Jan Kosiński

The paper identifies and characterizes a non-canonical bacterial protein translocation-folding complex using cryo-electron tomography, proteomics, and structure prediction in Mycoplasma pneumoniae.

The paper maps virus-host protein contact sites during influenza A virus infection, revealing new interactions and paraspeckle disassembly as a novel infection mechanism.

This paper introduces rejection sampling to improve macromolecular complex detection in cryo-electron tomography by incorporating biological constraints, achieving performance comparable to deep learning methods without annotated data.

The paper introduces HMFF, a method integrating experimental data into membrane simulations, and applies it to various biological systems to enable multi-scale modeling.

AlphaPulldown2 provides a pipeline for high-throughput protein structural modeling with automated workflows and improved data management.

This study shows that Dictyostelium discoideum uses nuclear pores as conduits for fluid flow during osmotic stress, distinct from passive diffusion and active transport.

This study improves AlphaFold 3 by integrating crosslink data as covalently-bound ligands, enhancing model accuracy for complexes with limited evolutionary data.

ColabSeg is a tool for processing and visualizing membrane segmentations from cryo-ET data, enabling high-throughput segmentation and training of CNNs.

The paper introduces pyTME, a Python library for template matching in cryo-EM data that improves efficiency and scalability for large datasets.

The paper evaluates template matching performance in cryo-ET by analyzing how template size, shape, and angular sampling affect ribosome detection in a ground truth dataset.

This study identifies potential antiviral drugs against influenza virus by computationally prioritizing drug-target pairs and validating their activity in cell lines.

AlphaPulldown is a Python package that simplifies protein-protein interaction screens and high-throughput modelling using AlphaFold-Multimer, offering a CLI, confidence scores, and a graphical tool.

The paper uses AI to create a dynamic model of the human nuclear pore complex, showing how the protein scaffold interacts with the nuclear envelope.

The paper introduces Assembline, a software package for integrative structural modelling of macromolecular complexes, enabling efficient sampling and analysis of conformational space.