Biology subjects

Eslami, M.

Publications and source records attributed to Eslami, M..

9 recordsLinked to original sources

bioRxiv2025

ICEPIC: A Toolkit to Discover Ice Binding Proteins from Sequence

Ice binding proteins, such as antifreeze proteins (AFPs) and ice nucleation proteins (INPs), are critical for survival in subzero environments and have wide-ranging applications in biotechnology, agriculture, and materials science. Current discovery methods for these proteins are constrained by low throughput and limited datasets that are not conducive for engineering. Here, we present a high-throughput, sequence-based model that leverages contextual embeddings from protein language models to predict ice binding potential, as well as the expression and activity potential of candidate proteins. Using a curated data corpus of over 18,000 ice binding proteins -- far larger than previous datasets -- we fine-tuned a ProtBERT-based model, achieving 99% accuracy for prediction of ice binding potential. Sensitivity analyses through targeted mutagenesis (alanine and threonine substitutions) confirmed the models biological significance, revealing functionally important residues and sequence patterns. Additionally, we developed an expression prediction model that achieved an R2 score of 0.64 and low false-negative rates in identifying highly expressible candidates in Pichia pastoris. An additional regression model trained to predict ice activity as measured by thermal hysteresis achieved an R2 score of at least 0.79 with a clear difference in prediction between ice binding and non-ice binding proteins. Our toolkit advances the predictive accuracy, interpretability, and scalability of ice binding protein discovery, offering a powerful tool for protein engineering in cold-environment applications. Significance StatementIce binding proteins enable organisms to survive freezing temperatures and are essential for applications in cryopreservation, agriculture, and materials science. However, discovering and engineering these proteins has been limited by small datasets and inadequate predictive tools. We developed a machine learning model trained on over 18,000 ice binding protein sequences to predict not only ice binding potential but also protein activity and expression in engineered hosts. This approach integrates advanced protein language models with biological context, enabling faster, more reliable discovery of ice binding proteins. Our platform advances rational protein design for real-world applications in climate resilience, biotechnology, and beyond.

synthetic biology↗

bioRxiv2025

Metadata Harmonization from Biological Datasets with Language Models

Biomedical research faces significant challenges in harmonizing metadata across diverse datasets due to inconsistent labeling and the lack of universally adopted ontologies. Conventional solutions, such as Common Data Elements, face adoption difficulties as they impede scientific progress by requiring researchers to navigate through thousands of standardized terms with subtle variations. Tools such as laboratory information management systems, while designed to enforce standardization, can hinder research progress when their rigid standards conflict with domain-specific documentation needs and evolving research practices. As a result of these challenges, researchers maintain their own annotation systems, leading to disconnected datasets that are difficult to integrate across studies. This study presents a novel approach using large language models to automatically standardize researcher annotations to standards within ontologies. The approach is applied to multiple domains such as oncology, alcohol research, and infectious disease. Data augmentation strategies are presented to align training data with the space of human representations. These strategies generate realistic variations of standard terms to simulate how researchers naturally document their work, especially valuable in domains lacking the extensive terminology mappings needed for training language models. Experiments with fine-tuned GPT-2 variants show up to 96% accuracy on in-dictionary tasks and 17% on out-of-dictionary tasks, outperforming traditional techniques and zero-shot GPT-4o applications. This implies that there can be up to a 96% reduction in metadata standardization labor if a term exists in an ontology. We also show a significant trade-off between domain-specific models versus those that aim to generalize across domains such as infectious disease or alcohol research. While larger models excel at generalization, fine-tuned models consistently outperform on domain-specific terminology. This approach enables more efficient and accurate research data integration across biomedical fields, though out-of-dictionary generalization remains a challenge across all model sizes.

bioinformatics↗

bioRxiv2023

An artificial intelligence method for phenotyping of OCT scans using unsupervised and self-supervised deep learning

Artificial intelligence (AI) has been increasingly used to analyze optical coherence tomography (OCT) images to better understand physiology and genetic architecture of ophthalmic diseases. However, to date, research has been limited by the inability to transfer OCT phenotypes from one dataset to another. In this work, we propose a new AI method for phenotyping and clustering of OCT-derived retinal layer thicknesses using unsupervised and self-supervised methods in a large clinical dataset using glaucoma as a model disease and subsequently transfer our phenotypes to a large biobank. The model includes a deep learning model, manifold learning, and a Gaussian mixture model. We also propose a correlation analysis for the performance evaluation of our model based on Pearson correlation coefficients. Our model was able to identify clinically meaningful OCT phenotypes and successfully transfer phenotypes from one dataset to another. Overall, our results will contribute to stronger research methodologies for future research in OCT imaging biomarkers, augment testing of OCT phenotypes in multiple datasets, and ultimately improve our understanding of pathophysiology and genetic architecture of ocular diseases.

bioinformatics↗

bioRxiv2022

AutoGater: A Weakly Supervised Neural Network Model to Gate Cells in Flow Cytometric Analyses

Flow cytometry is a useful and efficient method for the rapid characterization of a cell population based on the optical and fluorescence properties of individual cells. Ideally, the cell population would consist of only healthy viable cells as dead cells can confound the analysis. Thus, separating out healthy cells from dying and dead cells, and any potential debris, is an important first step in analysis of flow cytometry data. While gating of debris can be conducted using measured optical properties, identifying dead and dying cells often requires utilizing fluorescent stains (e.g. Sytox, a nucleic acid stain that stains cells with compromised cell membranes) to identify cells that should be excluded from downstream analyses. These stains prolong the experimental preparation process and use a flow cytometers fluorescence channels that could otherwise be used to measure additional fluorescent markers within the cells (e.g. reporter proteins). Here we outline a stain-free method for identifying viable cells for downstream processing by gating cells that are dying or dead. AutoGater is a weakly supervised deep learning model that can separate healthy populations from unhealthy and dead populations using only light-scatter channels. In addition, AutoGater harmonizes different measurements of dead cells such as Sytox and CFUs.

bioinformatics↗

bioRxiv2022

Robustness and reproducibility of simple and complex synthetic logic circuit designs using a DBTL loop

Computational tools addressing various components of design-build-test-learn loops (DBTL) for the construction of synthetic genetic networks exist, but do not generally cover the entire DBTL loop. This manuscript introduces an end-to-end sequence of tools that together form a DBTL loop called DART (Design Assemble Round Trip). DART provides rational selection and refinement of genetic parts to construct and test a circuit. Computational support for experimental process, metadata management, standardized data collection, and reproducible data analysis is provided via the previously published Round Trip (RT) test-learn loop. The primary focus of this work is on the Design Assemble (DA) part of the tool chain, which improves on previous techniques by screening up to thousands of network topologies for robust performance using a novel robustness score derived from dynamical behavior based on circuit topology only. In addition, novel experimental support software is introduced for the assembly of genetic circuits. A complete design-through-analysis sequence is presented using several OR and NOR circuit designs, with and without structural redundancy, that are implemented in budding yeast. The execution of DART tested the predictions of the design tools, specifically with regard to robust and reproducible performance under different experimental conditions. The data analysis depended on a novel application of machine learning techniques to segment bimodal flow cytometry distributions. Evidence is presented that, in some cases, a more complex build may impart more robustness and reproducibility across experimental conditions.

synthetic biology↗

bioRxiv2021

Prediction of Whole-Cell Transcriptional Response with Machine Learning

Applications in synthetic and systems biology can benefit from measuring whole-cell response to biochemical perturbations. Execution of experiments to cover all possible combinations of perturbations is infeasible. In this paper, we present the host response model (HRM), a machine learning approach that takes the cell response to single perturbations as the input and predicts the whole cell transcriptional response to the combination of inducers. We find that the HRM is able to qualitatively predict the directionality of dysregulation to a combination of inducers with an accuracy of >90% using data from single inducers. We further find that the use of known prior, known cell regulatory networks doubles the predictive performance of the HRM (an R2 from 0.3 to 0.65). This tool will significantly reduce the number of high-throughput sequencing experiments that need to be run to characterize the transcriptional impact of the combination of perturbations on the host.

bioinformatics↗

bioRxiv2021

Impact of sterilization on a conjugated polymer based bioelectronic patch

Translation into the clinic of organic bioelectronic devices having conjugated polymers as the active material will hinge on their long-term operation in vivo. This will require the device to be subject to clinically approved sterilization techniques without a deterioration in its physical and electronic properties. To date, there remains a gap in the literature addressing the impact of this critical pre-operative procedure on the properties of conjugated polymers. This study aims to address this gap by assessing the physical and electronic properties of a sterilized porous bioelectronic patch having polyaniline as the conjugated polymer. The patch was sterilized by autoclave, ethylene oxide and gamma ({gamma}-) irradiation at 15, 25, and 50 kGy doses. Autoclaving resulted in cracking and macroscopic degradation of the patch, while patches sterilized by {gamma}-irradiation at 50 kGy exhibited reduced mechanical and electronic properties, attributed to chain scission and non-uniform crosslinking caused by the high dose irradiation. Ethylene oxide and {gamma}-irradiation at 15 and 25 kGy sterilization appeared to be the most effective at maintaining the mechanical and electronic properties of the patch, as well as inducing a minimal immune response as revealed by a receding fibrotic capsule after 4 weeks implantation. Our findings pave the way towards closing the gap for the translation of organic bioelectronic devices from acute to long-term in vivo models.

bioengineering↗

bioRxiv2020

Estimating the density of small population of leopard Panthera pardus using multi-session photographic-sampling data

West Asian drylands host a number of threatened large carnivores, including the leopard (Panthera pardus) which is limited to spatially scattered landscapes with generally low primary productivity. While conservation efforts have focused on these areas for several decades, reliable population density estimates are missing. Spatially-explicit capture-recapture (SECR) methodology, incorporating animal movement in density estimates, is widely used to monitor populations of large carnivores. We employed multi-session SECR modeling to estimate the density of a small population of leopard (Panthera pardus) in a mountainous stretch surrounded by deserts in central Iran. During 6724 camera trap nights, we detected eight and five independent leopards in 2012 and 2016 sessions, respectively. The top performing model demonstrated density estimates of 1.6 (95% CI = 0.9-2.9) and 1.0 (95% CI = 0.6-1.6) independent leopards/100 km2 in 2012 and 2016, respectively. Both sex and season had substantial effects on spatial scale ({sigma}), with larger movements for males and during winter. Currently available estimates in arid regions represent some of the lowest densities across the leopard global range. These small populations are vulnerable to demographic stochasticity. Monitoring temporal changes in population density and composition can inform conservation priorities.

ecology↗

bioRxiv2017

afpCOOL: An Accurate Tool for Antifreeze Protein Detection

Various cold-adapted organisms produce antifreeze proteins (AFPs), which prevent to freeze of cell fluids by resisting the growth of the ice crystal. AFPs are currently being recognized in various organisms that are living in extremely low temperatures. AFPs have several important applications in increasing freeze tolerance of plants; maintain the tissue in frozen conditions and producing cold-hardy plants using transgenic technology. Substantial differences in the sequence and structure of the AFPs, pose a challenge for researcher to identify these proteins. In this paper, we proposed a novel method for identifying AFPs using support vector machine (SVM) by incorporating 4 types of features. Results on two benchmark datasets revealed the strength of the proposed method in AFP prediction. Also, according to the results on an independent test set, our method outperformed the current state-of-the-art methods. The further analysis showed the non-satisfactory performance of the BLAST in AFP detection: more than 62% of the BLAST searches have specificity less than 10% and there is no any BLAST search with sensitivity higher than 10%. These results reveal the urgent need for an accurate tool for AFP detection. In addition, the comparison results of the discrimination power of different feature types disclosed that evolutionary features and amino acid composition are the most contributing features in AFP detection. This method has been implemented as a stand-alone tool, namely afpCOOL, for various operating systems to predict AFPs with a user friendly graphical interface.\n\nAvailabilityafpCOOL is freely available at http://bioinf.modares.ac.ir:8080/AFPCOOL/page/afpcool.isp\n\nContactDr Zahiri zahiri@modares.ac.ir

bioinformatics↗