Biology subjects

Rahnavard, A.

Publications and source records attributed to Rahnavard, A..

13 recordsLinked to original sources

bioRxiv2026

waveome: a toolkit for longitudinal omics analysis using Gaussian processes

Identifying meaningful associations from small-sample longitudinal data is challenging, especially in low signal-to-noise environments where the Gaussian likelihood assumption does not hold. We introduce two methods to algorithmically perform variable selection with sparse, irregularly sampled, longitudinal count data with over-dispersion to characterize nonlinear relationships between omics measurements and covariates of interest using Gaussian processes. The first is an additive non-greedy search-based method, while the second is a penalization approach using Horseshoe priors on kernel hyperparameters. In simulation studies, both methods outperform conventional statistical models in terms of distributional fit and exhibit a trade-off in feature selection. Applying the penalized variant to a real-world Crohns disease cohort, we recover well-established biomarkers, such as short-chain fatty acids, secondary bile acids, and specific lipid species, and uncover novel candidates for cross-sectional and temporal disease severity. Both methods are implemented in an open-source Python library, waveome, offering a robust set of tools for longitudinal biomarker discovery.

bioinformatics↗

bioRxiv2026

Guided tokenization and domain knowledge enhance genomic language models' performance

Adapting language models to genomic and metagenomic sequences presents unique challenges, particularly in tokenization and task-specific generalization. Standard methods, such as fixed-length k-mers or byte pair encoding, often fail to preserve biologically meaningful patterns essential for downstream tasks. We introduce Guided Tokenization (GT), a strategy that prioritizes biologically and statistically important subsequences based on importance scores, model attention, and class distributions. Combined with domain adaptation, which incorporates prior domain specific biological knowledge, this approach improves both representation quality and classification accuracy in compact genomic language models (gLMs). GT enhances biological awareness in genomic language models, particularly for effective small and mid-sized models across key tasks, including DNA sequence read classification, promoter detection, antimicrobial resistance classification, and targeted amplicon taxonomic profiling. Our results highlight the promise of guided tokenization and domain-aware modeling for building efficient, biologically grounded language models for scalable genomic applications.

bioinformatics↗

bioRxiv2025

Curating 16S rRNA databases enhances taxonomic accuracy and computational efficiency in microbial profiling

The 16S rRNA gene serves as the gold standard molecular marker for microbial profiling, yet taxonomic assignment accuracy depends critically on reference database quality. Substantial heterogeneity exists among databases in sequence coverage, curation standards, and taxonomic nomenclature, leading to conflicting taxonomic assignments. Despite previous comparisons highlighting performance differences, the impact of database preprocessing, including sequence cleaning and redundancy removal, on taxonomic classification remains understudied. To improve database quality, we implemented novel cleaning approaches to remove nested sequences, duplicate sequences, and correct missing taxonomic nomenclature. We compared four major 16S databases (SILVA, Greengenes2, RefSeq, and MIMt) using 69 mock communities and the DADA2 analysis pipeline for microbial genus-level profiling. Database size was significantly reduced after cleaning: SILVA reduced from 452,055 to 291,733 sequences, Greengenes2 from 337,506 to 277,982 sequences, MIMt from 48,749 to 34,734 sequences, and RefSeq from 27,376 to 25,970 sequences. Greengenes2, MIMt, and RefSeq, which exhibited comparable performance, consistently outperformed SILVA in recall, precision, and abundance estimation accuracy across all sample types. Cleaning SILVA improved computational efficiency by up to 50% while maintaining classification performance. We provide a benchmarking framework with the cleaned databases as resources for accurate 16S rRNA analysis and profiling.

bioinformatics↗

bioRxiv2025

resLens: genomic language models to enhance antibiotic resistance gene detection

The rise of antibiotic resistance necessitates advanced tools to detect and analyze antibiotic resistance genes (ARGs). We present resLens, a family of genomic language models that leverage latent genomic representations to enhance ARG detection and analysis. Unlike alignment-based methods constrained by reference databases, resLens fine-tunes a pre-trained DNA language model on curated ARG datasets, achieving competitive or superior performance in classifying resistance genes across multiple evaluation scenarios, including when ARGs exhibit sequences and mechanisms of resistance dissimilar to those in reference datasets.

bioinformatics↗

bioRxiv2025

cellSight: Characterizing dynamics of cells using single-cell RNA-sequencing

Single-cell analysis has transformed our understanding of cellular diversity, offering insights into complex biological systems. Yet, manual data processing in single-cell studies poses challenges, including inefficiency, human error, and limited scalability. To address these issues, we propose the automated workflow cellSight, which integrates high-throughput sequencing in a user-friendly platform. By automating tasks like cell type clustering, feature extraction, and data normalization, cellSight reduces researcher workload, promoting focus on data interpretation and hypothesis generation. Its standardized analysis pipelines and quality control metrics enhance reproducibility, enabling collaboration across studies. Moreover, cellSights adaptability supports integration with emerging technologies, keeping pace with advancements in single-cell genomics. cellSight accelerates discoveries in single-cell biology, driving impactful insights and clinical translation. It is available with documentation and tutorials at https://github.com/omicsEye/cellSight.

bioinformatics↗

bioRxiv2025

Meta-analytic microbiome target discovery for immune checkpoint inhibitor response in advanced melanoma

Immune checkpoint inhibitors (ICIs) have revolutionized melanoma treatment, yet patient responses remain highly variable, underscoring the need for predictive biomarkers. Emerging evidence suggests that gut microbiome composition influences ICI efficacy, though findings remain inconsistent across studies. Here, we present a meta-analysis of seven melanoma-associated microbiome cohorts (N=678) using a standardized computational pipeline to integrate microbial species, biosynthetic gene clusters (BGCs), and functional pathways. We identify Faecalibacterium SGB15346 as a key species enriched in responders, alongside RiPP biosynthetic class and pathways involved in short-chain fatty acid fermentation. Conversely, dTDP-sugar biosynthesis correlates with non-response. Our results highlight microbial signatures and metabolic pathways associated with ICI outcomes, offering potential targets for microbiome-based interventions in personalized immunotherapy.

microbiology↗

bioRxiv2025

seqLens: optimizing language models for genomic predictions

Understanding evolutionary variation in genomic sequences through the lens of language modeling has the potential to revolutionize biological research. Yet to maximize the utility of language modeling in genomics, we must overcome computational challenges in tokenization and model architecture adapted to diverse genomic features across evolutionary timescales. In this study, we investigated key elements in genomic language modeling (gLM), including tokenization, pretraining datasets, fine-tuning approaches, pooling methods, and domain adaptation, and applied the language models to diverse genomic data. We gathered two evolutionarily distinct pretraining datasets: one consisting of 19,551 reference genomes, including over 18,000 prokaryotic genomes (115B nucleotides) and the remainder eukaryotic genomes, and another more balanced dataset with 1,354 genomes, including 1,166 prokaryotic and 188 eukaryotic reference genomes (180B nucleotides). We trained five byte-pair encoding tokenizers and pretrained 52 gLMs, systematically comparing different architectures, hyperparameters, and classification heads. We introduce seqLens, a family of models based on disentangled attention with relative positional encoding, which outperforms relatively similar-sized models in 13 of 19 benchmarking phenotypic predictions. We further explore continual pretraining, domain adaptation, and parameter-efficient fine-tuning methods to assess trade-offs between computational efficiency and accuracy. Our findings demonstrate that relevant pretraining data significantly boost performance, alternative pooling techniques can enhance classification, tokenizers with larger vocabulary sizes negatively impact generalization, and gLMs are capable of understanding evolutionary relationships. These insights provide a foundation for optimizing genomic language models for identifying diverse evolutionary genomic features and improving genome annotations.

bioinformatics↗

bioRxiv2024

Deep skin fibroblast-mediated macrophage recruitment supports acute wound healing

Epithelial and immune cells have long been appreciated for their contribution to the early immune response after injury; however, much less is known about the role of mesenchymal cells. Using single nuclei RNA-sequencing, we defined changes in gene expression associated with inflammation at 1-day post-wounding (dpw) in mouse skin. Compared to keratinocytes and myeloid cells, we detected enriched expression of pro-inflammatory genes in fibroblasts associated with deeper layers of the skin. In particular, SCA1+ fibroblasts were enriched for numerous chemokines, including CCL2, CCL7, and IL33 compared to SCA1-fibroblasts. Genetic deletion of Ccl2 in fibroblasts resulted in fewer wound bed macrophages and monocytes during injury-induced inflammation with reduced revascularization and re-epithelialization during the proliferation phase of healing. These findings highlight the important contribution of deep skin fibroblast-derived factors to injury-induced inflammation and the impact of immune cell dysregulation on subsequent tissue repair.

cell biology↗

bioRxiv2024

Discovering genotype-phenotype relationships with machine learning and the Visual Physiology Opsin Database (VPOD)

BackgroundPredicting phenotypes from genetic variation is foundational for fields as diverse as bioengineering and global change biology, highlighting the importance of efficient methods to predict gene functions. Linking genetic changes to phenotypic changes has been a goal of decades of experimental work, especially for some model gene families including light-sensitive opsin proteins. Opsins can be expressed in vitro to measure light absorption parameters, including {lambda}max - the wavelength of maximum absorbance - which strongly affects organismal phenotypes like color vision. Despite extensive research on opsins, the data remain dispersed, uncompiled, and often challenging to access, thereby precluding systematic and comprehensive analyses of the intricate relationships between genotype and phenotype. ResultsHere, we report a newly compiled database of all heterologously expressed opsin genes with {lambda}max phenotypes called the Visual Physiology Opsin Database (VPOD). VPOD_1.0 contains 864 unique opsin genotypes and corresponding {lambda}max phenotypes collected across all animals from 73 separate publications. We use VPOD data and deepBreaks to show regression-based machine learning (ML) models often reliably predict {lambda}max, account for non-additive effects of mutations on function, and identify functionally critical amino acid sites. ConclusionThe ability to reliably predict functions from gene sequences alone using ML will allow robust exploration of molecular-evolutionary patterns governing phenotype, will inform functional and evolutionary connections to an organisms ecological niche, and may be used more broadly for de-novo protein design. Together, our database, phenotype predictions, and model comparisons lay the groundwork for future research applicable to families of genes with quantifiable and comparable phenotypes. Key PointsO_LIWe introduce the Visual Physiology Opsin Database (VPOD_1.0), which includes 864 unique animal opsin genotypes and corresponding {lambda}max phenotypes from 73 separate publications. C_LIO_LIWe demonstrate that regression-based ML models can reliably predict {lambda}max from gene sequence alone, predict non-additive effects of mutations on function, and identify functionally critical amino acid sites. C_LIO_LIWe provide an approach that lays the groundwork for future robust exploration of molecular-evolutionary patterns governing phenotype, with potential broader applications to any family of genes with quantifiable and comparable phenotypes. C_LI

bioinformatics↗

bioRxiv2023

Beneficial metabolic effects of PAHSAs depend on the gut microbiota in diet-induced obese mice

Dietary lipids play an essential role in regulating the function of the gut microbiota and gastrointestinal tract, and these luminal interactions contribute to mediating host metabolism. PAHSAs are a class of lipids with anti-diabetic and anti-inflammatory properties, but whether the gut microbiota contributes to their beneficial effects on host metabolism is unknown. Here, we report that treating high fat diet (HFD)-fed germ-free mice with PAHSAs does not improve insulin sensitivity. However, transfer of feces from PAHSA-treated, but not Vehicle-treated, chow-fed mice increases insulin-sensitivity in HFD-fed germ free mice. Thus, the gut microbiota is necessary for and can transmit the insulin-sensitizing effects of PAHSAs in HFD-fed germ-free mice. Functional analyses of the cecal metagenome and lipidome of PAHSA-treated mice identified multiple lipid species that associate with the gut commensal Bacteroides thetaiotaomicron (Bt) and with insulin sensitivity resulting from PAHSA treatment. Bt supplementation in HFD-fed female mice prevented weight gain, reduced adiposity, improved glucose tolerance, fortified the colonic mucus barrier and reduced systemic inflammation versus chow-fed controls, effects that were not observed in HFD-fed male mice. Furthermore, ovariectomy partially reversed the beneficial Bt effects on host metabolism, indicating a role for sex hormones in mediating probiotic effects. Altogether, these studies highlight the fact that lipids can modulate the gut microbiota resulting in improvement in host metabolism and that PAHSA-induced changes in the microbiota result in at least some of their insulin-sensitizing effects in female mice.

physiology↗

bioRxiv2021

Differential expression of single-cell RNA-seq data using Tweedie models

SO_SCPLOWUMMARYC_SCPLOWThe performance of computational methods and software to identify differentially expressed genes in single-cell RNA-sequencing (scRNA-seq) has been shown to be influenced by several factors, including the choice of the normalization method used and the choice of the experimental platform (or library preparation protocol) to profile gene expression in individual cells. Currently, it is up to the practitioner to choose the most appropriate differential expression (DE) method out of over 100 DE tools available to date, each relying on their own assumptions to model scRNA-seq data. Here, we propose to use generalized linear models with the Tweedie distribution that can flexibly capture a large dynamic range of observed scRNA-seq data across experimental platforms induced by heavy tails, sparsity, or different count distributions to model the technological variability in scRNA-seq expression profiles. We also propose a zero-inflated Tweedie model that allows zero probability mass to exceed a traditional Tweedie distribution to model zero-inflated scRNA-seq data with excessive zero counts. Using both synthetic and published plate- and droplet-based scRNA-seq datasets, we performed a systematic benchmark evaluation of more than 10 representative DE methods and demonstrate that our method (Tweedieverse) outperforms the state-of-the-art DE approaches across experimental platforms in terms of statistical power and false discovery rate control. Our open-source software (R package) is available at https://github.com/himelmallick/Tweedieverse.

bioinformatics↗

bioRxiv2021

{-}Multivariable Association Discovery in Population-scale Meta-omics Studies

It is challenging to associate features such as human health outcomes, diet, environmental conditions, or other metadata to microbial community measurements, due in part to their quantitative properties. Microbiome multi-omics are typically noisy, sparse (zero-inflated), high-dimensional, extremely non-normal, and often in the form of count or compositional measurements. Here we introduce an optimized combination of novel and established methodology to assess multivariable association of microbial community features with complex metadata in population-scale observational studies. Our approach, MaAsLin 2 (Microbiome Multivariable Associations with Linear Models), uses general linear models to accommodate a wide variety of modern epidemiological studies, including cross-sectional and longitudinal designs, as well as a variety of data types (e.g. counts and relative abundances) with or without covariates and repeated measurements. To construct this method, we conducted a large-scale evaluation of a broad range of scenarios under which straightforward identification of meta-omics associations can be challenging. These simulation studies reveal that MaAsLin 2s linear model preserves statistical power in the presence of repeated measures and multiple covariates, while accounting for the nuances of meta-omics features and controlling false discovery. We also applied MaAsLin 2 to a microbial multi-omics dataset from the Integrative Human Microbiome (HMP2) project which, in addition to reproducing established results, revealed a unique, integrated landscape of inflammatory bowel disease (IBD) across multiple time points and omics profiles.

microbiology↗

bioRxiv2018

netome: a computational framework for metabolite profiling and omics network analysis

SummaryAdvances in metabolomics technologies have enabled comprehensive analyses of associations between metabolites and human disease and have provided a means to study biochemical pathways and processes in detail using model systems. Liquid chromatography tandem mass spectrometry (LC-MS) is an analytical technique commonly used by metabolomics labs to measure hundreds of metabolites of known identity and thousands of \"peaks\" from yet to be identified compounds that are tracked by their measured masses and chromatographic retention times. netome is a computational framework that provides tools for analyzing processed LC-MS data. In this framework, we develop and provide various computational resources including individual software modules to inspect and adjust trends in raw data, align unknown peaks between separately acquired data sets, and to remove redundancies in nontargeted LC-MS data arising from multiple ionization products of a single metabolite. These tools are deployed through computing resources such as web servers and virtual machines with detailed documentation in order to support researchers.\n\nAvailability and implementationnetome is publicly available with extensive documentation and support via issue tracker at https://broadinstitute.github.io/netome under the MIT license. netome includes a set of computational methods that have been designed to execute quality control and post-raw data processing tasks for metabolomics data (e.g. scaling and clustering metabolite abundances), as well as statistical association testing in a network manner (e.g., testing relationship between metabolites and microbes). Each individual tool is available with source code, workshop-oriented documentation which includes instructions for installation and using tools with demonstration examples, and a web server with all services. We also provide a complete image of the netome package with all pre-installed dependencies and support for Google Compute Engine and Amazon EC2. All tools and related services are maintained, and upon new developments, new modules will be added to the environment.\n\nContactrah@broadinstitute.org, clary@broadinstitute.org\n\nSupplementary informationSupplementary data are available at Bioinformatics online.

bioinformatics↗