Biology subjects

Pouryahya, M.

Publications and source records attributed to Pouryahya, M..

12 recordsLinked to original sources

bioRxiv2026

Predicting Antibody Self-Association with Sequence Structure Fusion Models: The Central Role of CSI-BLI in Early Developability Screening

Antibody-based biologics are expanding rapidly, yet challenges in development from self-association, high viscosity, aggregation, and unfavorable clearance underscore the need for accurate in silico screening. Clone self-interaction biolayer interferometry (CSI-BLI) is a plate-based, low-material assay of weak, reversible self-association that serves as an early proxy for high-concentration viscosity and a complementary predictor of in vivo clearance. In a 246-mAb panel, CSI-BLI moderately correlates with viscosity; further, in hFcRn Tg32 mice (41 antibodies), CSI-BLI strongly associates with clearance. Here, we present an end-to-end framework that distinguishes high versus low self-interacting clones (CSI-BLI class) by coupling a fine-tuned protein language model (ESM-2) with residue-aligned 3D context from AlphaFold-predicted structures encoded as residue graphs. Disentangled multi-stream attention fuses sequence content, chain-aware positional information, and structural signals to capture spatially proximate interactions that are distant in sequence. Edit-distance-controlled splits across 1499 IgGs and 988 VHHs assess generalization. The structure-aware model achieves the highest hold-out performance (VHH-Fc F1 = 0.76; IgG F1 = 0.57), while a sequence-only disentangled variant outperforms a standard PLM baseline without structural inputs. Complementary biophysical feature-based models, built from AlphaFold structures and sequence/structure-derived physicochemical descriptors with cluster-aware selection, deliver robust, interpretable performance (VHH F1 = 0.72; IgG F1 = 0.57), with SHAP analyses highlighting charge/dipole, hydrophobicity, and aggregation-propensity drivers across CDRs and frameworks. This interaction-aware sequence-structure framework, supported by interpretable feature models, is extensible to other developability endpoints and broader protein classification tasks where joint modeling of language-derived representations and residue-level geometry is advantageous.

bioinformatics↗

bioRxiv2026

ACCURATE PREDICTION OF ASPARAGINE DEAMIDATION IN BIOLOGICS USING ADVANCED MACHINE LEARNING MODELS

The spontaneous deamidation of asparagine residues remains a major obstacle to the stability and efficacy of protein therapeutics. Currently available models in the literature for predicting deamidation liabilities can suffer from limited generalizability, likely due to biases such as sequence similarity within datasets. In this study, we built machine learning models using protein language models (e.g., ESM2) and graph neural networks (GNNs), trained on a comprehensive dataset of 591 asparagine sites from over 105 protein molecules. To address the critical issue of data leakage, we implemented a peptide grouping strategy yielding more accurate estimates of model performance for novel deamidation sites. Our analysis shows that, when sequence similarity bias is controlled, protein language models match traditional feature-based models that use amino acid composition, k-mers, PSSMs, and predicted secondary structure/solvent accessibility, while offering substantial computational advantages. Additionally, our GNN-based pipeline further increases prediction accuracy by up to 8% compared to language model-only tools and delivers a 15-25% improvement over motif-based approaches. This methodological framework enables more reliable and rapid in-silico prediction of deamidation liabilities, potentially reducing costly late-stage interventions in protein therapeutic development and is generalizable to the modeling of additional protein post-translational modifications.

bioengineering↗

bioRxiv2025

Accelerating Antibody Development: Sequence and Structure-Based Models for Predicting Developability Properties through Size Exclusion Chromatography

Experimental screening for biopharmaceutical developability properties typically relies on resource-intensive, and time-consuming assays such as size exclusion chromatography (SEC). This study highlights the potential of in silico models to accelerate the screening process by exploring sequence and structure-based machine learning techniques. Specifically, we compared surrogate models based on pre-computed features extracted from sequence and predicted structure with sequence-based approaches using protein language models (PLMs) like ESM-2. In addition to different end-to-end fine-tuning strategies for PLM, we have also investigated the integration of the structural information of the antibodies into the prediction pipeline through graph neural networks (GNN). We applied these different methods for predicting protein aggregation propensity using a dataset of approximately 1200 Immunoglobulin G (IgG1) molecules. Through this empirical evaluation, our study identifies the most effective in silico approach for predicting developability properties for SEC assays, thereby adding insights to existing screening efforts for accelerating the antibody development process.

bioinformatics↗

bioRxiv2020

Periodicity Scoring of Time Series Encodes Dynamical Behavior of the Tumor Suppressor p53

In this paper, we analyze the dynamical behavior of the tumor suppressor protein p53, an essential player in the cellular stress response, which prevents a cell from dividing if severe DNA damage is present. When this response system is malfunctioning, e.g. due to mutations in p53, uncontrolled cell proliferation may lead to the development of cancer. Understanding the behavior of p53 is thus crucial to prevent its failing. It has been shown in various experiments that periodicity of the p53 signal is one of the main descriptors of its dynamics, and that its pulsing behavior (regular vs. spontaneous) indicates the level and type of cellular stress. In the present work, we introduce an algorithm to score the local periodicity of a given time series (such as the p53 signal), which we call Detrended Autocorrelation Periodicity Scoring (DAPS). It applies pitch detection (via autocorrelation) on sliding windows of the entire time series to describe the overall periodicity by a distribution of localized pitch scores. We apply DAPS to the p53 time series obtained from single cell experiments and establish a correlation between the periodicity scoring of a cells p53 signal and the number of cell division events. In particular, we show that high periodicity scoring of p53 is correlated to a low number of cell divisions and vice versa. We show similar results with a more computationally intensive state-of-the-art periodicity scoring algorithm based on topology known as Sw1PerS. This correlation has two major implications: It demonstrates that periodicity scoring of the p53 signal is a good descriptor for cellular stress, and it connects the high variability of p53 periodicity observed in cell populations to the variability in the number of cell division events. AMS subject classification92C42, 92C37, 62M10

bioinformatics↗

bioRxiv2019

Optimal Mass Transport for Robust Texture Analysis

The emerging field of radiomics, which consists of transforming standard-of-care images to quantifiable scalar statistics, endeavors to reveal the information hidden in these macroscopic images. This field of research has found different applications ranging from phenotyping and tumor classification to outcome prediction and treatment planning. Texture analysis, which often consists of reducing spatial texture matrices to summary scalar features, has been shown to be important in many of the latter applications. However, as pointed out in many studies, some of the derived texture statistics are strongly correlated and tend to contribute redundant information; and are also sensitive to the parameters used in their computation, e.g., the number of gray intensity levels. In the present study, we propose first to consider texture matrices, with an emphasis on gray-level co-occurrence matrix (GLCM), as a non-parametric multivariate objects. The proposed modeling approach avoids evaluating redundant and strongly correlated features and also prevents the feature processing steps. Then, via the Wasserstein distance from optimal mass transport theory, we propose to compare these spatial objects to identify computerized tomography slices with dental artifacts in head and neck cancer. We demonstrate the robustness of the proposed classification approach with respect to the GLCM extraction parameters and the size of the training set. Comparisons with the random forest classifier, which is constructed on scalar texture features, demonstrates the efficiency and robustness of the proposed algorithm.

bioengineering↗

bioRxiv2019

Reproducibility test of radiomics using network analysis and Wasserstein K-means algorithm

PurposeTo construct robust and validated radiomic predictive models, the development of a reliable method that can identify reproducible radiomic features robust to varying image acquisition methods and other scanner parameters should be preceded with rigorous validation. Due to the property of high correlation present between radiomic features, we hypothesize that reproducible radiomic features across different datasets that are obtained from different image acquisition settings preserve some level of connectivity between features in the form of a network.\n\nMethodsWe propose a regularized partial correlation network to identify robust and reproducible radiomic features. This approach was tested on two radiomic feature sets generated with two different reconstruction methods from a cohort of 47 lung cancer patients. The commonality of the resulting two networks was assessed. A largest common network component from the two networks was tested on phantom data consisting of 5 cancer samples. We further propose a novel K-means algorithm coupled with the optimal mass transport (OMT) theory to cluster samples. This approach following the regularized partial correlation analysis was tested on computed tomography (CT) scans from 77 head and neck cancer patients that were downloaded from The Cancer Imaging Archive (TCIA) and validated on CT scans from 83 head and neck cancer patients treated at our institution.\n\nResultsCommon radiomic features were found in relatively large network components between the resulting two partial correlation networks from a cohort of 47 lung cancer patients. The similarity of network components in terms of the common number of radiomic features was statistically significant. For phantom data, the Wasserstein distance on a largest common network component from the lung cancer data was much smaller than the Wasserstein distance on the same network using random radiomic features, implying the reliability of those radiomic features present in the network. Further analysis using the proposed Wasserstein K-means algorithm on TCIA head and neck cancer data showed that the resulting clusters separate tumor subsites and this was validated on our institution data.\n\nConclusionsWe showed that a network-based analysis enables identifying reproducible radiomic features. This was validated using phantom data and external data via the Wasserstein distance metric and the proposed Wasserstein K-means method.

biophysics↗

bioRxiv2019

Network-based clustering for drug sensitivity prediction in cancer cell lines

The study of large-scale pharmacogenomics provides an unprecedented opportunity to develop computational models that can accurately predict large cohorts of cell lines and drugs. In this work, we present a novel method for predicting drug sensitivity in cancer cell lines which considers both cell line genomic features and drug chemical features. Our network-based approach combines the theory of optimal mass transport (OMT) with machine learning techniques. It starts with unsupervised clustering of both cell line and drug data, followed by the prediction of drug sensitivity in the paired cluster of cell lines and drugs. We show that prior clustering of the heterogenous cell lines and structurally diverse drugs significantly improves the accuracy of the prediction. In addition, it facilities the interpretability of the results and identification of molecular biomarkers which are significant for both clustering of the cell lines and predicting the drug response.

bioinformatics↗

bioRxiv2019

Hierarchical Network Exploration using Gaussian MixtureModels

We present a framework based on optimal mass transport to construct, for a given network, a reduction hierarchy which can be used for interactive data exploration and community detection. Given a network and a set of numerical data samples for each node, we calculate a new computationally-efficient comparison metric between Gaussian Mixture Models, the Gaussian Mixture Transport distance, to determine a series of merge simplifications of the network. If only a network is given, numerical samples are synthesized from the network topology. The method has its basis in the local connection structure of the network, as well as the joint distribution of the data associated with neighboring nodes.\n\nThe analysis is benchmarked on networks with known community structures. We also analyze gene regulatory networks, including the PANTHER curated database and networks inferred from the GTEx lung and breast tissue RNA profiles. Gene Ontology annotations from the EBI GOA database are ranked and superimposed to explain the salient gene modules. We find that several gene modules related to highly specific biological processes are well-coordinated in such tissues. We also find that 18 of the 50 genes of the PAM50 breast-tumor prognostic signature appear among the highly coordinated genes in a single gene module, in both the breast and lung samples. Moreover these 18 are precisely the subset of the PAM50 recently identified as the basal-like markers.

bioinformatics↗

bioRxiv2019

A Novel Integrative Multiomics Method Reveals a Hypoxia-Related Subgroup of Breast Cancer with Significantly Decreased Survival

The remarkable growth of multi-platform genomic profiles has led to the multiomics data integration challenge. The effective integration of such data provides a comprehensive view of the molecular complexity of cancer tumors and can significantly improve clinical out-come predictions. In this study, we present a novel network-based integration method of multiomics data as well as a clustering technique involving the Wasserstein (Earth Movers) distance from the theory of optimal mass transport. We applied our proposed method of integrative Wasserstein-based clustering (iWCluster) to invasive breast carcinoma from The Cancer Genome Atlas (TCGA) project. The subtypes were characterized by the concordant effect of mRNA expression, DNA copy number alteration, and DNA methylation as well as the interaction network connectivity of the gene products. iW-Cluster is substantially more effective in distinguishing clusters with different survival rates as compared to isolated one-dimensional conventional omics analysis. Applying iWCluster to breast cancer TCGA data successfully recovered the known PAM50 molecular subtypes. In addition, iWCluster preserves the gene-specific data, which enables us to interpret the results and perform further analysis of significant genes for a specific cluster. The gene ontology enrichment analysis of significant genes in our substantially low survival sub-group leads to the well-known phenomenon of tumor hypoxia and the transcription factor ETS1 whose expression is induced by hypoxia. Increased expression of ETS1 is associated with an increased risk of recurrence and worse prognosis in breast cancer. Consequently, we believe iWCluster has the potential to discover novel subtypes by accentuating the genes that have concordant multiomics measurements in their interaction network, which are challenging to find without the network inference or with single omics analysis.

genomics↗

bioRxiv2018

Topological Data Analysis of PAM50 and 21-Gene Breast Cancer Assays

We introduce a classification of breast tumors into 7 classes which are more clearly defined by interpretable mRNA signatures along the PAM50 gene set than the 5 traditional PAM50 intrinsic subtypes. Each intrinsic subtype is partially concordant with one of our classes, and the 2 additional classes correspond to division of the classes concordant with the Luminal B and the Normal intrinsic subtypes along expression of the Her2 gene group. Our Normal class shows similarity with the myoepithelial mammary cell phenotype, including TP63 expression (specificity: 80.8% and sensitivity: 82.8%), and exhibits the best overall survival (89.6% at 5 years). Though Luminal A tumors are traditionally considered the least aggressive, our analysis shows that only the Luminal A tumors which are now classified as myoepithelial have this phenotype, while tumors in our luminal class (concordant with Luminal A) may be more aggressive than previously thought. We also find that patients with Basal tumors surviving to 48 months exhibit favorable survival rates when certain markers for B-lymphocytes are present and poor survival rates when they are absent, which is consistent with recent findings.

cancer biology↗

bioRxiv2018

Molecular phenotyping using networks, diffusion, and topology: soft tissue sarcoma

Many biological datasets are high-dimensional yet manifest an underlying order. In this paper, we describe an unsupervised data analysis methodology that operates in the setting of a multivariate dataset and a network which expresses influence between the variables of the given set. The technique involves network geometry employing the Wasserstein distance, global spectral analysis in the form of diffusion maps, and topological data analysis using the Mapper algorithm. The prototypical application is to gene expression profiles obtained from RNA-Seq experiments on a collection of tissue samples, considering only genes whose protein products participate in a known pathway or network of interest. Employing the technique, we discern several coherent states or signatures displayed by the gene expression profiles of the sarcomas in the Cancer Genome Atlas along the p53 signaling network. The signatures substantially recover the leiomyosarcoma, dedifferentiated liposarcoma (DDLPS), and synovial sarcoma histological subtype diagnoses, but they also include a new signature defined by simultaneous activation and inactivation of about a dozen genes, including activation of fibrinolysis inhibitor SERPINE1/PAI and inactivation of p53-family tumor suppressor gene P73 along with cyclin dependent kinase inhibitor 2A CDKN2A/P14ARF.

cancer biology↗

bioRxiv2017

Characterizing Cancer Drug Response andBiological Correlates: A Geometric NetworkApproach

In the present work, we consider a geometric network approach to study common biological features of anticancer drug response. We use for this purpose the panel of 60 human cell lines (NCI-60) provided by the National Cancer Institute. Our study suggests that utilization of mathematical tools for network-based analysis can provide novel insights into drug response and cancer biology. We adopted a discrete notion of Ricci curvature to measure the robustness of biological networks constructed with a pre-treatment gene expression dataset and coupled the results with the GI50 response of the cell lines to the drugs. The link between network robustness and Ricci curvature was implemented using the theory of optimal mass transport. Our hypothesis behind this idea is that robustness in the biological network contributes to tumor drug resistance, thereby enabling us to predict the effectiveness and sensitivity of drugs in the cell lines. Based on the resulting drug response ranking, we assessed the impact of genes that are likely associated with individual drug response. For important genes identified, we performed a gene ontology enrichment analysis using a curated bioinformatics database which resulted in very plausible biological processes associated with drug response across cell lines and cell types from the biological and literature viewpoint. These results demonstrate the potential of using the mathematical network analysis in assessing drug response and in identifying relevant genomic biomarkers and biological processes for precision medicine.

cancer biology↗