Aleksej Zelezniak

Cross-family learning transfers thermal adaptation between enzymes

This paper shows that thermal adaptation can be transferred between unrelated enzymes using a generative framework, improving thermostability without evolutionary homology.

Joint Variable Selection for Omic Biomarkers in Time-to-Event Data

The paper introduces vampW, a Bayesian method for joint variable selection in omic biomarkers and time-to-event data, showing reduced false discoveries and improved prediction.

Genome-wide analysis of 439 mass spectrometry-based proteomic profiles in a population of 15,035 Scottish individuals

The paper performs genome-wide association studies on 439 high-abundance serum proteins in 15,035 Scottish individuals, identifying 1,553 SNP signals and 13 putative causal protein-outcome associations.

The Role of Metabolism in Shaping Enzyme Structures Over 400 Million Years of Evolution

This study uses AlphaFold2 to analyze how metabolism influenced enzyme structural evolution across 400 million years by examining 11,269 enzyme structures and their metabolic reactions.

Extensive PFAS accumulation by human gut bacteria

This study shows human gut bacteria can accumulate PFAS, with some species like Bacteroides uniformis accumulating them to high concentrations, and identifies mechanisms involving transmembrane transport and efflux pumps.

Single Distal Mutation Enhances Activity of known PETases via Stabilisation of PET-binding

The paper reports an engineered PETase, Fast_2.9, with enhanced activity and stability for PET degradation, identified through high-throughput screening of PET nanoparticles.

Learning deep representations of enzyme thermal adaptation

The paper trains a deep learning model to predict enzyme thermal adaptation using a dataset of enzymes labeled with optimal growth temperatures.

Supervised generative design of regulatory DNA forgene expression control

This paper introduces ExpressionGAN, a deep learning model that generates functional regulatory DNA with specified expression levels, outperforming random mutagenesis approaches.

Transcriptome analysis of EPO and GFP HEK293 Cell-lines Reveal Shifts in Energy and ER Capacity Support Improved Erythropoietin Production in HEK293F Cells

The paper compares transcriptomic profiles of HEK293 cell lines producing erythropoietin or GFP to identify factors associated with improved erythropoietin production.

Plastic-degrading potential across the global microbiome correlates with recent pollution trends

The paper assesses the global potential of microorganisms to degrade plastics by analyzing metagenomes and correlates enzyme abundance with pollution trends.

metaGEM: reconstruction of genome scale metabolic models directly from metagenomes.

metaGEM reconstructs genome-scale metabolic models directly from metagenomes, demonstrating comparable metabolic representation to reference-based models and capturing intraspecies diversity in gut microbiomes.

Expanding functional protein sequence space using generative adversarial networks

This paper introduces ProteinGAN, a GAN that generates functional protein sequences, showing 24% of generated sequences are soluble and active in vitro.

Benchmarking accuracy and precision of intensity-based absolute quantification of protein abundances in Saccharomyces cerevisiae

This study benchmarks the accuracy and precision of intensity-based protein quantification methods in yeast using mass spectrometry.

Bayesian genome scale modelling identifies thermal determinants of yeast metabolism

This paper uses Bayesian modeling to identify temperature-dependent enzymes in yeast metabolism, revealing squalene epoxidase as a key rate-limiting enzyme and demonstrating its role in thermotolerance.

Gene expression is encoded in all parts of a co-evolving interacting gene regulatory structure

The paper uses deep learning to show that gene expression levels are encoded in both coding and non-coding regions of the gene regulatory structure across multiple organisms.

Parallel factor analysis enables quantification and identification of highly-convolved data independent-acquired protein spectra

PARADIAS deconvolves DIA spectra to improve quantification and identification of proteins by analyzing more ion data than traditional methods.

Scanning SWATH acquisition enables high-throughput proteomics with chromatographic gradients as fast as 30 seconds

This paper introduces Scanning SWATH, a proteomics method that enables high-throughput quantification by using fast chromatographic gradients and improves precursor identification.

Precise label-free quantitative proteomes in high-throughput by microLC and data-independent SWATH acquisition

This paper introduces a label-free quantitative proteomics platform using microLC and SWATH acquisition, achieving high throughput and precision in protein quantification.