Stavros Makrodimitris

The paper develops a data-driven framework to detect cancer using cfDNA fragmentation patterns at TSSs, outperforming existing methods in distinguishing cancer from healthy samples.

This paper reports on the fourth CAFA challenge, showing improvements in Gene Ontology term prediction but highlighting challenges in new evaluation settings.

The paper introduces MeD-seq, a method to profile circulating tumor DNA by analyzing DNA methylation and copy number changes, showing its effectiveness in detecting cancer-specific signals in plasma samples.

This study compares linear and non-linear joint embedding methods for bulk and single-cell multi-omics, finding non-linear methods excel in imputation and joint embeddings improve performance in supervised tasks.

This study benchmarks variational autoencoders on cancer transcriptomics data to evaluate hyperparameter effects and latent space biological relevance.

This paper improves methylation-based cell type deconvolution of cfDNA by introducing CelFEER, which achieves higher accuracy in estimating cell type proportions and detects significant differences in skeletal muscle cfDNA between ALS patients and controls.

The paper demonstrates that using universal contextualized protein embeddings from SeqVec allows effective cross-species molecular function prediction, especially for understudied species.

The paper evaluates how incomplete PPI networks contribute to functional annotation across species, showing improved performance with additional data sources like text mining.

The paper shows that unsupervised protein sequence embeddings outperform traditional features in predicting molecular function, with a simple model achieving state-of-the-art results.

The paper introduces MLC, a method that improves gene function prediction by weighting RNA-Seq samples based on Gene Ontology terms, outperforming standard Pearson correlation.