SIMON: Open-Source Knowledge Discovery Platform

Tomić, Adriana; Tomić, Ivan; Kühn, Max; Spreng, Rachel L.; Dahora, Lindsay C.; Seaton, Kelly E.; Tomaras, Georgia D.; Hill, Jennifer; Duggal, Niharika Arora; Pollock, Ross D.; Lazarus, Norman R.; Harridge, Stephen; Lord, Janet M.; Khatri, Purvesh; Pollard, Andrew J.; Davis, Mark M.

doi:10.1016/j.patter.2020.100178

Cited by 17 publications

(10 citation statements)

References 54 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We also applied machine learning (ML) to define clinical and PB protein markers that predict disease outcome (recovered vs fatal) or progression (early vs late death). First, the combined clinical and PB dataset was partitioned in training, validation and test sets for building models using 171 ML algorithms in SIMON (Sequential Iterative Modeling ‘‘Over Night’’) (Tomic et al, 2019; Tomic et al, 2021) to identify the best performing ML predictive models and order the best clinical/PB parameters for prediction by feature selection. Based on metrics of performance, we selected the random forest (RF) model and sorted the clinical/PB parameters in descending order by their permutation-based importance (mean decrease accuracy or increase out-of-bag error), averaged over 50 RF runs ( Figures S2C-F ) (Ganggayah et al, 2019; Jiang, 2020; Ludemann et al, 2006; Speiser et al, 2019; Tuleau-Malot, 2022; Wickham, 2020; Wiener, 2002).…”

Section: Resultsmentioning

confidence: 99%

Disease trajectories in hospitalized COVID-19 patients are predicted by clinical and peripheral blood signatures representing distinct lung pathologies

Da Silva Filho,

Herder,

Gibbins

et al. 2023

Preprint

View full text Add to dashboard Cite

SummaryLinking clinical biomarkers and lung pathology still is necessary to understand COVID-19 pathogenesis and the basis of progression to lethal outcomes. Resolving these knowledge gaps enables optimal treatment approaches of severe COVID-19. We present an integrated analysis of longitudinal clinical parameters, blood biomarkers and lung pathology in COVID-19 patients from the Brazilian Amazon. We identified core signatures differentiating severe recovered patients and fatal cases with distinct disease trajectories. Progression to early death was characterized by rapid and intense endothelial and myeloid activation, presence of thrombi, mostly driven by SARS-CoV-2+macrophages. Progression to late death was associated with systemic cytotoxicity, interferon and Th17 signatures and fibrosis, apoptosis, and abundant SARS-CoV-2+epithelial cells in the lung. Progression to recovery was associated with pro-lymphogenic and Th2-mediated responses. Integration of ante-mortem clinical and blood biomarkers with post-mortem lung-specific signatures defined predictors of disease progression, identifying potential targets for more precise and effective treatments.

show abstract

Section: Resultsmentioning

confidence: 99%

Disease trajectories in hospitalized COVID-19 patients are predicted by clinical and peripheral blood signatures representing distinct lung pathologies

Da Silva Filho,

Herder,

Gibbins

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Tailored ML frameworks and platforms that account for the idiosyncrasies of the underlying data have been published for applications in genomics 39,40 , proteomics 41,42 , biomedicine 43 , and chemistry 44 . Their creation recognizes the infeasibility to define, implement, and train appropriate ML models by relying solely on generic ML frameworks such as scikit-learn 45 or PyTorch 46 .…”

Section: Introductionmentioning

confidence: 99%

The immuneML ecosystem for machine learning analysis of adaptive immune receptor repertoires

et al. 2021

View full text Add to dashboard Cite

Adaptive immune receptor repertoires (AIRR) are key targets for biomedical research as they record past and ongoing adaptive immune responses. The capacity of machine learning (ML) to identify complex discriminative sequence patterns renders it an ideal approach for AIRR-based diagnostic and therapeutic discovery. To date, widespread adoption of AIRR ML has been inhibited by a lack of reproducibility, transparency, and interoperability. immuneML ( immuneml.uio.no ) addresses these concerns by implementing each step of the AIRR ML process in an extensible, open-source software ecosystem that is based on fully specified and shareable workflows. To facilitate widespread user adoption, immuneML is available as a command-line tool and through an intuitive Galaxy web interface, and extensive documentation of workflows is provided. We demonstrate the broad applicability of immuneML by (i) reproducing a large-scale study on immune state prediction, (ii) developing, integrating, and applying a novel method for antigen specificity prediction, and (iii) showcasing streamlined interpretability-focused benchmarking of AIRR ML. 1.

show abstract

“…Pairwise comparisons within groups were performed using Wilcoxon matched-pairs signed-rank tests. Principal component analysis (PCA) was performed using SIMON software version 0.2.1 (https://genular.org) (37). Before PCA was performed, the data was pre-processed (center/scale), missing values were median imputed and variables with fewer than 5 unique values were removed.…”

Section: Discussionmentioning

confidence: 99%

Omicron BA.1/BA.2 infections in triple-vaccinated individuals enhance a diverse repertoire of mucosal and blood immune responses

Hornsby

Nicols

Longet

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Pronounced immune escape by the SARS-CoV-2 Omicron variant has resulted in large numbers of individuals with hybrid immunity, generated through a combination of vaccination and infection. Based primarily on circulating neutralizing antibody (NAb) data, concerns have been raised that omicron breakthrough infections in triple-vaccinated individuals result in poor induction of omicron-specific immunity, and that a history of prior SARS-CoV-2 in particular is associated with profound immune dampening. Taking a broader and comprehensive approach, we characterized mucosal and blood immunity to both spike and non-spike antigens following BA.1/BA.2 infections in triple mRNA-vaccinated individuals, with and without a history of previous SARS-CoV-2 infection. We find that the majority of individuals increase BA.1/BA.2/BA.5-specific NAb following infection, but confirm that the magnitude of increase and post-omicron titres are indeed higher in those who were infection-naive. In contrast, significant increases in nasal antibody responses are seen regardless of prior infection history, including neutralizing activity against BA.5 spike. Spike-specific T cells increase only in infection-naive vaccinees; however, post-omicron T cell responses are still significantly higher in previously-infected individuals, who appear to have maximally induced responses with a CD8+ phenotype of high cytotoxic potential after their 3rd mRNA vaccine dose. Antibody and T cell responses to non-spike antigens also increase significantly regardless of prior infection status, with a boost seen in previously-infected individuals to immunity primed by their first infection. These findings suggest that hybrid immunity induced by omicron breakthrough infections is highly dynamic, complex, and compartmentalised, with significant immune enhancement that can help protect against COVID-19 caused by future omicron variants.

show abstract

SIMON: Open-Source Knowledge Discovery Platform

Cited by 17 publications

References 54 publications

Disease trajectories in hospitalized COVID-19 patients are predicted by clinical and peripheral blood signatures representing distinct lung pathologies

Disease trajectories in hospitalized COVID-19 patients are predicted by clinical and peripheral blood signatures representing distinct lung pathologies

The immuneML ecosystem for machine learning analysis of adaptive immune receptor repertoires

Omicron BA.1/BA.2 infections in triple-vaccinated individuals enhance a diverse repertoire of mucosal and blood immune responses

Contact Info

Product

Resources

About