Markus J. V. Vähä-Koskela

This paper introduces a Transformer-derived transcriptomic score to identify drug-tolerant persister-like states in AML, showing external association with ex-vivo drug response.

This study benchmarks single-cell drug response models across 26 datasets, showing performance varies with dataset type, imbalance, and model design, with scDEAL showing competitive results.

This study uses scRNA-seq to analyze AML patient samples across disease stages, identifying immune dysregulation and potential therapeutic targets.

This study shows that clemastine eradicates cancer-associated fibroblasts, halts pancreatic cancer progression, and synergizes with gemcitabine to inhibit metastasis.

The paper compares whole exome sequencing and single-cell DNA sequencing in AML to assess clonal heterogeneity and evolution, finding both methods have strengths and limitations in variant detection and clonal analysis.

The paper develops a machine learning model to identify patient-specific multi-targeting therapies using single-cell transcriptomic data for hematological cancers and solid tumors.

This paper uses BERT to identify and extract drug-target interactions from PubMed, finding 2.1 million previously unlisted studies with ~99% accuracy in detecting drug-target pairs.

This paper introduces a drug repurposing method using chemical and genomic data to score compounds for 674 cancer types and other diseases, with results matching ongoing clinical trials.

MICHA is a pipeline that FAIRifies chemosensitivity assays by automatically annotating compounds, samples, reagents, and data processing methods, enabling interoperable and reusable drug screening data.

This paper develops a computational-experimental approach to design patient-specific drug combinations for AML that are synergistic and minimally toxic.