Biology subjects

Samee, M. A. H.

Publications and source records attributed to Samee, M. A. H..

At least 19 recordsLinked to original sources

bioRxiv2026

SPARC: A Graph-based Optimization Framework for Directional Trajectory Reconstruction Across Ordered Single-Cell Conditions

Single-cell transcriptomics has enabled systematic profiling of cellular states across ordered biological contexts, including developmental stages, treatment phases, disease progression, and anatomical compartments. A central challenge is to reconstruct trajectories that respect the directionality imposed by biology or experimental design. Existing trajectory inference methods reconstruct cell-state progressions from latent-space geometry but do not enforce external biological ordering during graph construction, yielding biologically inadmissible transitions. An emerging paradigm of optimal-transport (OT) approaches partially addresses this limitation by incorporating experimental ordering into probabilistic state-to-state correspondences, yet their pairwise formulation cannot resolve whether a given state is an intermediate state or a terminal state along a multi-step progression. In multi-timepoint settings, OT typically estimates couplings only betweenadjacent timepoints and then chains these locally solved couplings to approximate long-range trajectories without a global optimization across all conditions simultaneously. Here we present SPARC, a graph-based optimization framework that quantifies similarity in a shared high-dimensional latent space and reconstruct directional trajectories under biological constraints. Global shortest-path optimization over this graph yields progression routes, from which SPARC derives path-based pseudotime identifies bottlenecks clusters, and detects gene temporal behavior. SPARC was evaluated across three complementary settings representing distinct trajectory-inference challenges. Its application to paired primary and lung metastatic osteosarcoma samples allows us to be the first to propose a "cross-organ bone-like microenvironment" hypothesis, in which osteoclastogenic signaling establishes a bone-like remodeling niche within the pulmonary metastatic lesion that promotes osteoclast differentiation and activity. The findings are independently recoverable in human osteosarcoma Visium HD spatial transcriptomics.

bioinformatics↗

bioRxiv2026

Kernelized approach enables explainable gene prioritizations for complex traits

Genome-wide association studies (GWAS) have identified numerous variant-trait associations; yet, assigning effector genes to GWAS loci remains challenging. Similarity-based machine-learning methods, such as PoPS, prioritize effector genes from shared functional profiles among trait-relevant genes. These models assign a prioritization score for each gene and nominate a single effector gene within a GWAS locus. However, the scores provide limited insight into why a gene was prioritized or whether the nomination is biologically plausible. To address this gap, we introduce Kernelized Polygenic Priority Score, K-PoPS, a kernelized reformulation of PoPS that enables gene-centric explanations by decomposing each prediction into contributions from training genes. For each prioritized gene, K-PoPS reports top contributor genes and an anchor score that quantifies support from a user-defined set of trait-relevant genes. Across 38 Pan-UK Biobank traits, the full-feature OLS implementation underlying K-PoPS improved closest-gene enrichment relative to default PoPS for 26 of 37 evaluable traits. Across 25 traits with curated anchor sets, predictions supported by anchor scores were more enriched for closest-gene proxies than unsupported predictions. When applying to blood level apolipoprotein B, K-PoPS nominated SCARB1 over UBC gene, and further provided convincing explanations that support this prediction. Using explanation evidence, K-PoPS identified multiple plausible effector genes within a dilated cardiomyopathy locus, contrary to the parsimonious assumption. In summary, K-PoPS provides a post hoc framework for examining and interpreting GWAS effector-gene nominations.

genomics↗

bioRxiv2025

Modeling Temporal Dependencies and Feature Interactions Reveal Novel Clinical and Molecular Insights into Alzheimer's Disease Progression

ObjectiveAlzheimers Disease remains a major public health challenge, requiring insights into feature interactions and temporal trends of feature importance. Community-wide data science competitions such as the TADPOLE Challenge provide platforms to benchmark predictive models using ADNI datasets. While top-performing models achieve accurate predictions, they often leave mechanistic questions unresolved. We introduce a framework that separately models static feature interactions and temporal dynamics, enabling complementary insights from longitudinal AD data. Materials and MethodsWe analyzed XGBoost with TreeSHAP for feature interactions and RNN-AD with Integrated Gradients for temporal trends. This two-branch design allows XGBoost to capture nonlinear cross-sectional interactions, while RNN captures evolving, time-dependent influences. The attributions are fused into a combined importance map. ResultsOur framework showed agreement between TreeSHAP and IG, highlighting FAQ, CDRSB, ADAS13, MMSE, and RAVLT variants as the most consistently important features across both branches. Temporal attribution analysis revealed stage-dependent trends: in CN, features such as DX:CN, FAQ, and RAVLT_immediate increased in importance with longer prediction horizons; in MCI, MidTemp and WholeBrain gained importance; and in AD, FAQ remained dominant. Feature-interaction analysis identified strong clinical-clinical interactions and secondary clinical-molecular interactions involving hippocampal and entorhinal volumes. DiscussionCombining interaction and temporal trends showed that RAVLT_immediate, FAQ, and DX-based features were the only markers consistently influential across both dimensions, indicating stable, cross-validated predictors of Alzheimers disease progression. Feature importance in AD prediction is dynamic, with early-time features often most influential. These insights support personalized monitoring, adaptive modeling, and mechanistic interpretability, enhancing patient-specific interventions and trial design. ConclusionThis work highlights feature interactions and temporal trends in AD prediction models, offering insights for personalized treatments and patient-specific trial designs. Our framework provides stable, cross-validated explanations that unify structural and temporal importance, enhancing trustworthiness and mechanistic interpretability in AD modeling.

neuroscience↗

bioRxiv2025

Hypoxia Imaging as a Radiomics Signature in Tumors Using Novel 19F-Eu-Based Contrast Agents for Magnetic Resonance Imaging

The ability to accurately identify and characterize regions of hypoxia has been an active area of study due to the biological ramifications of low oxygen tension in different pathological conditions, including inflammation, infections, wound healing, cardiovascular conditions, kidney and pulmonary diseases, hepatic and neurological toxicities, and cancer. Although hypoxia contributes a significant role in these conditions, the ability to accurately and instantaneously monitor the presence of hypoxia in vivo and correlate comprehensive analysis of hypoxia-dependent molecular signatures in response to treatments regimens is lacking. With the advent of hypoxia-responsive contrast agents for magnetic resonance imaging (MRI), including the recent development and characterization by our team of a novel probe that involves both 19F and EuII, the capability to integrate direct hypoxia imaging in real-time with new technologies that enable spatial transcriptomic profiling has become a possibility. To assess the capability of this agent for studying hypoxia, we used osteosarcoma as a model. In this preliminary study we demonstrate two major results: First, we show that convection-enhanced delivery (CED) is a reliable and robust methodology to distribute MRI contrast agents throughout a tumor. Second, we show that integration of direct hypoxia-detecting imaging modalities and spatial biology enables real-time in vivo insights into biology and identification of biomarkers directly applicable to disease development and response to therapy. The ability to identify and define hypoxia-mediated biology using direct rather than indirect MRI methods has extremely significant implications in the care of a wide-range of pathological conditions. Consequently, the framework outlined in these preliminary studies is applicable to other pathological conditions and provides the basis for direct in vivo hypoxia detection, monitoring, and biological analyses with translational applications to patient care and management.

cancer biology↗

bioRxiv2025

Supervised Factorization to Associate Spatial Transcriptomics with Complementary Molecular Readouts

Spatial Transcriptomics enables studying gene expression data within spatial context of tissues. Yet understanding how spatial molecular phenomena influence transcriptional patterns remains a key challenge. We propose a novel supervised Non-negative Matrix Factorization (NMF) framework, where supervision is selectively and explicitly applied to guide the learning of a supervised spatial factor. This distinguishes our method from prior approaches by enforcing spatial alignment only on a targeted component of the factorization, enabling biologically interpretable associations between gene expression and spatial molecular events. This approach also enables the identification of genes whose expression patterns are spatially correlated with molecular events of interest. Applied to datasets involving Alzheimers Disease (AD) and Myocardial Infarction (MI), our method successfully discovered supervised spatial factor associated with disease related signal. In the case of Alzheimers Disease (AD), we have presented a spatial decay model to represent how the influence of amyloid-beta plaque signals diminishes with distance, and used this as a supervision signal during matrix factorization. Applied across both disease contexts, our method successfully identified biologically meaningful gene sets associated with disease progression. By ranking genes based on their contribution to the supervised spatial factor, the framework highlights candidate genes potentially involved in disease-related processes.

bioinformatics↗

bioRxiv2025

ORANGE: A Machine Learning Approach for Modeling Tissue-Specific Aging from Transcriptomic Data

Despite aging being a fundamental biological process which profoundly influences health and disease, the interplay between tissue-specific aging and mortality remains underexplored. This study applies machine learning on GTEx transcriptomic data to model tissue-specific biological ages across 12 different types of tissues and introduces an age-gap metric to quantify deviations from the chronological age. We use several modeling techniques optimized with three feature selection strategies: Pearson correlation, age-related differentially expressed genes, and tissue-enriched genes (expressed at least fourfold higher in a specific tissue). Among these, Pearson correlation combined with elastic net regression yields the best performance, with models achieving an average RMSE of 6.44 years and an R2 of 0.64. To quantify deviations from chronological age relative to the population, we train neural networks to regress predicted ages against chronological ages, and subtract their outputs from the predicted ages to calculate a metric which we call the age-gap. Age-gap statistics reveal significant tissue-specific aging patterns, identifying extreme agers and correlations between extreme aging and mortality. About 20% of subjects are found to exhibit extreme aging in one tissue, while 1% show multi-organ aging. Further analysis reveals that accelerated aging in specific tissues correlates with with greater risk of death from illness. These findings greatly emphasize the role of transcriptomics in aging research and its implications for health and longevity.

bioinformatics↗

bioRxiv2024

Quantifying Pathological Progression from Single-Cell Data

The surge in single-cell datasets and reference atlases has enabled the comparison of cell states across conditions, yet a gap persists in quantifying pathological shifts from healthy cell states. To address this gap, we introduce single-cell Pathological Shift Scoring (scPSS) which provides a statistical measure for how much a "query" cell from a diseased sample has been shifted away from a reference group of healthy cells. In scPSS, The distance of a query cell to its k-th nearest reference cell is considered as its pathological shift score. Euclidean distances in the top n principal component space of the gene expressions are used to measure distances between cells. The p-value of a query pathological shift score belonging to the null distribution of intra-reference cell shift scores provides a statistical significance measure of the query cell being in the reference cell group. This makes our method both simple and statistically rigorous. Comparative evaluations against a state-of-the-art contrastive variational inference model, modified for shift scores, demonstrate our methods accuracy and efficiency. Additionally, we have also shown that the aggregation of cell-level pathological scores from scPSS can be used to predict health conditions at the individual level.

genomics↗

bioRxiv2024

GAGER: gene regulatory network assisted gene expression restoration

Gene regulatory networks are crucial for cellular function, and disruptions in transcription factor (TF) regulation often lead to diseases. However, identifying TFs to transition a source cell state to a desired target state remains challenging. We present a method to identify key TFs whose perturbation can restore gene expressions in a source state to target levels. Its effectiveness is demonstrated on datasets from yeast TF knockouts, cardiomyocytes from hypoplastic left heart syndrome patients, and mouse models of neurodegeneration. The method accurately identifies knocked-out TFs in the yeast dataset. In the cardiomyocyte dataset, it pinpoints TFs that, though not differentially expressed in many cases, exert significant regulatory influence on downstream differentially expressed genes. Finally, in the mouse model dataset, it identifies disease stage-specific TFs, improving similarity between healthy and diseased states at various time points. Unlike traditional approaches relying on differential expression analysis, our method uses network-based prioritization for more targeted and biologically relevant TF selection. These findings highlight its potential as a therapeutic tool for precise TF targeting to normalize gene expressions in diseased states.

bioinformatics↗

bioRxiv2024

Reference-free Analysis of scRNA-seq Data Reveals Elevated rRNA and mtRNA Transcription during Neurogenesis in Axolotl

Analysis of single-cell RNA-seq data is typically performed on a gene expression matrix estimated by aligning reads to a reference transcriptome. However, this approach is difficult to apply to organisms with no or incomplete reference transcriptomes. In addition, events deviating from the reference remain undetected. Here we present a reference-free method to analyze single-cell RNA-seq data based on k-mers. We assess the performance of our method on a metastatic renal cell carcinoma dataset and find that it is largely able to capture differentially expressed genes. We then analyze a recently generated dataset to study neurogenesis in Axolotl and observe increased levels of transcription of rRNA and mtRNA during neurogenesis as well as a miRNA with previously predicted links to neuronal development. We also detect lncRNAs and intron retention in heart disease-related genes in diseased cardiomyocytes in an analysis of a congenital heart disease dataset.

bioinformatics↗

bioRxiv2023

Imputation approaches and quality standards for single-cell epigenetic age predictions

DNA methylation describes the addition of methyl groups, often between CpG dinucleotides. Single-cell bisulfite sequencing technologies allow the measurement of DNA methylation levels within individual cells. Epigenetic clocks are statistical models for computing biological age from DNA methylation levels, and have been used for detecting age variations in various disease contexts. However, there have been no attempts to apply epigenetic clocks to single-cell methylation data in humans. Thus, we questioned whether pre-built epigenetic clocks could be applied to single-cell methylation data; if so, how can we perform data quality control and imputation. We concluded that 1) linear regression-based epigenetic clocks can be applied to bisulfite-sequencing data, 2) data quality control can be used to reach the desired level of prediction accuracy, 3) first-principle imputation strategies could be used for missing data on selected CpG methylation sites, and 4) machine learning-based imputation tools could be used for accuracy-based age predictions. We built the first training-free, reference data-free framework for estimating epigenetic age in human single-cells, which would provide a foundation for future single-cell methylation-based age analyses.

bioinformatics↗

bioRxiv2023

Probabilistic clustering of cells using single-cell RNA-seq data

Single-cell RNA sequencing is a modern technology for analyzing cellular heterogeneity. A key challenge is to cluster a heterogeneous sample of different cell types into multiple different homogeneous groups. Although there exist a number of clustering methods, they do not perform well consistently across various datasets. Moreover, most of them are not based on probabilistic approaches making it difficult to assess uncertainties in their results. Therefore, in spite of having large cell atlases, it is often quite difficult to map cells to types. In addition, many of the methods require prior knowledge such as marker gene information for each type. Also due to technological limitations, dropouts of gene expressions may occur in the data which is not taken into account in other methods. Here we present a probabilistic method named CellHorizon for clustering scRNA-seq data that is based on a generative model, handles dropouts and works without any prior marker gene information. Experiments reveal that our method outperforms current state-of-the-art methods overall on six gold standard datasets.

bioinformatics↗

bioRxiv2023

CC-Tempo: A cell-cell communication aware temporal model of cellular dynamics

Delineating the mechanisms underlying cell state changes is key to gaining insights into organismal development and disease prognosis. Various methods have been proposed to study cellular differentiation and cell fate specification. However, they either do not incorporate temporal information or do not consider the vital role of intercellular communication in cellular differentiation and cell fate determination. Furthermore, many of these methods lack interpretability, making it difficult to identify the critical genes and pathways that influence the differentiation process. Here we propose CC-Tempo, a cell-cell communication-aware model of cellular dynamics that leverages intercellular communication scores and can help identify important genes and pathways crucial for different stages of differentiation in various lineages. While previous studies have indicated that scRNA-seq data alone may not suffice for accurately predicting cell fates, CC-Tempo demonstrates that incorporating intercellular communication significantly enhances the performance of such models. CC-Tempo can predict the significance of genes and pathways at different stages of the differentiation process. By perturbing these genes in silico, CC-Tempo reveals their efficacy for manipulating cell fate, which can be crucial for defining efficient reprogramming factors.

developmental biology↗

bioRxiv2023

Hippo-deficient cardiac fibroblasts differentiate into osteochondroprogenitors

Cardiac fibrosis, a common pathophysiology associated with various heart diseases, occurs from the excess deposition of extracellular matrix (ECM)1. Cardiac fibroblasts (CFs) are the primary cells that produce, degrade, and remodel ECM during homeostasis and tissue repair2. Upon injury, CFs gain plasticity to differentiate into myofibroblasts3 and adipocyte-like4,5 and osteoblast-like6 cells, promoting fibrosis and impairing heart function7. How CFs maintain their cell state during homeostasis and adapt plasticity upon injury are not well defined. Recent studies have shown that Hippo signalling in CFs regulates cardiac fibrosis and inflammation8-11. Here, we used single-nucleus RNA sequencing (snRNA-seq) and spatially resolved transcriptomic profiling (ST) to investigate how the cell state was altered in the absence of Hippo signaling and how Hippo-deficient CFs interact with macrophages during cardiac fibrosis. We found that Hippo-deficient CFs differentiate into osteochondroprogenitors (OCPs), suggesting that Hippo restricts CF plasticity. Furthermore, Hippo-deficient CFs colocalized with macrophages, suggesting their intercellular communications. Indeed, we identified several ligand-receptor pairs between the Hippo-deficient CFs and macrophages. Blocking the Hippo-deficient CF-induced CSF1 signaling abolished macrophage expansion. Interestingly, blocking macrophage expansion also reduced OCP differentiation of Hippo-deficient CFs, indicating that macrophages promote CF plasticity.

developmental biology↗

bioRxiv2023

CRISPR-DIPOFF: An Interpretable Deep LearningApproach for CRISPR Cas-9 Off-Target Prediction

CRISPR Cas-9 is a groundbreaking gene-editing tool that harnesses bacterial defense systems to alter DNA sequences accurately. This innovative technology holds vast promise in multiple domains like biotechnology, agriculture, and medicine. However, such power does not come without its own peril, and one such issue is the potential for unintended modifications (Off-Target), which highlights the need for accurate prediction and mitigation strategies. Though previous studies have demonstrated improvement in Off-Target prediction capability with the application of deep learning, they often struggle with the precision-recall trade-off, limiting their effectiveness and do not provide proper interpretation of the complex decision-making process of their models. To address these limitations, we have thoroughly explored deep learning networks, particularly the recurrent neural network (RNN) and transformer based models, leveraging their established success in handling sequence data. Furthermore, we have employed genetic algorithm for hyperparameter tuning to optimize these models performance. The results from our experiments demonstrate significant performance improvement compared to the current state-of-the-art in Off-Target prediction, highlighting the efficacy of our approach. Furthermore, leveraging the power of the integrated gradient method, we make an effort to interpret our models resulting in a detailed analysis and understanding of the underlying factors that contribute to Off-Target predictions, in particular the presence of two sub-regions in the seed region of sgRNA which extends the established biological hypothesis of Off-Target effects. To the best of our knowledge, our model can be considered as the first model combining high efficacy, interpretability, and a desirable balance between precision and recall.

bioinformatics↗

bioRxiv2022

Identifying Genes with Location Dependent Noise Variance in Spatial Transcriptomics Data

Spatial transcriptomics (ST) holds the promise to identify the existence and extent of spatial variation of gene expression in complex tissues. Such analyses could help identify gene expression signatures that distinguish between healthy and disease samples. Existing tools to detect spatially variable genes assume a constant noise variance across location. This assumption might miss important biological signals when the variance could change across spatial locations, e.g., in the tumor microenvironment. In this paper, we propose NoVaTeST, a framework to identify genes with location-dependent noise variance in ST data. NoVaTeST can model gene expression as a function of spatial location with a spatially variable noise. We then compare the model to one with constant noise to detect genes that show significant spatial variation in noise. Our results show genes detected by NoVaTeST provide complimentary information to existing tools while providing important biological insights.

systems biology↗

bioRxiv2022

Gene panel design for spatial transcriptomics with prioritized gene sets

A fundamental limitation of the emerging single-cell spatial transcriptomics (sc-ST) technologies is their panel size. Being based on fluorescence in situ hybridization, an sc-ST dataset can profile only a pre-determined panel of a few hundred genes. This often forces biologists to build panels from only the marker genes of different cell types and forgo other genes of interest, e.g., genes encoding ligand-receptor complexes or genes in specific pathways. We propose scGIST- a deep neural network that designs sc-ST panels through constrained feature selection. On four datasets, scGIST outperformed alternative methods in terms of cell type detection accuracy. Moreover, unlike other methods, scGIST allows genes of interest to be prioritized for inclusion in the panel while staying within the its size constraint. We demonstrate through diverse use cases that scGIST includes large fractions of prioritized genes without compromising cell type prediction efficacy making it a valuable addition to sc-STs algorithmic toolbox.

bioinformatics↗

bioRxiv2022

Epstein-Barr virus nuclear antigen proteins deploy diverse mechanisms to bind the human genome

Viral transcription factors (vTFs) are known to bind the human genome and impact critical biological processes. However, a vTF can infect multiple cell-types, and the extent to which they deploy the same mechanisms for DNA binding across different cell-types is poorly understood. We used the Epstein-Barr virus Nuclear Antigen Proteins 2 and 3 (EBNA2 and EBNA3) to address this gap, the two widely studied vTFs associated with cancers, as case studies. We analyze multiple ChIP-seq datasets of these vTFs from different human cell-lines using a state-of-the-art convolutional neural network model. We show that each vTF uses both cell-type specific and cell-type independent cofactors to bind the human genome. Interestingly, each vTF requires [~]10 cofactors consistently across all binding peaks. Such dependency on a large number of cofactors is in contrast to human TFs since human TFs typically co-bind with 2-3 cofactors. Our de novo motif finding approach also reveals novel sequence motifs at the ChIP peaks of these vTFs. These critical signals were missed by previous studies doing enrichment analysis using known human TFs. Finally, we find that although a vTF impacts similar biological processes across cell-types, it also impacts distinct biological processes in a cell-type specific manner. Our study provides a roadmap for integrative analysis of vTF binding and pinpointing their diverse mechanisms of targeting the human genome across different human cell-types.

genomics↗

bioRxiv2022

DeepBend: An Interpretable Model of DNA Bendability

The bendability of genomic DNA impacts chromatin packaging and protein-DNA binding. However, beyond a handful of known sequence motifs, such as certain dinucleotides and poly(A)/poly(T) sequences, we do not have a comprehensive understanding of the motifs influencing DNA bendability. Recent high-throughput technologies like Loop-Seq offer an opportunity to address this gap but the lack of accurate and interpretable machine learning models still poses a significant challenge. Here we introduce DeepBend, a convolutional neural network model built as a visible neural network where we designed the convolutions to directly capture the motifs underlying DNA bendability and how their periodic occurrences or relative arrangements modulate bendability. Through extensive benchmarking on Loop-Seq data, we show that DeepBend consistently performs on par with alternative machine learning and deep learning models while giving an extra edge through mechanistic interpretations. Besides confirming the known motifs of DNA bendability, DeepBend also revealed several novel motifs and showed how the spatial patterns of motif occurrences influence bendability. DeepBends genome-wide prediction of bendability further showed how bendability is linked to chromatin conformation and revealed the motifs controlling bendability of topologically associated domains and their boundaries.

bioinformatics↗