Fig 1.
(A). Summary of the workflow used to generate cell to cell communication results from single cell RNA sequence h5ad data from the Tabula Muris Senis dataset.
(B): Structure of CCC analysis between ages. The main strategy is two-phase comparison where all ligand-receptor pairs are categorised according to their presence in two-phase ages.
Fig 2.
Single cell annotated UMAP projections, labelled by cell type (left) and time point (right).
(A): Heart (B): Kidney (C): Liver (D): Lung.
Fig 3.
Shrink and expand (SE) score pattern.
SE score of 4 organs in different time points to show how CCC shrinks or expands over time, the y-axis indicates the SE score, and the x-axis shows the different time points.
Fig 4.
Kidney gain ratios heatmap reflecting the ratio of newly generated CCC inference units.
The colour represents the value of the gain ratio: 1 indicates the communication content was completely new in the next time point, and 0 means no new communication content formed by the time point. Y axis: types of communication, X axis: time points. The green and blue colours highlight the T cell and collecting duct cell types of communication discussed in the main text.
Fig 5.
Kidney correlation heatmap of the gain ratio in Fig 4 representing types of communication had the same trend for the new communication generation.
There were roughly two clusters of types of communication that had obvious correlations. The T cell correlation cluster was coloured in yellow.
Fig 6.
Kidney consensus change binary pattern, white blocks indicate there is a significant change in the shared ligand-receptor pairs between two consecutive ages while the black blocks indicate no significant changes.
Fig 7.
Kidney potential loss ratios heatmap reflecting the ratio of potential lost CCC inference units, the color represents the value of the potential loss ratio, 1 represents that communication content was completely lost in the previous time point, and 0 means all communication in the previous time point was preserved in the next time point.
Fig 8.
Endothelial cell to Endothelial cell interaction changes in all organs.
Ligand-receptors that were shortlisted and had more than one time point change were represented as the key pairs. Colours: change categories, Y-axis: ligand-receptor pairs, X-axis: time points.
Fig 9.
Organ-level molecule analysis of the kidney.
(A): high-frequency ligand-receptor pairs, x-axis: time points with different categories of changes, y-axis: ligand (left) to receptor (right), colours show categories of changes, and the size of the dot indicates the frequency of changes in the organ (the number of types of communication for each particular pair). Most changes of pair within a single time point did not possess more than one categorical change. The pairs shown here are a subset of the top 20 pairs (see methods) for ease of visualisation, with all shown in Fig 7 in S1 Fig. (B): high-frequency single molecule patterns for the top 20 single molecules selected based on the frequency within each time point on each categorical change. The single molecules shown here are a subset of top 20 single molecules (see methods), with all the top 20 single molecules in Fig 8 in S1 Fig.