Table 1.
Blood samples and associated clonotypes of publically available TCRβ sequences.
Figure 1.
PCA of Kidera Factor representation of unique CDR3 sequences.
3D scatterplot on the first three eigenvectors of the Principal Component Space obtained from Kidera Factor representations of the unique CDR3 sequences in the unseparated T cell Female sample. In this sample there are 78,391 unique Kidera Factor representations of 86,225 unique CDR3 sequences (Table 1).
Figure 2.
Diversity of Vβ CDR3 Kidera Factors.
The full set of CDR3 Kidera Factors from an unseparated T cell sample was contained in a small fraction of the theoretical total possible diversity bounded by the maximum possible Kidera Factor values for each PCA. The example shown is from the Female unseparated sample which contained 78,391 unique Kidera Factors.
Figure 3.
PCA of naïve and memory T cell samples from the same donor and unseparated T cell samples from different donors.
The first three principal components of the naïve and memory samples from Male 1 sample 2 (A) and the first two principal components of the same analysis (B). Unseparated samples from Male 1 (sample 1) and Female on the first three principal components (C) and the first two principal component axes (D). The slight shift of the male points compared to the female points suggests that these samples contain different physicochemical properties (sample size = 5000).
Figure 4.
MDS representation of similarity between samples.
Unseparated T cell samples from Female donor compared with Male 1, sample 1 (A) (2-D stress value 16.12%) and from Male 1 compared with Male 2 (B). Note the north-west south-east separation between the male and female samples and the two male samples. Naïve T cell sample compared with memory T cells from Male 1 (C) (2-D stress value 13.70%) and repeated one week later (D) (2-D stress value 14.52%); and unseparated T cell samples from Male 1 taken one week apart (E) (2-D stress value 11.72%).
Figure 5.
ANOSIM R-statistic and null distribution.
ANOSIM statistic for unseparated T cell samples from Male 1 compared with Female, p<0.001 (A); Male 1 compared with Male 2, p ≈ 0.001 (B); Naïve T cell sample compared with memory T cells from Male 1, p≈1 (C); repeated one week later, p≈1 (D); and unseparated T cell samples from Male 1 taken one week apart, p≈1 (E).
Table 2.
ANOSIM R statistic (in italics below the diagonal) and p-values (above the diagonal) for all possible pairwise combinations of the eight human blood samples. Statistically different values highlighted in bold italic.
Figure 6.
Distribution of the p-values in pairwise samples.
The distribution of the p-values after bootstrapping non aggregated subsamples from both male1 unseparated samples (Figure 6A) and the male1 unseparated sample1 and female samples (Figure 6B). The relative uniform frequency of the p-values from 0 to 1 in (A) suggests that significant p-values would occur only by chance in these subsamples. This contrasts with the high number of significant p-values (p<0.05) in the distributions between the male1 and female samples (B).