Gianluca Pollastri

The paper develops a high-quality dataset for protein sub-cellular localisation prediction, showing performance improvements and highlighting data leakage issues in existing models.

The paper introduces a classification framework for disordered flexible linkers and creates a comprehensive dataset of 1640 domain linkers, annotated with structural and sequence features.

LINKER-Pred provides deep learning tools to predict disordered flexible linkers in proteins using sequence data, achieving state-of-the-art performance on benchmarks.

PUNCH2 is a novel IDR predictor that improves accuracy using sequence embeddings and convolutional layers, outperforming existing methods on the CAID2 dataset.

This paper introduces a random forest-based method to infer epigenetic marks and assign chromatin states with fewer experimental data.

The paper evaluates the role of evolutionary profiles in protein secondary structure prediction and introduces a new protocol to fairly compare profile-based predictors.

Porter 5 predicts ab initio protein secondary structure at 84% 3-class accuracy and 73% 8-class accuracy on an independent test set, available as web and standalone tools.