Interpretable and Predictive Deep Neural Network Modeling of the SARS-CoV-2 Spike Protein Sequence to Predict COVID-19 Disease Severity

Sokhansanj, Bahrad A.; Zhao, Zhengqiao; Rosen, Gail

doi:10.3390/biology11121786

Cited by 4 publications

(3 citation statements)

References 120 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Instead, the most recent approaches to biomedical text extraction tasks have employed transformer-based techniques, as reviewed in [ 36 ] and [ 37–39 ]; they report that current works are mainly focused on connections between entities [ 40 , 41 ]. Very few works addressed results’ explainability combined with transformers in this domain [ 42 , 43 ].…”

Section: Related Workmentioning

confidence: 99%

CoVEffect: interactive system for mining the effects of SARS-CoV-2 mutations and variants based on deep learning

Serna García,

Al Khalaf,

Invernici

et al. 2022

GigaScience

View full text Add to dashboard Cite

Background Literature about SARS-CoV-2 widely discusses the effects of variations that have spread in the past 3 years. Such information is dispersed in the texts of several research articles, hindering the possibility of practically integrating it with related datasets (e.g., millions of SARS-CoV-2 sequences available to the community). We aim to fill this gap, by mining literature abstracts to extract—for each variant/mutation—its related effects (in epidemiological, immunological, clinical, or viral kinetics terms) with labeled higher/lower levels in relation to the nonmutated virus. Results The proposed framework comprises (i) the provisioning of abstracts from a COVID-19–related big data corpus (CORD-19) and (ii) the identification of mutation/variant effects in abstracts using a GPT2-based prediction model. The above techniques enable the prediction of mutations/variants with their effects and levels in 2 distinct scenarios: (i) the batch annotation of the most relevant CORD-19 abstracts and (ii) the on-demand annotation of any user-selected CORD-19 abstract through the CoVEffect web application (http://gmql.eu/coveffect), which assists expert users with semiautomated data labeling. On the interface, users can inspect the predictions and correct them; user inputs can then extend the training dataset used by the prediction model. Our prototype model was trained through a carefully designed process, using a minimal and highly diversified pool of samples. Conclusions The CoVEffect interface serves for the assisted annotation of abstracts, allowing the download of curated datasets for further use in data integration or analysis pipelines. The overall framework can be adapted to resolve similar unstructured-to-structured text translation tasks, which are typical of biomedical domains.

show abstract

Section: Related Workmentioning

confidence: 99%

CoVEffect: interactive system for mining the effects of SARS-CoV-2 mutations and variants based on deep learning

Serna García,

Al Khalaf,

Invernici

et al. 2022

GigaScience

View full text Add to dashboard Cite

show abstract

“…Machine learning models have been successfully applied to complex phenomena, including cardiovascular diseases [ 26 ], and to make reliable predictions for emerging COVID-19 variants. For instance, as in [ 27 ], where the authors introduced a novel interpretable deep learning architecture to predict SARS-Cov-2 disease severity.…”

Section: Introductionmentioning

confidence: 99%

An Epidemiological Analysis for Assessing and Evaluating COVID-19 Based on Data Analytics in Latin American Countries

Leiva,

Alcudia,

Montano

et al. 2023

Biology

View full text Add to dashboard Cite

This research provides a detailed analysis of the COVID-19 spread across 14 Latin American countries. Using time-series analysis and epidemic models, we identify diverse outbreak patterns, which seem not to be influenced by geographical location or country size, suggesting the influence of other determining factors. Our study uncovers significant discrepancies between the number recorded COVID-19 cases and the real epidemiological situation, emphasizing the crucial need for accurate data handling and continuous surveillance in managing epidemics. The absence of a clear correlation between the country size and the confirmed cases, as well as with the fatalities, further underscores the multifaceted influences on COVID-19 impact beyond population size. Despite the decreased real-time reproduction number indicating quarantine effectiveness in most countries, we note a resurgence in infection rates upon resumption of daily activities. These insights spotlight the challenge of balancing public health measures with economic and social activities. Our core findings provide novel insights, applicable to guiding epidemic control strategies and informing decision-making processes in combatting the pandemic.

show abstract

“…First, there is a dire need for improved surveillance of viral evolution. Predictive tools linking viral sequence to immune evasion potential are critical at this point, and a number of groups have demonstrated that such approaches are indeed technically feasible [86][87][88][89][90][91][92] . These methods should be deployed at scale-global monitoring of viral evolution and accurate threat forecasting are critical for preventing a sudden mortality event.…”

mentioning

confidence: 99%

The gray swan: model-based assessment of the risk of sudden failure of hybrid immunity to SARS-CoV-2

Stoddard¹,

Yuan²,

Sarkar

et al. 2023

Preprint

View full text Add to dashboard Cite

In the fourth year of the COVID-19 pandemic, public health authorities worldwide have adopted a strategy of learning to live with SARS-CoV-2. This has involved the removal of measures for limiting viral spread, resulting in a large burden of recurrent SARS-CoV-2 infections. Crucial for managing this burden is the concept of the so-called wall of hybrid immunity, through repeated reinfections and vaccine boosters, to reduce the risk of severe disease and death. Protection against both infection and severe disease is provided by the induction of neutralizing antibodies (nAbs) against SARS-CoV-2. However, pharmacokinetic (PK) waning and rapid viral evolution both degrade nAb binding titers. The recent emergence of variants with strongly immune evasive potential against both the vaccinal and natural immune responses raises the question of whether the wall of population-level immunity can be maintained in the face of large jumps in nAb binding potency. Here we use an agent-based simulation to address this question. Our findings suggest large jumps in viral evolution may cause failure of population immunity resulting in sudden increases in mortality. As a rise in mortality will only become apparent in the weeks following a wave of disease, reactive public health strategies will not be able to provide meaningful risk mitigation. Learning to live with the virus could thus lead to large death tolls with very little warning. Our work points to the importance of proactive management strategies for the ongoing pandemic, and to the need for multifactorial approaches to COVID-19 disease control.

show abstract

Interpretable and Predictive Deep Neural Network Modeling of the SARS-CoV-2 Spike Protein Sequence to Predict COVID-19 Disease Severity

Cited by 4 publications

References 120 publications

CoVEffect: interactive system for mining the effects of SARS-CoV-2 mutations and variants based on deep learning

CoVEffect: interactive system for mining the effects of SARS-CoV-2 mutations and variants based on deep learning

An Epidemiological Analysis for Assessing and Evaluating COVID-19 Based on Data Analytics in Latin American Countries

The gray swan: model-based assessment of the risk of sudden failure of hybrid immunity to SARS-CoV-2

Contact Info

Product

Resources

About