Interpreting Links in a coverage plot

[aleenah23]

Hi,

I have recently started using Signac and I am facing some difficulty in interpreting the coverage plots. I have multimodal data from multiple samples that I have jointly analysed.

I ran the LinkPeaks function on the combined object then visualised the coverage plot with both the gene expression and peaks. In the "Links" section, I am a little bit confused regarding what it's showing. Based on the documentations and a discussion with a colleague, we assume the following:

• At every peak for a specific gene, the relationship between gene expression and accessibility is calculated. We end up with a correlation coefficient/ expected coefficient.
• Peaks with similar coefficients are linked.
• Linked peaks ultimately are the ones that strongly correlate with gene expression and therefore most likely regulate it.
• The intensity of the blue linking lines is indicative of the strength of correlation. Based on the documentation we assume that is possibly based on the z-score.

I was just wondering if this is a correct way to interpret/ read the plot. Any feedback or insights would be appreciated.

Thanks,
Aleen

[timoast]

At every peak for a specific gene, the relationship between gene expression and accessibility is calculated. We end up with a correlation coefficient/ expected coefficient.

Correct

Peaks with similar coefficients are linked.

No, not sure what this means. Peaks-gene pairs with a correlation are linked.

Linked peaks ultimately are the ones that strongly correlate with gene expression and therefore most likely regulate it.

Yes that's the goal

The intensity of the blue linking lines is indicative of the strength of correlation. Based on the documentation we assume that is possibly based on the z-score.

Yes (score is the correlation coefficient)

Please check the Signac paper for more information: https://www.nature.com/articles/s41592-021-01282-5

[aleenah23]

Thanks very much!