design scale models that are guaranteed to reduce false positives compared to equivalent normalizations. At least in the context of ALDEx2, we recommend using scale models over normalizations in all practical situations.

bioinformatics↗

bioRxiv2023

Normalizations are not what you think; Explicit Scale Simulation in ALDEx2

In high-throughput sequencing (HTS) studies, sample-to-sample variation in sequencing depth is driven by technical factors, and not by variation in the scale (e.g., total size, microbial load, or total mRNA expression) of the underlying biological systems. Typically a statistical normalization is used to remove unwanted technical variation in the data or the parameters of the model to enable analyses that are reliant on scale; e.g., differential abundance and differential expression analyses. We recently showed that all normalizations make implicit assumptions about the unmeasured system scale and that errors in these assumptions can dramatically increase false positive and false negative rates. We demonstrated that these errors can be mitigated by accounting for uncertainty about scale using a scale model, which we integrated into the ALDEx2 R package. This article provides new insights into those methods, focusing on the application to transcriptomic analysis. Here we provide transcriptomic case studies demonstrating how scale models, rather than traditional normalizations, can reduce false positive and false negative rates in practice while enhancing the transparency and reproducibility of analyses. We show that these scale models replace the need for dual cutoff approaches often used to address the disconnect between practical and statistical significance. We demonstrate the utility of that scale models built based on known housekeeping genes in complex metatranscriptomic datasets. Thus this work provides example and practical guidance on how to incorporate scale into transcriptomic analysis.

bioinformatics↗

bioRxiv2023

Changes in the gut mycobiome are associated with Type 2 diabetes mellitus and metformin treatment across populations

The human gut teems with a diverse ecosystem of microbes, yet non-bacterial portions of that community are overlooked in studies of metabolic diseases firmly linked to gut bacteria. Type 2 diabetes mellitus (T2D) associates with compositional shifts in the gut bacterial microbiome and fungal mycobiome, but whether T2D and/or pharmaceutical treatments underpin the community change is unresolved. To differentiate these effects, we curated a gut mycobiome cohort to-date spanning 1,000 human samples across 5 countries and a murine experimental model. We use Bayesian multinomial logistic normal models to show that metformin and T2D both associate with shifts in the relative abundance of distinct gut fungi. T2D associates with shifts in the Saccharomycetes and Sordariomycetes fungal classes, while the genera Fusarium and Tetrapisipora most consistently associate with metformin treatment. We confirmed the impact of metformin on individual gut fungi by administering metformin to healthy mice. Thus, metformin and T2D account for subtle, but significant and distinct variation in the gut mycobiome across human populations. This work highlights for the first time that oral pharmaceuticals can confound associations of gut fungi with T2D and warrants the need to consider pharmaceutical interventions in investigations of linkages between metabolic diseases and gut microbial inhabitants.

microbiology↗

bioRxiv2023

Quantitative Geometric Modeling of Blood Cells from X-ray Histotomograms of Whole Zebrafish Larvae

Tissue phenotyping is foundational to understanding and assessing the cellular aspects of disease in organismal context and an important adjunct to molecular studies in the dissection of gene function, chemical effects, and disease. As a first step toward computational tissue phenotyping, we explore the potential of cellular phenotyping from 3-Dimensional (3D), 0.74 {micro}m isotropic voxel resolution, whole zebrafish larval images derived from X-ray histotomography, a form of micro-CT customized for histopathology. As proof of principle towards computational tissue phenotyping of cells, we created a semi-automated mechanism for the segmentation of blood cells in the vascular spaces of zebrafish larvae, followed by modeling and extraction of quantitative geometric parameters. Manually segmented cells were used to train a random forest classifier for blood cells, enabling the use of a generalized cellular segmentation algorithm for the accurate segmentation of blood cells. These models were used to create an automated data segmentation and analysis pipeline to guide the steps in a 3D workflow including blood cell region prediction, cell boundary extraction, and statistical characterization of 3D geometric and cytological features. We were able to distinguish blood cells at two stages in development (4- and 5-days-post-fertilization) and wild-type vs. polA2 huli hutu (hht) mutants. The application of geometric modeling across cell types to and across organisms and sample types may comprise a valuable foundation for computational phenotyping that is more open, informative, rapid, objective, and reproducible.

systems biology↗

bioRxiv2023

Phosphorylation barcodes direct biased chemokine signaling at CXCR3

G protein-coupled receptor (GPCR) biased agonism, the activation of some signaling pathways over others, is thought to largely be due to differential receptor phosphorylation, or "phosphorylation barcodes." At chemokine receptors, ligands act as "biased agonists" with complex signaling profiles, which contributes to the limited success in pharmacologically targeting these receptors. Here, mass spectrometry-based global phosphoproteomics revealed that CXCR3 chemokines generate different phosphorylation barcodes associated with differential transducer activation. Chemokine stimulation resulted in distinct changes throughout the kinome in global phosphoproteomic studies. Mutation of CXCR3 phosphosites altered {beta}-arrestin conformation in cellular assays and was confirmed by molecular dynamics simulations. T cells expressing phosphorylation-deficient CXCR3 mutants resulted in agonist- and receptor-specific chemotactic profiles. Our results demonstrate that CXCR3 chemokines are non-redundant and act as biased agonists through differential encoding of phosphorylation barcodes and lead to distinct physiological processes.

biochemistry↗

bioRxiv2023

Addressing Erroneous Scale Assumptions in Microbe and Gene Set Enrichment Analysis

By applying Differential Set Analysis (DSA) to sequence count data, researchers can determine whether groups of microbes or genes are differentially enriched. Yet these data lack information about the scale (i.e., size) of the biological system under study, leading some authors to call these data compositional (i.e., proportional). In this article we show that commonly used DSA methods make strong, implicit assumptions about the unmeasured system scale. We show that even small errors in these assumptions can lead to false positive rates as high as 70%. To mitigate this problem, we introduce a sensitivity analysis framework to identify when modeling results are robust to such errors and when they are suspect. Unlike standard benchmarking studies, our methods do not require ground-truth knowledge and can therefore be applied to both simulated and real data.

bioinformatics↗

bioRxiv2021

Decoding the olfactory map: targeted transcriptomics link olfactory sensory neurons to glomeruli

Sensory processing in vertebrate olfactory systems is organized across olfactory bulb glomeruli, wherein axons of peripheral sensory neurons expressing the same olfactory receptor co-terminate to transmit receptor-specific activity to central neurons. Understanding how receptors map to glomeruli is therefore critical to understanding olfaction. High-throughput spatial transcriptomics is a rapidly advancing field, but low-abundance olfactory receptor expression within glomeruli has previously precluded high-throughput mapping of receptors to glomeruli. Here we combined spatial sectioning along the anteroposterior, dorsoventral, and mediolateral axes with target capture enrichment sequencing to overcome low-abundance target expression. This strategy allowed us to spatially map 86% of olfactory receptors across the olfactory bulb and uncover a relationship between OR sequence and glomerular position. ONE SENTENCE SUMMARYTargeted enrichment of olfactory receptor mRNA in olfactory bulb sections determines spatial positions for murine glomeruli.

neuroscience↗

bioRxiv2021

Microbiota responses to different prebiotics are conserved within individuals and associated with habitual fiber intake

BackgroundShort-chain fatty acids (SCFAs) derived from gut bacteria are associated with protective roles in diseases ranging from obesity to colorectal cancers. Intake of microbially accessible dietary fibers (prebiotics) lead to varying effects on SCFA production in human studies, and gut microbial responses to nutritional interventions vary by individual. It is therefore possible that prebiotic therapies will require customizing to individuals. ResultsHere, we explored prebiotic personalization by conducting a three-way crossover study of three prebiotic treatments in healthy adults. We found that within individuals, metabolic responses were correlated across the three prebiotics. Individual identity, rather than prebiotic choice, was also the major determinant of SCFA response. Across individuals, prebiotic response was inversely related to basal fecal SCFA concentration, which, in turn, was associated with habitual fiber intake. Experimental measures of gut microbial SCFA production for each participant also negatively correlated with fiber consumption, supporting a model in which individuals gut microbiota are limited in their overall capacity to produce fecal SCFAs from fiber. ConclusionsOur findings support developing personalized prebiotic regimens that focus on selecting individuals who stand to benefit, and that such individuals are likely to be deficient in fiber intake.

microbiology↗

bioRxiv2020

Short-chain fatty acid production by gut microbiota from children with obesity is linked to bacterial community composition and prebiotic choice

Pediatric obesity remains a public health burden and continues to increase in prevalence. The gut microbiota plays a causal role in obesity and is a promising therapeutic target. Specifically, the microbial production of short-chain fatty acids (SCFA) from the fermentation of otherwise indigestible dietary carbohydrates may protect against pediatric obesity and metabolic syndrome. Still, it has not been demonstrated that therapies involving microbiota-targeting carbohydrates, known as prebiotics, will enhance gut bacterial SCFA production in children and adolescents with obesity (age 10-18). Here, we used an in vitro system to examine the SCFA production by fecal microbiota from 17 children with obesity when exposed to five different commercially available over-the-counter (OTC) prebiotic supplements. We found microbiota from all 17 patients actively metabolized most prebiotics. Still, supplements varied in their acidogenic potential. Significant inter-donor variation also existed in SCFA production, which 16S rRNA sequencing supported as being associated with differences in the host microbiota composition. Last, we found that neither fecal SCFA concentration, microbiota SCFA production capacity, nor markers of obesity positively correlated with one another. Together, these in vitro findings suggest the hypothesis that OTC prebiotic supplements may be unequal in their ability to stimulate SCFA production in children and adolescents with obesity, and that the most acidogenic prebiotic may differ across individuals. IMPORTANCEPediatric obesity remains a major public health problem in the US, where 17% of children and adolescents are obese, and rates of pediatric severe obesity are increasing. Children and adolescents with obesity face higher health risks, and non-invasive therapies for pediatric obesity often have limited success. The human gut microbiome has been implicated in adult obesity, and microbiota-directed therapies can aid weight loss in adults with obesity. However, less is known about the microbiome in pediatric obesity, and microbiota-directed therapies are understudied in children and adolescents. Our research has two important findings: 1) dietary prebiotics (fiber) cause the microbiota from adolescents with obesity to produce more SCFA, and 2) the effectiveness of each prebiotic is donor-dependent. Together, these findings suggest that prebiotic supplements could help children and adolescents with obesity, but that these therapies may not be one-size-fits-all.

microbiology↗

bioRxiv2019

High-throughput isolation and culture of human gut bacteria with droplet microfluidics

Isolation and culture of gut bacteria enable testing for microbial roles in disease and may also lead to novel therapeutics. However, the diversity of human gut microbial communities (microbiota) impedes comprehensive experimental studies of individual bacterial taxa. Here, we combine advances in droplet microfluidics and high-throughput DNA sequencing to develop a platform for isolating and assaying microbiota members in picoliter droplets (MicDrop). MicDrop can be used to create millions of distinct bacterial colonies in a single experiment while using off-the-shelf parts compact enough to fit in an anaerobic chamber. In proof-of-concept experiments, we used the platform to characterize inter-individual metabolic variation among hundreds of polysaccharide-degrading gut bacteria from nine stool donors. We also used MicDrop to test the hypothesis that growth kinetics of individual gut bacterial taxa are associated with longterm community dynamics in an artificial gut. These demonstrations suggest the MicDrop platform could support future diagnostic efforts to personalize microbiota-directed therapies, as well as to provide comprehensive new insights into the ecology of human gut microbiota.

microbiology↗

bioRxiv2019

Measuring and Mitigating PCR Bias in Microbiome Data

PCR amplification plays a central role in the measurement of mixed microbial communities via high-throughput sequencing. Yet PCR is also known to be a common source of bias in microbiome data. Here we present a paired modeling and experimental approach to characterize and mitigate PCR bias in microbiome studies. We use experimental data from mock bacterial communities to validate our approach and human gut microbiota samples to characterize PCR bias under real-world conditions. Our results suggest that PCR can bias estimates of microbial relative abundances by a factor of 2-4 but that this bias can be mitigated using simple Bayesian multinomial logistic-normal linear models.\n\nAuthor summaryHigh-throughput sequencing is often used to profile host-associated microbial communities. Many processing steps are required to transform a community of bacteria into a pool of DNA suitable for sequencing. One important step is amplification where, to create enough DNA for sequencing, DNA from many different bacteria are repeatedly copied using a technique called Polymerase Chain Reaction (PCR). However, PCR is known to introduce bias as DNA from some bacteria are more efficiently copied than others. Here we introduce an experimental procedure that allows this bias to be measured and computational techniques that allow this bias to be mitigated in sequencing data.

genomics↗

bioRxiv2018

Naught all zeros in sequence count data are the same

Genomic studies feature multivariate count data from high-throughput DNA sequencing experiments, which often contain many zero values. These zeros can cause artifacts for statistical analyses and multiple modeling approaches have been developed in response. Here, we apply common zero-handling models to gene-expression and microbiome datasets and show models disagree on average by 46% in terms of identifying the most differentially expressed sequences. Next, to rationally examine how different zero handling models behave, we developed a conceptual framework outlining four types of processes that may give rise to zero values in sequence count data. Last, we performed simulations to test how zero handling models behave in the presence of these different zero generating processes. Our simulations showed that simple count models are sufficient across multiple processes, even when the true underlying process is unknown. On the other hand, a common zero handling technique known as "zero-inflation" was only suitable under a zero generating process associated with an unlikely set of biological and experimental conditions. In concert, our work here suggests several specific guidelines for developing and choosing state-of-the-art models for analyzing sparse sequence count data.

bioinformatics↗

bioRxiv2018

Statistical Considerations in the Design and Analysis of Longitudinal Microbiome Studies

Longitudinal studies of microbial communities have emphasized that host-associated microbiota are highly dynamic as well as underscoring the potential biomedical relevance of understanding these dynamics. Despite this increasing appreciation, statistical challenges in the design and analysis of longitudinal microbiome studies such as sequence counting, technical variation, signal aliasing, contamination, sparsity, missing data, and algorithmic scalability remain. In this review we discuss these challenges and highlight current progress in the field. Where possible, we try to provide guidelines for best practices as well as discuss how to tailor design and analysis to the hypothesis and ecosystem under study. Overall, this review is intended to serve as an introduction to longitudinal microbiome studies for both statisticians new to the microbiome field as well as biologists with little prior experience with longitudinal study design and analysis.

microbiology↗