Efficient COVID-19 testing via contextual model based compressive sensing.

Area of Science:

• Computational biology
• Epidemiology
• Graph signal processing

Background:

• The COVID-19 pandemic necessitates rapid and efficient widespread testing to control its spread and inform public health policies.
• Current PCR-based testing methods are time-consuming and limit the scale of testing, hindering effective pandemic management.
• Traditional group testing methods have limitations, particularly concerning sparsity requirements.

Purpose of the Study:

• To develop a novel approach for identifying individuals affected by COVID-19 using a reduced number of tests.
• To improve the efficiency of diagnostic testing by leveraging contextual information and advanced modeling techniques.
• To offer a more resource-effective strategy for large-scale disease surveillance.

Main Methods:

• Utilizing contextual information (age, symptoms, contacts, underlying conditions) to construct a graph-based model.
• Applying model-based compressive sensing (CS) integrated with Discrete Graph Signal Processing on Graphs (DSPG).
• Developing a contextual model to optimize CS efficiency for reduced sample requirements and improved recovery quality.

Main Results:

• The proposed method significantly reduces the number of required tests compared to individual and traditional group testing.
• Achieved an increase in testing speed (individuals per test ratio) of over 15 times.
• Maintained a low error rate of less than 5%, outperforming traditional compressive sensing methods.

Conclusions:

• The graph-based compressive sensing approach offers a highly efficient and accurate method for COVID-19 testing.
• This strategy conserves valuable time and resources, crucial for managing widespread infectious diseases.
• The model's applicability extends beyond group testing limitations, offering a versatile solution for public health challenges.