Biology subjects

Conejo-Garcia, J.

Publications and source records attributed to Conejo-Garcia, J..

6 recordsLinked to original sources

bioRxiv2025

A Spatial Multi-Omic Framework Identifies Gliomas Permissive to TIL Expansion

Tumor-infiltrating lymphocyte (TIL) therapy is effective in several tumor types; however, its feasibility in immune subversive tumors like glioblastoma is unclear. We expanded TILs from glioblastoma specimens and observed marked variability in yield, composition, function, and TCR clonality. By interrogating TILs expanded ex vivo alongside their sourced glioma tissue, we sought to identify determinants of successful (TIL+) vs unsuccessful TIL expansion (TIL-). Expanded TILs were predominantly effector memory CD4 cells and exhibited oligoclonal TCR enrichment. Despite similar T cell abundance in TIL vs TIL- tumors, TIL tumors exhibited distinct spatial organization and cellular interactions, including increased endothelial-immune interactions/ proximity and enrichment of vascular-associated niches. CD4 T cells localized near CD68 macrophages in TIL tumors, while they were positioned near CD163 CD206 macrophages in TIL- tumors. Thus, TIL expansion and functionality are linked to spatial organization and myeloid context; these features may enable biomarker-driven stratification for future TIL therapy in gliomas.

cancer biology↗

bioRxiv2024

An Artificial Intelligence Model for Profiling the Landscape of Antigen-binding Affinities of Massive BCR Sequencing Data

The interaction between antigens and antibodies (B cell receptors, BCRs) is the key step underlying the function of the humoral immune system in cancers and other biological contexts. The capability to profile the landscape of antigen-binding affinities of a vast number of BCRs will provide a powerful tool to reveal novel insights at unprecedented levels and will yield powerful tools for translational development. However, current experimental approaches for detecting antibody-antigen interactions are costly and time-consuming, and can only achieve low-to-mid throughput. On the other hand, bioinformatics tools in the field of antibody informatics mostly focus on optimization of antibodies given known binding antigens, which is a very different research question and of limited scope. In this work, we developed an innovative Artificial Intelligence tool, Cmai, to address the prediction of binding between antibodies and antigens that can be scaled to high-throughput sequencing data. We devised a biomarker metric based on the output from Cmai applied to high-throughput BCR sequencing data to predict the landscape of the antigen-binding affinities of the BCRs. We found that extracellular antigens on malignant tumor cells are inducing B cell infiltrations, and the infiltrating B cells have a greater tendency to co-localize with tumor cells expressing these antigens. We further found that the abundance of tumor antigen-targeting antibodies is predictive of immune-checkpoint inhibitor (ICI) treatment response. On the other hand, we found that, during immune-related adverse events (irAEs) caused by ICI treatment, humoral immunity is preferentially responsive to intracellular antigens from the organs affected by the irAEs. Overall, we developed a powerful tool for elucidating the antigen binding landscape of BCR repertoires, and it shall also propell antibody-centric translational applications for cancers and other diseases.

immunology↗

bioRxiv2023

BERLIN: Basic Explorer for single-cell RNAseq analysis and cell Lineage Determination.

Single-cell RNA sequencing has revolutionized the study of immuno-oncology, cancer biology, and developmental biology by enabling the joint characterization of gene expression and cellular heterogeneity in a single platform. As of July 2023, the Gene Expression Omnibus now contains over 4000 published single-cell data sets, providing an invaluable opportunity for reanalysis to identify new cell types or cellular states as well as their defining transcriptional programs. To facilitate the reprocessing of these public datasets, we have devised a single-cell RNA sequencing analysis framework for data retrieval, quality control, expression normalization, dimension reduction, cell clustering, and data integration. Additionally, we have developed a Shiny App visualization platform that enables the exploration of gene expression, cell type annotations, and cell lineages through a user interface. We performed a re-analysis of single-cell RNAseq data generated from acute myeloid leukemia and tumor-reactive lymphocytes and found our pipeline to faithfully recapitulated the cell type assignment as well as expected lineage trajectories. Altogether, we present BERLIN, a single-cell RNAseq analysis pipeline that facilitates the integration and public dissemination of results from the reanalysis.

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↗

bioRxiv2019

Estimation of immune cell content in tumor using single-cell RNA-seq reference data

BackgroundThe rapid development of single-cell RNA sequencing (scRNA-seq) provides unprecedented opportunities to study the tumor ecosystem that involves a heterogeneous mixture of cell types. However, the majority of previous and current studies related to translational and molecular oncology have only focused on the bulk tumor and there is a wealth of gene expression data accumulated with matched clinical outcomes.\n\nResultsIn this paper, we introduce a scheme for characterizing cell compositions from bulk tumor gene expression by integrating signatures learned from scRNA-seq data. We derived the reference expression matrix to each cell type based on cell subpopulations identified in head and neck cancer dataset. Our results suggest that scRNA-Req-derived reference matrix outperforms the existing gene panel and reference matrix with respect to distinguishing immune cell subtypes.\n\nConclusionsFindings and resources created from this study enable future and secondary analysis of tumor RNA mixtures in head and neck cancer for a more accurate cellular deconvolution, and can facilitate the profiling of the immune infiltration in other solid tumors due to the expression homogeneity observed in immune cells.

cancer biology↗

bioRxiv2019

DNA methylation-based immune cell deconvolution in solid tumors

Understanding of the tumor microenvironment (TME) structure is likely to have a profound and immediate impact on therapeutic interventions as well as the development of signatures for diagnostic and prognostic evaluations. DNA methylation arrays represent one of the most reproducible molecular assays across replicates and studies, but its value of profiling tumor-infiltrating immune lymphocytes (TILs) hasnt been intensively investigated. Here we report a model-based evaluation of tumor TIL levels using DNA methylation profiles. By employing a hybrid method of stability selection and elastic net, we show that methylation array data in ten TCGA cancer types provide a strikingly accurate prediction of immune cell abundance, in particular the levels of T cells, B cells and cytotoxic cells in skin cutaneous melanoma (SKCM). The immune-informative CpG sites showed significant prognostic values, representing important candidates for further functional validation. Further, we present regression models each using only ten CpG sites to estimate the levels of infiltrated immune cell types in melanoma. To validate these models, we performed matched methylation EPIC array and RNA-seq on 30 new melanoma samples. We observed high concordance on methylation and gene expression predicted tumor immune infiltration levels in our new dataset. Our study demonstrated that DNA methylation data is a valuable resource in reliably evaluating tumor immune responses. The selected methylation panels provide candidate targets for future clinical researches. Our prediction models are easy to implement and will provide reference for future clinical practices.

bioinformatics↗