Biology subjects

Jayadeva, J.

Publications and source records attributed to Jayadeva, J..

3 recordsLinked to original sources

bioRxiv2021

Gene expression based inference of drug resistance in cancer

Inter and intra-tumoral heterogeneity are major stumbling blocks in the treatment of cancer and are responsible for imparting differential drug responses in cancer patients. Recently, the availability of large-scale drug screening datasets has provided an opportunity for predicting appropriate patient-tailored therapies by employing machine learning approaches. In this study, we report a predictive modeling approach to infer treatment response in cancers using gene expression data. In particular, we demonstrate the benefits of considering integrated chemogenomics approach, utilizing the molecular drug descriptors and pathway activity information as opposed to gene expression levels. We performed extensive validation of our approach on tissue-derived single-cell and bulk expression data. Further, we constructed several prostate cancer cell lines and xenografts, exposed to differential treatment conditions to assess the predictability of the outcomes. Our approach was further assessed on pan-cancer RNA-sequencing data from The Cancer Genome Atlas (TCGA) archives, as well as an independent clinical trial study describing the treatment journey of three melanoma patients. To summarise, we benchmarked the proposed approach on cancer RNA-seq data, obtained from cell lines, xenografts, as well as humans. We concluded that pathway-activity patterns in cancer cells are reasonably indicative of drug resistance, and therefore can be leveraged in personalized treatment recommendations.

cancer biology↗

bioRxiv2020

Deep learning discerns cancer mutation exclusivity

The exclusivity of a vast majority of cancer mutations remains poorly understood, despite the availability of large amounts of whole genome and exome sequencing data. In clinical settings, this markedly hinders the identification of the previously uncharacterized deleterious mutations due to the unavailability of matched normal samples. We employed state of the art deep learning algorithms for cross-exome learning of mutational embeddings and demonstrated their utility in sequence based detection of cancer-specific Single Nucleotide Variants (SNVs).

bioinformatics↗

bioRxiv2017

FORKS: Finding Orderings Robustly using K-means and Steiner trees

Recent advances in single cell RNA-seq technologies have provided researchers with unprecedented details of transcriptomic variation across individual cells. However, it has not been straightforward to infer differentiation trajectories from such data, due to the parameter-sensitivity of existing methods. Here, we present Finding Orderings Robustly using k-means and Steiner trees (FORKS), an algorithm that pseudo-temporally orders cells and thereby infers bifurcating state trajectories. FORKS, which is a generic method, can be applied to both single-cell and bulk differentiation data. It is a semi-supervised approach, in that it requires the user to specify the starting point of the time course. We systematically benchmarked FORKS and eight other pseudo-time estimation algorithms on six benchmark datasets, and found it to be more accurate, more reproducible, and more memory-efficient than existing methods for pseudo-temporal ordering. Another major advantage of our approach is its robustness - FORKS can be used with default parameter settings on a wide range of datasets.

bioinformatics↗