Dima Kozakov

This paper shows that focusing on complementarity-determining regions (CDRs) in antibody sequences improves binding affinity prediction using region-aware pretraining strategies.

The paper identifies phenotype-specific signaling networks in glioblastoma cells, showing that invasive spread is linked to DNA damage response and AKT-mTOR activation, with 14-3-3 proteins mediating phosphorylation and predictive phospho-site signatures for patient survival.

The paper introduces a chemical-genetic platform called SLIP to identify actionable sites on effector proteins for targeted protein degradation.

The paper fine-tunes AlphaFold using high-resolution MHC-peptide crystal structures to improve prediction accuracy for MHC-peptide complex structures, outperforming existing methods in RMSD and lDDT scores.

This study enhances AlphaFold for predicting phosphopeptide-protein complex structures, achieving higher accuracy in lDDT scores for peptides.

The paper fine-tunes protein language models on peptide-MHC sequences to predict MHC-peptide binding, outperforming the existing NetMHCpan4.1 method.

The paper introduces a multi-factor clustering method to classify diffraction data from dynamic proteins into polymorphs, using unit cell constants and intensities to cluster mixed-state data from chymotrypsinogen crystals.

Splitting AlphaFold2 complex models, re-docking with ClusPro, AlphaFold2 refinement and predicted-aligned-error ranking raised successful docking-benchmark targets from 88 (five AlphaFold2 models) to 123 (top 10) of 204.

The paper maps EMT transitions through multi-omic profiling, identifies key regulatory factors, and reveals discordance between molecular layers in human mammary epithelial cells.

The paper develops a high-resolution peptide-protein docking method called PIPER-FlexPepDock, validated on known complexes to show accuracy and efficiency.