Biology subjects

Stransky, B.

Publications and source records attributed to Stransky, B..

4 recordsLinked to original sources

bioRxiv2025

Explainable AI and Multiclassifiers for Staging Biomarker Discovery in Lung Squamous Cell Carcinoma

Lung cancer is one of the most common and lethal types of cancer worldwide. Among its subtypes, lung squamous cell carcinoma (LUSC) is one of the most frequent. Identifying biomarkers for LUSC represents a significant challenge due to its high molecular heterogeneity. However, this promising search may elucidate biological mechanisms and reveal potential therapeutic targets. In this context, the present study used gene expression data from TCGA-LUSC, combined with feature selection, data balancing, machine learning, and explainable artificial intelligence (XAI) to identify possible biomarkers related to staging. The employed methods demonstrated robust classification metrics and highlighting random forest, which achieved an accuracy of 0.91. The use of data balancing and feature selection techniques proved to be crucial in the classification process. In addition, it was possible to identify the 16 most relevant genes selected by random forest using the SHapley Additive Explanations (SHAP) method. Among them, three genes (MYOSLID, IMPDH1P8, and COL9A3) were chosen by all successful classifiers, positioning themselves as potential staging biomarkers and possible molecular therapeutic targets for LUSC.

cancer biology↗

bioRxiv2023

Machine Learning Gene Signature to Metastatic ccRCC based on ceRNA Network

Renal carcinoma is a pathology of silent and multifactorial development characterized by a high rate of metastases in patients. After several studies have elucidated the activity of coding genes in the metastatic progression of renal carcinoma, new studies seek to evaluate the association of non-coding genes, such as competitive endogenous RNA (ceRNA). Thus, this study aims to build a gene signature for clear cell renal cell carcinoma (ccRCC) associated with metastatic development from a ceRNA network and to analyze the probable biological functions performed by the participants of the signature. Using ccRCC data from The Cancer Genome Atlas (TCGA), we constructed the ceRNA network with the differentially expressed genes, assembled nine gene signatures from eight feature selection techniques, and analyzed the evaluation metrics of the classification models in the benchmarking process. With the signature, we performed somatic and copy number alteration analysis, survival and metastatic progression risk analysis, and functional annotation analysis. In this study, we present an 11-gene signature (SNHG15, AF117829.1, hsa-miR-130a-3p, hsa-mir-381-3p, BTBD11, INSR, HECW2, RFLNB, PTTG1, HMMR, RASD1). Validation using the external dataset of the International Cancer Genome Consortium (ICGC-RECA) made it possible to assess the generalization of the signature, which showed an Area Under Curve of 81.5%. The genomic analysis identified the signature participants on chromosomes with highly mutated regions (G-index > 2). The hsa-miR-130a-3p, AF117829.1, hsa-miR-381-3p, and PTTG1 had a significant relationship between expression and patient survival, and the first two had a significant association with metastatic development. In addition, functional annotation resulted in relevant pathways for tumor development, such as PI3K/AKT, TNF, FoxO, RNA polymerase two transcription regulation, and cell control. Finally, by analyzing the connections of the signature genes within the ceRNA network in conjunction with studies in the literature, it was possible to obtain an overview of their activities within the ccRCC. Therefore, this gene signature identified new coding and non-coding genes and could act as potential biomarkers for a better understanding of renal carcinoma and in the development of future treatments in the clinical area.

bioinformatics↗

bioRxiv2020

A new genome-wide method to identify genes with bimodal gene expression

A new method is presented to detect bimodality in gene expression data using the Gaussian Mixture Models to cluster samples in each mode. We have used the method to search for bimodal genes in data from 25 tumor types available from The Cancer Genome Atlas. The method identified 554 genes with bimodal gene expression, of which 46 were identified in more than one cancer type. To further illustrate the impact of the method, we show that 96 out of the 554 genes with bimodal expression patterns presented different prognosis when patients belonging to the two expression peaks are compared. The software to execute the method and the corresponding documentation are available at https://github.com/LabBiosystemUFRN/Bimodality_Genes.

genomics↗

bioRxiv2020

A small interfering RNA (siRNA) database for SARS-CoV-2

Coronavirus disease 2019 (COVID-19) rapidly transformed into a global pandemic, for which a demand for developing antivirals capable of targeting the SARS-CoV-2 RNA genome and blocking the activity of its genes has emerged. In this work, we propose a database of SARS-CoV-2 targets for siRNA approaches, aiming to speed the design process by providing a broad set of possible targets and siRNA sequences. Beyond target sequences, it also displays more than 170 features, including thermodynamic information, base context, target genes and alignment information of sequences against the human genome, and diverse SARS-CoV-2 strains, to assess whether siRNAs targets bind or not off-target sequences. This dataset is available as a set of four tables in a single spreadsheet file, each table corresponding to sequences of 18, 19, 20, and 21 nucleotides length, respectively, aiming to meet the diversity of technology and expertise among labs around the world concerning siRNAs design of varied sizes, more specifically between 18 and 21nt length. We hope that this database helps to speed the development of new target antivirals for SARS-CoV-2, contributing to more rapid and effective responses to the COVID-19 pandemic.

bioinformatics↗