Biology subjects

Tastan, O.

Publications and source records attributed to Tastan, O..

At least 19 recordsLinked to original sources

bioRxiv2026

A Systematic Evaluation of Single-Cell Batch Integration Metrics and sBEE: A Robust New Metric

BackgroundSingle-cell RNA sequencing (scRNA-seq) datasets generated across laboratories and experimental conditions often exhibit batch effects that obscure biological variation. Numerous computational methods have been developed to integrate such datasets, making robust benchmarking essential. Evaluation metrics play a central role in assessing integration quality. However, existing metrics each capture only specific aspects of batch mixing and often rely on assumptions that are violated in practice, such as every cell type being present in every batch, balanced batch composition, or simple cluster geometries. Consequently, benchmarking studies frequently report discordant rankings of integration methods, complicating interpretation and method selection. ResultsWe systematically evaluate widely used batch integration metrics using controlled synthetic scenarios that isolate common integration challenges, including imbalanced batch composition, partial cell-type overlap, heterogeneous cluster densities, and varying cluster geometries. Our analysis reveals distinct strengths, limitations, and failure modes, demonstrating that assessments of batch mixing can change substantially across different data characteristics. These observations motivated the development of sBEE (single-cell Batch Effect Evaluator), a metric that assesses batch mixing at the single-cell level using two complementary measures: local neighborhood composition and cross-batch distance relationships. Across diverse simulated scenarios, sBEE produces stable, interpretable evaluations and remains robust to the failure modes identified in existing metrics. We further validate these findings on multiple real-world scRNA-seq datasets. ConclusionsOur work provides a comprehensive characterization of the behavior of widely used batch integration metrics and introduces sBEE, a robust metric that combines complementary neighborhood and distance-based information to provide more reliable assessments of batch mixing. By identifying the conditions under which existing metrics succeed or fail and providing a unified alternative, sBEE enables more consistent benchmarking of single-cell integration methods. Code and datasets are available at https://github.com/tastanlab/sBEE.

bioinformatics↗

bioRxiv2025

DARKIN: A zero-shot benchmark for phosphosite-dark kinase association using protein language models

MotivationProtein Language Models (pLMs) have emerged as powerful tools for capturing the intricate information encoded in protein sequences, facilitating various downstream protein prediction tasks. With numerous pLMs available, there is a critical need for diverse benchmarks to systematically evaluate their performance across biologically relevant tasks. Here, we introduce DARKIN, a zero-shot classification benchmark designed to assign phosphosites to understudied kinases, termed dark kinases. Kinases, which catalyze phosphorylation, are central to cellular signaling pathways. While phosphoproteomics enables the large-scale identification of phosphosites, determining the cognate kinase responsible for the phosphorylation event remains an experimental challenge. ResultsIn DARKIN, we prepared training, validation, and test folds that respect the zero-shot nature of this classification problem, incorporating stratification based on kinase groups and sequence similarity. We evaluated multiple pLMs using two zero-shot classifiers: a novel, training-free k-NN-based method, and a bilinear classifier. Our findings indicate that ESM, ProtT5-XL, and SaProt exhibit superior performance on this task. DARKIN provides a challenging benchmark for assessing pLM efficacy and fosters deeper exploration of under-characterized (dark) kinases by offering a biologically relevant test bed. ImplementationThe DARKIN benchmark data and the scripts for generating additional splits are publicly available at: https://github.com/tastanlab/darkin Contactotastan@sabanciuniv.edu Supplementary informationSupplementary data are available at Bioinformatics online.

bioinformatics↗

bioRxiv2025

One-Hot News: Drug Synergy Models Take a Shortcut

Combinatorial drug therapy holds great promise for tackling complex diseases, but the vast number of possible drug combinations makes exhaustive experimental testing infeasible. Computational models have been developed to guide experimental screens by assigning synergy scores to drug pair-cell line combinations, where they take input structural and chemical information on drugs and molecular features of cell lines. The premise of these models is that they leverage this biological and chemical information to predict synergy measurements. In this study, we demonstrate that replacing drug and cell line representations with simple one-hot encodings results in comparable or even slightly improved performance across diverse published drug combination models. This unexpected finding suggests that current models use these representations primarily as identifiers and exploit covariation in the synergy labels. Our synthetic data experiments show that models can learn from the true features; however, when drugs and cell lines recur across drug-drug-cell triplets, this repeating structure impairs feature-based learning. While the current synergy prediction models can still aid in prioritizing drug pairs within a panel of tested drugs and cell lines, our results highlight the need for better strategies to learn from intended features and generalize to unseen drugs and cell lines.

bioinformatics↗

bioRxiv2025

DeepSynBa: Actionable Drug Combination Prediction with Complete Dose-Response Profiles

Many cancer monotherapies demonstrate limited clinical efficacy, making combination therapies a relevant treatment strategy. The extensive number of potential drug combinations and context-specific response profiles complicates the prediction of drug combination responses. Existing computational models are typically trained to predict a single aggregated synergy score, which summarises drug responses across different dosage combinations, such as Bliss or Loewe scores. This oversimplification of the drug-response surface leads to high prediction uncertainty and limited actionability, as these models fail to distinguish between potency and efficacy. We introduce DeepSynBa, an actionable model that predicts the complete dose-response matrix of drug pairs instead of relying on an aggregated synergy score. This is achieved by predicting parameters describing the response surface as an intermediate layer in the model. Evaluated on the NCI-ALMANAC and the ONeil datasets, DeepSynBa outperforms the state-of-the-art methods in the dose-response matrix prediction task across most evaluation scenarios, including testing on novel drug combinations, cell lines, and drugs, across nine different tissue types. We also show that DeepSynBa yields reliable synergy score predictions. More importantly, DeepSynBa can predict drug combination responses across different dosages for untested combinations. The intermediate dose-response parameter layer enables the separation of efficacy from potency, informing the selection of dosage ranges that optimise efficacy while limiting off-target toxicity in experimental screens. The predictive capability and the downstream actionability make DeepSynBa a powerful tool for advancing drug combination research beyond the limitations of the current approaches. The code and the dataset for DeepSynBa are available at https://github.com/hikuru/DeepSynBa.

bioinformatics↗

bioRxiv2024

PHACTboost: A Phylogeny-aware Boosting Algorithm to Compute the Pathogenicity of Missense Mutations

Most algorithms that are used to predict the effects of variants rely on evolutionary conservation. However, a majority of such techniques compute evolutionary conservation by solely using the alignment of multiple sequences while overlooking the evolutionary context of substitution events. We had introduced PHACT, a scoring-based pathogenicity predictor for missense mutations that can leverage phylogenetic trees, in our previous study. By building on this foundation, we now propose PHACTboost, a gradient boosting tree-based classifier that combines PHACT scores with information from multiple sequence alignments, phylogenetic trees, and ancestral reconstruction. The results of comprehensive experiments on carefully constructed sets of variants demonstrated that PHACTboost can outperform 40 prevalent pathogenicity predictors reported in the dbNSFP, including conventional tools, meta-predictors, and deep learning-based approaches as well as state-of-the-art tools, AlphaMissense, EVE, and CPT-1. The superiority of PHACTboost over these methods was particularly evident in case of hard variants for which different pathogenicity predictors offered conflicting results. We provide predictions of 219 million missense variants over 20,191 proteins. PHACTboost can improve our understanding of genetic diseases and facilitate more accurate diagnoses.

bioinformatics↗

bioRxiv2023

Deep Phenotyping of Sleep in Drosophila

Animal behavior depends on internal state. While subtle movements can signify significant changes in internal state, computational methods for analyzing these "microbehaviors" are lacking. Here, we present FlyVISTA, a machine-learning platform to characterize microbehaviors in freely-moving flies, which we use to perform deep phenotyping of sleep. This platform comprises a high-resolution closed-loop video imaging system, coupled with a deep-learning network to annotate 35 body parts, and a computational pipeline to extract behaviors from high-dimensional data. FlyVISTA reveals the distinct spatiotemporal dynamics of sleep-associated microbehaviors in flies. We further show that stimulation of dorsal fan-shaped body neurons induces micromovements, not sleep, whereas activating R5 ring neurons triggers rhythmic proboscis extension followed by persistent sleep. Importantly, we identify a novel microbehavior ("haltere switch") exclusively seen during quiescence that indicates a deeper sleep stage. These findings enable the rigorous analysis of sleep in Drosophila and set the stage for computational analyses of microbehaviors.

neuroscience↗

bioRxiv2023

SUMOnet: Deep Sequential Prediction of SUMOylation Sites

SUMOylation is a reversible post-translational protein modification in which SUMOs (small ubiquitin-like modifiers) covalently attach to a specific lysine residue of the target protein. This process is vital for many cellular events. Aberrant SUMOylation is associated with several diseases, including Alzheimers, cancer, and diabetes. Therefore, accurate identification of SUMOylation sites is essential to understanding cellular processes and pathologies that arise with their disruption. We present three deep neural architectures, SUMOnets, that take the peptide sequence centered on the candidate SUMOylation site as input and predict whether the lysine could be SUMOylated. Each of these models, SUMOnet-1, -2, and -3 relies on different compositions of deep sequential learning architectural units, such as bidirectional Gated Recurrent Units(biGRUs) and convolutional layers. We evaluate these models on the benchmark dataset with three different input peptide representations of the input sequence. SUMOnet-3 achieves 75.8% AUPR and 87% AUC scores, corresponding to approximately 5% improvement over the closest state-of-the-art SUMOylation predictor and 16% improvement over GPS-SUMO, the most widely adopted tool. We also evaluate models on a challenging subset of the test data formed based on the absence and presence of known SUMOylation motifs. Even though the performances of all methods degrade in these cases, SUMOnet-3 remains the best predictor in these challenging cases. Availability and ImplementationThe SUMOnet-3 framework is available as an open-source project and a Python library at https://github.com/berkedilekoglu/SUMOnet.

bioinformatics↗

bioRxiv2023

From Cell-Lines to Cancer Patients: Personalized Drug Synergy Prediction

Combination drug therapies are effective treatments for cancer. However, the genetic heterogeneity of the patients and exponentially large space of drug pairings pose significant challenges for finding the right combination for a specific patient. Current in silico prediction methods can be instrumental in reducing the vast number of candidate drug combinations. However, existing powerful methods are trained with cancer cell line gene expression data, which limits their applicability in clinical settings. While synergy measurements on cell lines models are available at large scale, patient-derived samples are too few to train a complex model. On the other hand, patient-specific single-drug response data are relatively more available. In this work, we propose a deep learning framework, Personalized Deep Synergy Predictor (PDSP), that enables us to use the patient-specific single drug response data for customizing patient drug synergy predictions. PDSP is first trained to learn synergy scores of drug pairs and their single drug presonses for a given cell line using drug structures and large scale cell line gene expression data. Then, the model is fine-tuned for patients with their patient gene expression data and associated single drug response measured on the patient ex vivo samples. In this study, we evaluate PDSP on data from three leukemia patients and observe that it improves the prediction accuracy by 27% compared to models trained on cancer cell line data. PDSP is built and available at https://github.com/hikuru/PDSP

bioinformatics↗

bioRxiv2021

MuDCoD: Multi-Subject Community Detection in Personalized Dynamic Gene Networks from Single Cell RNA Sequencing

MotivationWith the wide availability of single-cell RNA-seq (scRNA-seq) technology, population-scale scRNA-seq datasets across multiple individuals and time points are emerging. While the initial investigations of these datasets tend to focus on standard analysis of clustering and differential expression, leveraging the power of scRNA-seq data at the personalized dynamic gene co-expression network level has the potential to unlock subject and/or time-specific network-level variation, which is critical for understanding phenotypic differences. Community detection from co-expression networks of multiple time points or conditions has been well-studied; however, none of the existing settings included networks from multiple subjects and multiple time points simultaneously. To address this, we develop MuDCoD for multi-subject community detection in personalized dynamic gene networks from scRNA-seq. MuDCoD builds on the spectral clustering framework and promotes information sharing among the networks of the subjects as well as networks at different time points. It clusters genes in the personalized dynamic gene networks and reveals gene communities that are variable or shared not only across time but also among subjects. ResultsEvaluation and benchmarking of MuDCoD against existing approaches reveal that MuDCoD effectively leverages apparent shared signals among networks of the subjects at individual time points, and performs robustly when there is no or little information sharing among the networks. Applications to population-scale scRNA-seq datasets of human-induced pluripotent stem cells during dopaminergic neuron differentiation and CD4+ T cell activation indicate that MuDCoD enables robust inference for identifying time-varying personalized gene modules. Our results illustrate how personalized dynamic community detection can aid in the exploration of subject-specific biological processes that vary across time. AvailabilityMuDCoD is publicly available at https://github.com/bo1929/MuDCoD as a Python package. Implementation includes simulation and real-data experiments together with extensive documentation. Contactkeles@stat.wisc.edu, otastan@sabanciuniv.edu Supplementary informationSupplementary data are available at Bioinformatics online.

genomics↗

bioRxiv2020

Uncovering complementary sets of variants for the prediction of quantitative phenotypes

MotivationGenome-wide association studies show that variants in individual genomic loci alone are not sufficient to explain the heritability of complex, quantitative phenotypes. Many computational methods have been developed to address this issue by considering subsets of loci that can collectively predict the phenotype. This problem can be considered a challenging instance of feature selection in which the number of dimensions (loci that are screened) is much larger than the number of samples. While currently available methods can achieve decent phenotype prediction performance, they either do not scale to large datasets or have parameters that require extensive tuning. ResultsWe propose a fast and simple algorithm, Macarons, to select a small, complementary subset of variants by avoiding redundant pairs that are in linkage disequilibrium. Our method features two interpretable parameters that control the time/performance trade-off without requiring parameter tuning. In our computational experiments, we show that Macarons consistently achieves similar or better prediction performance than state-of-the-art selection methods while having a simpler premise and being at least 2 orders of magnitude faster. Overall, Macarons can seamlessly scale to the human genome with ~107 variants in a matter of minutes while taking the dependencies between the variants into account. ConclusionMacarons can offer a reasonable trade-off between phenotype predictivity, runtime and the complementarity of the selected subsets. The framework we present can be generalized to other high-dimensional feature selection problems within and beyond biomedical applications. AvailabilityMacarons is implemented in Matlab and the source code is available at: https://github.com/serhan-yilmaz/macarons

bioinformatics↗

bioRxiv2020

Identifying Cross-Cancer Similar Patients via a Semi-Supervised Deep Clustering Approach

The treatment decisions for a cancer patient are typically based on the patients diagnosed cancer type. With the characterization of cancer tumors at the molecular level, there have been reports of patients being similar despite being diagnosed with different cancer types. Motivated from these observations, we aim at discovering cross-cancer patients, which we define as patients whose tumors are more similar to patient tumors diagnosed with another cancer type. We develop DeepCrossCancer to identify cross-cancer patients that always co-cluster with the other patient from another cancer type. The input to DeepCrossCancer is the transcriptomic profiles of the patient tumors, the age, and sex of the patient. To solve the clustering problem, we use a semi-supervised deep learning-based clustering method in which the clustering task is supervised by cancer type labels and the survival times of the patients. Applying the method to patient data from nine different cancers, we discover 20 cross-cancer patients that consistently co-cluster. By analyzing the predictive genes of the cross-cancer patients and other genomic information available for the patient such as somatic mutations and copy number variations, we identify striking genomic similarities across these patients providing support. The detection of cross-cancer patients opens up possibilities for transferring clinical decisions across patients at a single patient level. The source code is available at github.com/tastanlab/DeepCrossCancer

bioinformatics↗

bioRxiv2020

Detecting Misannotated Long Non-coding RNAs with Training Dynamics of Deep Sequence Classification

Long non-coding RNAs (lncRNAs) are the largest class of non-coding RNAs (ncRNAs). However, recent experimental evidence has shown that some lncRNAs contain small open reading frames (sORFs) that are translated into functional micropeptides. Current methods to detect misannotated lncRNAs rely on ribosome-profiling (ribo-seq) experiments, which are expensive and cell-type dependent. In addition, while very accurate machine learning models have been trained to distinguish between coding and non-coding sequences, little attention has been paid to the increasing evidence about the incorrect ground-truth labels of some lncRNAs in the underlying training datasets. We present a framework that leverages deep learning models training dynamics to determine whether a given lncRNA transcript is misannotated. Our models achieve AUC scores > 91% and AUPR > 93% in classifying non-coding vs. coding sequences while allowing us to identify possible misannotated lncRNAs present in the dataset. Our results overlap significantly with a set of experimentally validated misannotated lncRNAs as well as with coding sORFs within lncRNAs found by a ribo-seq dataset. The general framework applied here offers promising potential for use in curating datasets used for training coding potential predictors and assisting experimental efforts in characterizing the hidden proteome encoded by misannotated lncRNAs. Source code is available at https://github.com/nabiafshan/DetectingMisannotatedLncRNAs.

genomics↗

bioRxiv2020

Functional Enrichment Analysis of Deregulated Long Non-Coding RNAs in Cancer Based on their Genomic Neighbors

The dysregulation of long non-coding RNAs (lncRNAs) expressions has been implicated in cancer. Since most of the lncRNAs are not functionally characterized well, investigating the set of perturbed lncRNAs are is challenging. Existing methods that inspect lncRNAs functionally rely on the co-expressed coding genes, which are far better characterized functionally. LncRNAs can be known to act as transcriptional regulators; they may activate or repress the neighborhoods coding genes on the genome. Based on this, in this work, we aim to analyze the deregulated lncRNAs in cancer by taking into account their ability to regulate nearby loci on the genome. We perform functional analysis on differentially expressed lncRNAs for 28 different cancers considering their adjacent coding genes. We identify that some deregulated lncRNAs are cancer-specific, but a substantial number of lncRNAs are shared across cancers. Next, we assess the similarities of the cancer types based on the functional enrichment of the deregulated lncRNA sets. We find some cancers are very similar in the functions and biological processes related to the deregulated lncRNAs. We observe that some of the cancers for which we find similarity can be linked through primary, metastatic site relations. We investigate the similarity of enriched functional terms for the deregulated lncRNAs and the mRNAs. We further assess the enriched functions similarity to the functions and processes that the known cancer driver genes take place. We believe that our methodology help to understand the impact of the lncRNAs in cancer functionally.

bioinformatics↗

bioRxiv2020

PRER: A Patient Representation with Pairwise Relative Expression of Proteins on Biological Networks

Changes in protein and gene expression levels are often used as features to predictive models such as survival prediction. A common strategy to aggregate information on individual proteins is to integrate the expression information with biological networks. We propose a novel patient representation in this work where we integrate proteins expression levels with the protein-protein interaction (PPI) networks. Patient representation with PRER (Pairwise Relative Expressions with Random walks) uses the neighborhood of a protein to capture the dysregulation patterns in protein abundance. Specifically, PRER computes a feature vector for a patient by comparing the source proteins protein expression level with other proteins levels in its neighborhood. This neighborhood of the source protein is derived using a biased random-walk strategy on the network. We test PRERs performance through a survival prediction task in 10 different cancers using random forest survival models. PRER representation yields a statistically significant predictive performance in 9 out of 10 cancer types when compared to a representation based on individual protein expression. We also identify important proteins that are not important in the models trained with the expression values but emerge as predictive in models trained with PRER features. The set of identified relations provides a valuable collection of biomarkers with high prognostic value. PRER representation can be used for other complex diseases and prediction tasks that use molecular expression profiles as input. PRER is freely available at: https://github.com/hikuru/PRER

systems biology↗

bioRxiv2020

MatchMaker: A Deep Learning Framework for Drug Synergy Prediction

Drug combination therapies have been a viable strategy for the treatment of complex diseases such as cancer due to increased efficacy and reduced side effects. However, experimentally validating all possible combinations for synergistic interaction even with high-throughout screens is intractable due to vast combinatorial search space. Computational techniques can reduce the number of combinations to be evaluated experimentally by prioritizing promising candidates. We present MatchMaker that predicts drug synergy scores using drug chemical structure information and gene expression profiles of cell lines in a deep learning framework. For the first time, our model utilizes the largest known drug combination dataset to date, DrugComb. We compare the performance of MatchMaker with the state-of-the-art models and observe up to ~ 20% correlation and ~ 40% mean squared error (MSE) improvements over the next best method. We investigate the cell types and drug pairs that are relatively harder to predict and present novel candidate pairs. MatchMaker is built and available at https://github.com/tastanlab/matchmaker

bioinformatics↗

bioRxiv2019

DeepSide: A Deep Learning Framework for Drug Side Effect Prediction

Drug failures due to unforeseen adverse effects at clinical trials pose health risks for the participants and lead to substantial financial losses. Side effect prediction algorithms have the potential to guide the drug design process. LINCS L1000 dataset provides a vast resource of cell line gene expression data perturbed by different drugs and creates a knowledge base for context specific features. The state-of-the-art approach that aims at using context specific information relies on only the high-quality experiments in LINCS L1000 and discards a large portion of the experiments. In this study, our goal is to boost the prediction performance by utilizing this data to its full extent. We experiment with 5 deep learning architectures. We find that a multi-modal architecture produces the best predictive performance among multi-layer perceptron-based architectures when drug chemical structure (CS), and the full set of drug perturbed gene expression profiles (GEX) are used as modalities. Overall, we observe that the CS is more informative than the GEX. A convolutional neural network-based model that uses only SMILES string representation of the drugs achieves the best results and provides 13.0% macro-AUC and 3.1% micro-AUC improvements over the state-of-the-art. We also show that the model is able to predict side effect-drug pairs that are reported in the literature but was missing in the ground truth side effect dataset. DeepSide is available at http://github.com/OnurUner/DeepSide.

bioinformatics↗

bioRxiv2019

PAMOGK: A Pathway Graph Kernel based Multi-Omics Clustering Approach for Discovering Cancer Patient Subgroups

Accurate classification of patients into molecular subgroups is critical for the development of effective therapeutics and for deciphering what drives these subgroups to cancer. The availability of multi-omics data cat-alogs for large cohorts of cancer patients provides multiple views into the molecular biology of the tumors with unprecedented resolution. We develop PAMOGK (Pathway based Multi Omic Graph Kernel clustering) that not only integrates multi-omics patient data with existing biological knowledge on pathways. We develop a novel graph kernel that evaluates patient similarities based on a single molecular alteration type in the context of a pathway. To corroborate multiple views of patients evaluated by hundreds of pathways and molecular alteration combinations, we use multi-view kernel clustering. Applying PAMOGK to kidney renal clear cell carcinoma (KIRC) patients results in four clusters with significantly different survival times (p-value = 1.24e-11). When we compare PAMOGK to eight other state-of-the-art multi-omics clustering methods, PAMOGK consistently outperforms these in terms of its ability to partition KIRC patients into groups with different survival distributions. The discovered patient subgroups also differ with respect to other clinical parameters such as tumor stage and grade, and primary tumor and metastasis tumor spreads. The pathways identified as important are highly relevant to KIRC. PAMOGK is available at github.com/tastanlab/pamogk

bioinformatics↗

bioRxiv2019

miRCoop: Identifying Cooperating miRNAs via Kernel Based Interaction Tests

Although miRNAs can cause widespread changes in expression programs, single miRNAs typically induce mild repression on their targets. Cooperativity is reported as one strategy to overcome this constraint. Expanding the catalog of synergistic miRNAs is critical for understanding gene regulation and for developing miRNA-based therapeutics. In this study, we develop miRCoop to identify synergistic miRNA pairs that have weak or no repression on the target mRNA, but when bound together, induce strong repression. miRCoop uses kernel-based interaction tests together with miRNA and mRNA target information. We apply our approach to kidney tumor patient data and identify 66 putative triplets. For 64 of these triplets, there is at least one common transcription factor that potentially regulates all participating RNAs of the triplet, supporting a functional association among them. Furthermore, we find that triplets are enriched for certain biological processes that are relevant to kidney cancer. Some of the synergistic miRNAs are very closely encoded in the genome, hinting a functional association among them. We believe miRCoop can aid our understanding of the complex regulatory interactions in different health and disease states of the cell and can help in designing miRNA-based therapies. Matlab code for the methodology is provided in https://github.com/guldenolgun/miRCoop.

bioinformatics↗