The incorporation of machine learning methods into proteomics workflows improves the identification of disease-relevant biomarkers and biological pathways. However, machine learning models, such as deep neural networks, typically suffer from lack of interpretability. Here, we present a deep learning approach to combine biological pathway analysis and biomarker identification to increase the interpretability of proteomics experiments. Our approach integrates a priori knowledge of the relationships between proteins and biological pathways and biological processes into sparse neural networks to create biologically informed neural networks. We employ these networks to differentiate between clinical subphenotypes of septic acute kidney injury and COVID-19, as well as acute respiratory distress syndrome of different aetiologies. To gain biological insight into the complex syndromes, we utilize feature attribution-methods to introspect the networks for the identification of proteins and pathways important for distinguishing between subtypes. The algorithms are implemented in a freely available open source Python-package (https://github.com/InfectionMedicineProteomics/BINN).
Interstitial lung diseases such as idiopathic pulmonary fibrosis (IPF) are caused by persistent micro-injuries to alveolar epithelial tissues accompanied by aberrant repair processes. IPF is currently treated with pirfenidone and nintedanib, compounds which slow the rate of disease progression but fail to target underlying pathophysiological mechanisms. The DNA repair protein 8-oxoguanine DNA glycosylase-1 (OGG1) has significant roles in the modulation of inflammation and metabolic syndromes. Currently, no pharmaceutical solutions targeting OGG1 have been utilized in the treatment of IPF. In this study we show Ogg1-targeting siRNA mitigates bleomycin-induced pulmonary fibrosis in male mice, highlighting OGG1 as a tractable target in lung fibrosis. The small molecule OGG1 inhibitor, TH5487, decreases myofibroblast transition and associated pro-fibrotic gene expressions in fibroblast cells. In addition, TH5487 decreases levels of pro-inflammatory mediators, inflammatory cell infiltration, and lung remodeling in a murine model of bleomycin-induced pulmonary fibrosis conducted in male C57BL6/J mice. OGG1 and SMAD7 interact to induce fibroblast proliferation and differentiation and display roles in fibrotic murine and IPF patient lung tissue. Taken together, these data suggest that TH5487 is a potentially clinically relevant treatment for IPF but further study in human trials is required.
Sepsis is the major cause of mortality across intensive care units globally, yet details of accompanying pathological molecular events remain unclear. This knowledge gap has resulted in ineffective biomarker development and suboptimal treatment regimens to prevent and manage organ dysfunction/damage. Here, we used pharmacoproteomics to score time-dependent treatment impact in a murine Escherichia coli sepsis model after administering beta-lactam antibiotic meropenem (Mem) and/or the immunomodulatory glucocorticoid methylprednisolone (Gcc). Three distinct proteome response patterns were identified, which depended on the underlying proteotype for each organ. Gcc enhanced some positive proteome responses of Mem, including superior reduction of the inflammatory response in kidneys and partial restoration of sepsis-induced metabolic dysfunction. Mem introduced sepsis-independent perturbations in the mitochondrial proteome that Gcc counteracted. We provide a strategy for the quantitative and organotypic assessment of treatment effects of candidate therapies in relationship to dosing, timing, and potential synergistic intervention combinations during sepsis.
Data independent acquisition mass spectrometry (DIA-MS) has recently emerged as an important method for the identification of blood-based biomarkers. However, the large search space required to identify novel biomarkers from the plasma proteome can introduce a high rate of false positives that compromise the accuracy of false discovery rates (FDR) using existing validation methods. We developed a generalized precursor scoring (GPS) method trained on 2.75 million precursors that can confidently control FDR while increasing the number of identified proteins in DIA-MS independent of the search space. We demonstrate how GPS can generalize to new data, increase protein identification rates, and increase the overall quantitative accuracy. Finally, we apply GPS to the identification of blood-based biomarkers and identify a panel of proteins that are highly accurate in discriminating between subphenotypes of septic acute kidney injury from undepleted plasma to showcase the utility of GPS in discovery DIA-MS proteomics.
The complex dynamics of protein expression in plasma during hyperacute HIV-1 infection and its relation to acute retroviral syndrome, viral control, and disease progression are largely unknown. Here, we quantify 1293 blood plasma proteins from 157 longitudinally linked plasma samples collected before, during, and after hyperacute HIV-1 infection of 54 participants from four sub-Saharan African countries. Six distinct longitudinal expression profiles are identified, of which four demonstrate a consistent decrease in protein levels following HIV-1 infection. Proteins involved in inflammatory responses, immune regulation, and cell motility are significantly altered during the transition from pre-infection to one month post-infection. Specifically, decreased ZYX and SCGB1A1 levels, and increased LILRA3 levels are associated with increased risk of acute retroviral syndrome; increased NAPA and RAN levels, and decreased ITIH4 levels with viral control; and increased HPN, PRKCB, and ITGB3 levels with increased risk of disease progression. Overall, this study provides insight into early host responses in hyperacute HIV-1 infection, and present potential biomarkers and mechanisms linked to HIV-1 disease progression and viral load.