Biology subjects

Fridley, B.

Publications and source records attributed to Fridley, B..

4 recordsLinked to original sources

bioRxiv2024

iModMix: Integrative Module Analysis for Multi-omics Data

SummaryIntegrative Module Analysis for Multi-omics Data (iModMix) is a biology-agnostic framework that enables the discovery of novel associations across any type of quantitative abundance data, including but not limited to transcriptomics, proteomics, and metabolomics. Instead of relying on pathway annotations or prior biological knowledge, iModMix constructs data-driven modules using graphical lasso to estimate sparse networks from omics features. These modules are summarized into eigenfeatures and correlated across datasets for horizontal integration, while preserving the distinct feature sets and interpretability of each omics type. iModMix operates directly on matrices containing expression or abundances for a wide range of features, including but not limited to genes, proteins, and metabolites. Because it does not rely on annotations (e.g., KEGG identifiers), it can seamlessly incorporate both identified and unidentified metabolites, addressing a key limitation of many existing metabolomics tools. iModMix is available as a user-friendly R Shiny application requiring no programming expertise (https://imodmix.moffitt.org), and as an R package for advanced users (https://github.com/biodatalab/iModMix). The tool includes several public and in-house datasets to illustrate its utility in identifying novel multi-omics relationships in diverse biological contexts. Availability and implementationShiny application: https://imodmix.moffitt.org. The R package and source code: https://github.com/biodatalab/iModMix.

bioinformatics↗

bioRxiv2023

Spatial transcriptomics analysis identifies a unique tumor-promoting function of the meningeal stroma in melanoma leptomeningeal disease

Leptomeningeal disease (LMD) remains a rapidly lethal complication for late-stage melanoma patients. The inaccessible nature of the disease site and lack of understanding of the biology of this unique metastatic site are major barriers to developing efficacious therapies for patients with melanoma LMD. Here, we characterize the tumor microenvironment of the leptomeningeal tissues and patient-matched extra-cranial metastatic sites using spatial transcriptomic analyses with in vitro and in vivo validation. We show the spatial landscape of melanoma LMD to be characterized by a lack of immune infiltration and instead exhibit a higher level of stromal involvement. We show that the tumor-stroma interactions at the leptomeninges activate pathways implicated in tumor-promoting signaling, mediated through upregulation of SERPINA3 at the tumor-stroma interface. Our functional experiments establish that the meningeal stroma is required for melanoma cells to survive in the CSF environment and that these interactions lead to a lack of MAPK inhibitor sensitivity in the tumor. We show that knocking down SERPINA3 or inhibiting the downstream IGR1R/PI3K/AKT axis results in re-sensitization of the tumor to MAPK-targeting therapy and tumor cell death in the leptomeningeal environment. Our data provides a spatial atlas of melanoma LMD, identifies the tumor-promoting role of meningeal stroma, and demonstrates a mechanism for overcoming microenvironment-mediated drug resistance unique to this metastatic site.

cancer biology↗

bioRxiv2021

Summarizing internal dynamics boosts differential analysis and functional interpretation of super enhancers

Super enhancers (SEs) are broad enhancer domains usually containing multiple constituent enhancers that hold elevated activities in gene regulation. Disruption in one or more constituent enhancers causes aberrant SE activities that lead to gene dysregulation in diseases. To quantify SE aberrations, differential analysis is performed to compare SE activities between cell conditions. The state-of-art strategy in estimating differential SEs relies on overall activities and neglect the changes in length and structure of SEs. Here, we propose a novel computational method to identify differential SEs by weighting the combinatorial effects of constituent-enhancer activities and locations (i.e., internal dynamics). In addition to overall activity changes, our method identified four novel classes of differential SEs with distinct enhancer structural alterations. We demonstrate that these structure alterations hold distinct regulatory impact, such as regulating different number of genes and modulating gene expression with different strengths, highlighting the differentiated regulatory roles of these unexplored SE features. When compared to the existing method, our method showed improved identification of differential SEs that were linked to better discernment of cell-type-specific SE activity and functional interpretation. We implement an R package, DASE, to facilitate the use of our method.

bioinformatics↗

bioRxiv2020

Wide and Deep Learning for Automatic Cell Type Identification

Cell type classification is an important problem in cancer research, especially with the advent of single cell technologies. Correctly identifying cells within the tumor microenvironment can provide oncologists with a snapshot of how a patients immune system is reacting to the tumor. Wide deep learning (WDL) is an approach to construct a cell-classification prediction model that can learn patterns within high-dimensional data (deep) and ensure that biologically relevant features (wide) remain in the final model. In this paper, we demonstrate that the use of regularization can prevent overfitting and adding a wide component to a neural network can result in a model with better predictive performance. In particular, we observed that a combination of dropout and{ell} 2 regularization can lead to a validation loss function that does not depend on the number of training iterations and does not experience a significant decrease in prediction accuracy compared to models with{ell} 1, dropout, or no regularization. Additionally, we show WDL can have superior classification accuracy when the training and testing of a model is completed data on that arise from the same cancer type, but from different platforms. More specifically, WDL compared to traditional deep learning models can substantially increase the overall cell type prediction accuracy (41 to 90%) and T-cell sub-types (CD4: 0 to 76%, and CD8: 61 to 96%) when the models were trained using melanoma data obtained from the 10X platform and tested on basal cell carcinoma data obtained using SMART-seq.

genomics↗