---

Overall correlation of mitochondrial DNA and SARS-CoV-2 across all samples using Theil-Sen estimator

Each point shows the correlation between the percent of all high-quality reads that align to SARS-CoV-2 and the percent of aligned mitochondrial reads for a specific species taken across all samples. The axes show the Spearman correlation and a measure of correlation based on the Theil-Sen estimator .

A critical aspect of this plot is the choice of what scale is used to calculate the correlations. The choices, which are adjustable using the options below the plot, are:

• Quantify the mitochondrial reads is as the percent of all high-quality (preprocessed) reads that map to the mitochondrial genome of that species, or as the percent of reads that align to any chordate mitochondrial genome that align to that species.
• Whether the axis scales show the raw percentages or the log10 of the percentages. When using a log scale, percentages of zero are set to half the lowest observed percentage across all samples and species.

The options at bottom allows you to show correlations computed only over samples collected on specific dates, isolated from particular sources, or that contain at least one SARS-CoV-2 read.

The species to highlight boxes allow you to click on a specific species to highlight in a red box on the plot.

This plot is modeled on that shown in Figure 4A of the original Liu et al (2022) Research Square pre-print, except it shows the Theil-Sen estimator correlation rather than the Pearson correlation.

See Bloom et al, bioRxiv, DOI 10.1101/2023.04.25.538336 for a paper describing this analysis.

See https://github.com/jbloom/Huanan_market_samples for results files and a fully reproducible computational pipeline.

See https://jbloom.github.io/Huanan_market_samples for links to all interactive plots.