Intra-host evolution during SARS-CoV-2 prolonged infection

Figures

• 

  Figure 1.. Characterization of patients with prolonged SARS-CoV-2 infection. (A) Time intervals between onset of patients’ symptoms (blue circles), first sample sequenced—T1 (orange circles), first positive serological test (black rhombuses), second sample sequenced—T2 (white circles), and last positive RT-PCR test (orange circles), respectively. The overlapping between dates are characterized by overlapping between circles as observed, for example, in some cases for white and purple circles. Blue lines represent the interval between onset of patients’ symptoms and first sample sequenced. Black lines indicate the difference between the two samples sequenced whereas in purple we show the difference between the second time sample and the last positive RT-PCR test. (B) Key clinical features for the patients analyzed in this study. Each row indicates two different features, first row: gender and occupation; second-to-fourth rows: symptoms, fifth-to-seventh rows: comorbidities. Patients are represented in the columns.

• 

  Figure 2.. Intra-host genetic evaluation of SARS-CoV-2 genomes. (A) Distribution of iSNVs across the SARS-CoV-2 genome. Vertical line represents the within-host iSNV frequency targeting the protein products of the virus. In red, we showed the lineage-defining sites of each lineage identified in our samples. Dashed line indicates 95 per cent of allele frequency (B) Comparison of unique mapped reads versus number of iSNVs with MAF > 5 per cent of frequency identified in each of the 66 samples. (C) Spearman’s correlation tests between the Ct values and number of iSNVs. (D) Bar plot showing the distribution of iSNVs across the 66 samples.

• 

  Figure 3.. Phylogenetic analysis of prolonged samples. Maximum Likelihood tree obtained with Consensus dataset analysis under GTR + I model, containing a consensus genome for each time point of the 33 patients plus 135 populational samples. Populational sample names were excluded from the figure for clarity. Text boxes indicate the Pangolin lineage classification. A different color represents each sampled lineage in the tree. Numbers indicate each patient’s samples. Orange and purple circles represent T1 and T2 samples, respectively. Patients whose two time point samples are clustered are highlighted in light-gray boxes. Red numbers indicate patients that are not monophyletic in the haplotypes tree. Inset on the right side of the figure indicates the number of SNPs between the consensus sequences of T1 and T2 for each patient.

• 

  Figure 4.. Differential mutational signatures and prolonged infection sample classification using machine learning models. (A) Distribution of the overall proportion of transitions and transversion in SARS-CoV-2 genomes in our study (B) A → U and (C) G → A proportions in samples from T1 and T2. (D) ROC curve showing a graphical representation of the relationship between sensitivity and specificity of the time point (T1 and T2) classification. The metrics table displays model performance. (E) Overall feature the importance and exhibits the most significant variables to separate the T1 and T2 classes.