Fig 1.
Sample molecular transmission network.
Clusters [shown as gray circles with black borderline such as C1, C2 and C3], and three clusters are expanded for more clarity, the nodes represent viral sequences and the edges between every pair of nodes show epidemiological relatedness.
Fig 2.
Relationship of the active elements in progressing HIV transmissions.
Fig 3.
People living with HIV, San Diego PIRC.
A. Racial distribution, B. Category of transmission risk, and C. Age distribution.
Table 1.
San Diego PIRC (Primary Infection Resource Consortium) data: 1998-2019.
Table 2.
Macro average performance measures of machine learning algorithms on extended feature set trained classifiers using San Diego cohort data (1998-2019).
Fig 4.
Accuracy measures of the trained models.
A. Cross validation—binary classification, B. ROC and Area Under Curve.
Table 3.
Predictive performance measures.
Fig 5.
Accuracy comparison for node assignments to 29-clusters incorporating metadata.
A. Cross validation, B. Stratified Cross Validation.
Fig 6.
Hybrid unsupervised/supervised HIV transmission reconstruction model pipeline.
Generating a labeled dataset from molecular data (blue dashed line box), reconstructing transmission network using labeled dataset (red dashed box), classification using augmented genetic and non genetic data (black dashed box).