Robust evaluation of deep learning-based representation methods for survival and gene essentiality prediction on bulk RNA-seq data.

Deep learning (DL) has shown potential to provide powerful representations of bulk RNA-seq data in cancer research. However, there is no consensus regarding the impact of design choices of DL approaches on the performance of the learned representation, including the model architecture, the training methodology and the various hyperparameters. To address this problem, we evaluate the performance of various design choices of DL representation learning methods using TCGA and DepMap pan-cancer datasets and assess their predictive power for survival and gene essentiality predictions. We demonstrate that baseline methods achieve comparable or superior performance compared to more complex models on survival predictions tasks. DL representation methods, however, are the most efficient to predict the gene essentiality of cell lines. We show that auto-encoders (AE) are consistently improved by techniques such as masking and multi-head training. Our results suggest that the impact of DL representations and of pretraining are highly task- and architecture-dependent, highlighting the need for adopting rigorous evaluation guidelines. These guidelines for robust evaluation are implemented in a pipeline made available to the research community.

DNA methylation disruption reshapes the hematopoietic differentiation landscape.

Mutations in genes involved in DNA methylation (DNAme; for example, TET2 and DNMT3A) are frequently observed in hematological malignancies1-3 and clonal hematopoiesis4,5. Applying single-cell sequencing to murine hematopoietic stem and progenitor cells, we observed that these mutations disrupt hematopoietic differentiation, causing opposite shifts in the frequencies of erythroid versus myelomonocytic progenitors following Tet2 or Dnmt3a loss. Notably, these shifts trace back to transcriptional priming skews in uncommitted hematopoietic stem cells. To reconcile genome-wide DNAme changes with specific erythroid versus myelomonocytic skews, we provide evidence in support of differential sensitivity of transcription factors due to biases in CpG enrichment in their binding motif. Single-cell transcriptomes with targeted genotyping showed similar skews in transcriptional priming of DNMT3A-mutated human clonal hematopoiesis bone marrow progenitors. These data show that DNAme shapes the topography of hematopoietic differentiation, and support a model in which genome-wide methylation changes are transduced to differentiation skews through biases in CpG enrichment of the transcription factor binding motif.