Biology subjects

Sivadasan, N.

Publications and source records attributed to Sivadasan, N..

5 recordsLinked to original sources

bioRxiv2026

Mechanistically informed adaptive dosing for cancer immunotherapy using AI-guided decision making

Optimizing dose and schedule remains a central challenge in oncology drug development, particularly for immunotherapies where fixed dosing regimens often fail to account for patient specific heterogeneity in tumor-immune dynamics. Here, we present a hybrid quantitative systems pharmacology-reinforcement learning-Monte Carlo Tree Search (QSP-RL-MCTS) framework for personalized immunotherapy dosing that formulates dose selection as a sequential decision-making problem. The approach integrates a mechanistic QSP model of prostate cancer immunotherapy, transcriptomics informed virtual patient populations and data driven AI system comprising reinforcement learning and Monte Carlo tree search. Reinforcement learning is used to learn adaptive generalized dosing policies that optimize treatment outcomes across the population, while Monte Carlo Tree Search provides forward-looking evaluation of RL predicted dosing trajectories to refine patient-specific decisions. On benchmarking against fixed dosing regimens of ipilimumab, the remission rate of the proposed model (95.2%) was comparable to the highest fixed dosing regimen of 10 mg/kg per dose while the median total dose (72 mg/kg) of the proposed model designed regimen was comparable to the lowest fixed dosing regimen of 3 mg/kg per dose. The model is generalizable across different dosing protocols and can be extended to predict optimal dose under different therapeutic scenarios. Analysis of the learned dosing trajectories enables stratification of patients into distinct response groups and identifies drug activity rate as the dominant determinant of long-term treatment outcome. These results demonstrate how mechanistically guided artificial intelligence can transform population-level dose optimization into patient-specific, biologically interpretable treatment strategies for precision immuno-oncology.

systems biology↗

bioRxiv2025

Mitosis detection in the wild using detection transformers

BackgroundIdentification of mitotic cells and its down-stream analysis, is an important parameter in understanding the pathology of cancer, predicting response to chemotherapy and overall survival. However, their reliable detection remains challenging due to morphological overlap with other cellular structures, resulting in variability and high levels of false positives. Current artificial intelligence (AI) algorithms further face limitations when confronted with tissue heterogeneity and often underperform in non-tumor, inflamed, or necrotic regions. To address these challenges, Track 1 of MIDOG 2025 challenge expands the scope of mitotic figure detection to include all tissue regions, encompassing both hotspot and non-hotspot areas, thereby promoting real-world clinical applicability. MethodsWe propose a generalized deep learning based mitotic detection model (MD model) for robust and accurate detection of mitotic figures using the MIDOG-2025 challenge dataset. Our model enables robust and accurate detection of mitotic figures and demonstrates strong generalization to domain shifts arising from diverse histological regions, variations in scanners, tumor subtypes and laboratory protocols. ResultsOur approach showed a consistent performance with an F1-score of 0.7769 with a high recall of 0.8222. Our approach outperforms the baselines and generalizes well across the tumor types on the preliminary test set. ConclusionOur approach makes the final predictions with reduced false positives and improved detection accuracy. It generalizes well to address the domain shifts caused by diverse histological regions across the different tumor types among others.

pathology↗

bioRxiv2023

A Robust Deep Learning Approach for Joint Nuclei Detection and Cell Classification in Pan-Cancer Histology Images

Advanced image processing methods have shown promise in computational pathology, including the extraction of crucial microscopic features from histology images. Accurate detection and classification of cell nuclei from whole-slide images (WSI) play a crucial role in capturing the molecular and morphological landscape of the tissue sample. They enable widespread downstream applications, including cancer diagnosis, prognosis, and discovery of novel markers. Robust nuclei detection and classification are challenging due to the high intra-class variability and inter-class similarity of the microscopic morphological features. This is further compounded by the domain shift arising due to the variability in tissue types, staining protocols, and image acquisition. Motivated by the ability of the recent deep learning techniques to learn complex patterns in a biasfree manner, we develop a novel and robust deep learning model TransNuc, based on vision transformers, for simultaneous detection and classification of cell nuclei from H&E stained WSI. We benchmarked TransNuc on the comprehensive Open Pan-cancer Histology Dataset (PanNuke), sampled from over 20,000 WSI, comprising 19 different tissue types and five clinically important cell classes, namely, Neoplastic, Epithelial, Inflammatory, Connective, and Dead cells. TransNuc exhibited superior performance compared to the state-of-theart, including Hover-Net and Micro-Net. TransNuc was able to learn robust feature representations and thereby perform consistently better for the abundant classes such as neoplastic, and the under-represented classes such as dead cells. Similar performance gains were also obtained for epithelial and connective classes that have a significant inter-class morphological similarity.

pathology↗

bioRxiv2022

MAGE: Strain Level Profiling of MetagenomeSamples

Metagenomic profiling from sequencing data aims to disentangle a microbial sample at lower ranks of taxonomy, such as species and strains. Deep taxonomic profiling involving accurate estimation of strain level abundances aids in precise quantification of the microbial composition, which plays a crucial role in various downstream analyses. Existing tools primarily focus on strain/subspecies identification and limit abundance estimation to the species level. Abundance quantification of the identified strains is challenging and remains largely unaddressed by the existing approaches. We propose a novel algorithm MAGE (Microbial Abundance GaugE), for accurately identifying constituent strains and quantifying strain level relative abundances. For accurate profiling, MAGE uses read mapping information and performs a novel local searchbased profiling guided by a constrained optimization based on maximum likelihood estimation. Unlike the existing approaches that often rely on strain-specific markers and homology information for deep profiling, MAGE works solely with read mapping information, which is the set of target strains from the reference collection for each mapped read. As part of MAGE, we provide an alignment-free and kmer-based read mapper that uses a compact and comprehensive index constructed using FM-index and R-index. We use a variety of evaluation metrics for validating abundances estimation quality. We performed several experiments using a variety of datasets, and MAGE exhibited superior performance compared to the existing tools on a wide range of performance metrics.

bioinformatics↗

bioRxiv2017

Sequence Alignment On Directed Graphs

Genomic variations in a reference collection are naturally represented as genome variation graphs. Such graphs encode common subsequences as vertices and the variations are captured using additional vertices and directed edges. The resulting graphs are directed graphs possibly with cycles. Existing algorithms for aligning sequences on such graphs make use of partial order alignment (POA) techniques that work on directed acyclic graphs (DAG). For this, acyclic extensions of the input graphs are first constructed through expensive loop unrolling steps (DAGification). Also, such graph extensions could have considerable blow up in their size and in the worst case the blow up factor is proportional to the input sequence length. We provide a novel alignment algorithm V-ALIGN that aligns the input sequence directly on the input graph while avoiding such expensive DAGification steps. V-ALIGN is based on a novel dynamic programming formulation that allows gapped alignment directly on the input graph. It supports affine and linear gaps. We also propose refinements to V-ALIGN for better performance in practice. In this, the time to fill the DP table has linear dependence on the sizes of the sequence, the graph and its feedback vertex set. We perform experiments to compare against the POA based alignment. For aligning short sequences, standard approaches restrict the expensive gapped alignment to small filtered subgraphs having high similarity to the input sequence. In such cases, the performance of V-ALIGN for gapped alignment on the filtered subgraph depends on the subgraph sizes.

genomics↗