TABLE 5.
Clustering of prolines in prion-forming Q/N-rich domains
| Group | No. of prolines in clustersa |
P valuee | |
|---|---|---|---|
| Predictedb | Observed | ||
| Prion-like domainsc | 50.8 | 68 | 0.0054 |
| Non-prion Q/N domainsd | 39.0 | 25 | 0.0073 |
a
That is, prolines for which one of the two residues following the proline or one of the two residues preceding it is also a proline.
b
Predictions are based on the total number of proline residues within each potential PFD, assuming that the proline residues are randomly distributed.
c
Includes both the 18 Q/N-rich domains shown by Alberti et al. to have prion-like activity in four different assays (1) and the other known yeast PFDs (Mot3, Cyc8, and Mca1).
d
The 18 Q/N-rich domains shown by Alberti et al. to lack prion-like activity in all four assays (1).
e
P value is calculated by chi-square analysis.