High-resolution proteomics identifies potential new markers of Zika and dengue infections
Allgoewer, K.; Zhao, A.; Maity, S.; Lashua, L.; Ramgopal, M.; Balkaran, B. N.; Liu, L.; Arevalo, M. T.; Ross, T. M.; Choi, H.; Ghedin, E.; Vogel, C.
Show abstract
Distinguishing between Zika and dengue virus infections is critical for treatment and anticipation of complications. However, existing biomarkers have high error rates. To identify new potential diagnostic signatures, we used next-generation proteomics to profile 122 serum samples from 62 Zika or dengue patients. We quantified >500 proteins and identified 26 proteins that were significantly differentially expressed. These proteins typically function in infection and wound healing, with several also linked to pregnancy and brain. Integrating machine learning approaches, we used 7 proteins to predict ZIKV infection correctly in 72% of the cases, outperforming other tools. The three most predictive proteins were Platelet Factor 4 Variant 1, Fibrinogen Alpha, and Gelsolin. Finally, we showed that temporal changes in protein signatures from the same patient can disambiguate some diagnoses and serve as indicators for past infections. Taken together, we demonstrate that serum proteomics can be highly valuable to diagnose even challenging samples.
Matching journals
The top 4 journals account for 50% of the predicted probability mass.
Similar papers in this journal
- Surveying the vampire bat (Desmodus rotundus) serum proteome: a resource for identifying immunological proteins and detecting pathogens 95%
- Deep proteomics network and machine learning analysis of human cerebrospinal fluid in Japanese encephalitis virus infection 94%
- Mass spectrometry-based sequencing of the anti-FLAG-M2 antibody using multiple proteases and a dual fragmentation scheme 93%
Similar papers in this journal
- Automated Spatially Targeted Optical Micro Proteomics (autoSTOMP) to determine protein complexity of subcellular structures 95%
- In Vivo Large Scale Mapping Of Protein Turnover In The Human Cerebrospinal Fluid 92%
- Sensitive identification of known and unknown protease activities by unsupervised linear motif deconvolution 92%
Similar papers in this journal
- SugarQuant: a streamlined pipeline for multiplexed quantitative site-specific N-glycoproteomics 92%
- Imputation of label-free quantitative mass spectrometry-based proteomics data using self-supervised deep learning 92%
- Saudi Arabian SARS-CoV-2 genomes implicate a mutant Nucleocapsid protein in modulating host interactions and increased viral load in COVID-19 patients 92%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.