Gavriilidis, G. | bioRxiv Science

bioRxiv2024

APNet, an explainable sparse deep learning model to discover differentially active drivers of severe COVID-19

MotivationComputational analyses of plasma proteomics provide translational insights into complex diseases such as COVID-19 by revealing molecules, cellular phenotypes, and signaling patterns that contribute to unfavorable clinical outcomes. Current in silico approaches dovetail differential expression, biostatistics, and machine learning, but often overlook nonlinear proteomic dynamics, like post-translational modifications, and provide limited biological interpretability beyond feature ranking. ResultsWe introduce APNet, a novel computational pipeline that combines differential activity analysis based on SJARACNe co-expression networks with PASNet, a biologically-informed sparse deep learning model to perform explainable predictions for COVID-19 severity. The APNet driver-pathway network ingests co-expression and classification weights to aid result interpretation and hypothesis generation. APNet outperforms alternative models in patient classification across three COVID-19 proteomic datasets, identifying predictive drivers and pathways, including some confirmed in single-cell omics and highlighting under-explored biomarker circuitries in COVID-19. Availability and ImplementationAPNets R, Python scripts and Cytoscape methodologies are available at https://github.com/BiodataAnalysisGroup/APNet Contactggeorav@certh.gr Supplementary informationSupplementary information can be accessed in Zenodo (10.5281/zenodo.10438830).

systems biology↗

bioRxiv2020

Protein citrullination was introduced into animals by horizontal gene transfer from cyanobacteria

Protein post-translational modifications (PTMs) add an enormous amount of sophistication to biological systems but their origins are largely unexplored. Citrullination, a key regulatory mechanism in human physiology and pathophysiology, is particularly enigmatic in an evolutionary context. The citrullinating enzymes peptidylarginine deiminases (PADIs) are ubiquitous across vertebrates but absent from yeast, worms and flies. Here, we map the surprising evolutionary trajectory of PADIs into the animal lineage. We present strong phylogenetic support for a clade encompassing animal and cyanobacterial PADIs that excludes fungal and other bacterial homologues. The animal and cyanobacterial PADIs share unique, functionally relevant synapomorphies that are absent from all other homologues. Molecular clock calculations and sequence divergence analyses using the fossil record estimate the last common ancestor of the cyanobacterial and animal PADIs to be approximately 1 billion years old, far younger than the 3.35-4.52 billion years known to separate bacterial and eukaryotic lineages. Under an assumption of vertical descent, PADI sequence change is anachronistically slow during this evolutionary time frame, even when compared to mitochondrial proteins, products of likely endosymbiont gene transfer and some of the most highly conserved proteins in life. The consilience of evidence indicates that PADIs were introduced from cyanobacteria into animals by horizontal gene transfer (HGT). The ancestral cyanobacterial protein is enzymatically active and can citrullinate eukaryotic proteins, suggesting that the PADI HGT event introduced a new catalytic capability into the regulatory repertoire of animals. This study reveals the unusual evolution of a pleiotropic protein modification with clear relevance in human physiology and disease.

evolutionary biology↗