Understanding the links between mutations in the BAF/PBAF chromatin-remodeling complexes and expression in Kidney ccRCC cancer


Session type:

Noora Al-Muftah1
1Qatar Foundation



BAF and PBAF are two chromatin-remodeling complexes that have roles in transcriptional regulation, DNA repair, and organization of the chromatin architecture and topology. These two complexes are composed of a total of 28 genes as subunits, with 15 genes mutual between the two. Recent studies have shown links between these two complexes and cancer, in which they act as tumour suppressors and also oncogenes in several malignancies. In addition, there is a high specificity between the subunit, namely gene mutated, and the cancer type. 


In this study, we aim to understand more about the biological basis of these two complexes by investigating the relationship between transcriptional gene expression and BAF and PBAF mutational status in 415 ccRCC samples from TCGA. First, we conduct a differential gene expression between the BAF and PBAF mutants (n = 156) and the controls (n = 261) and evaluate the results according to the FDR and |logFC| values, in addition to conducting gene ontology analysis on the significant sets of genes. Next, we implement four different machine learning predictor models to predict the BAF/PBAF mutational status (i.e. case or control) from the RNA-Seq expression data of 20,531 genes. The types of models implemented are random forests, support vector machines, and logistic regression. 


We evaluated both the computational performance and biological relevance of the results. The results from the differential gene expression analysis and applying the predictor models present are not strongly informative of the relationship between BAF and PBAF mutational status and expression. 


This results of this study drive the need for future work to address some of the limitations in the study design and strengthen the analysis in order to have a better understanding of the link between BAF and PBAF complexes and transcriptional expression levels.