Protein stickiness, rather than number of functional protein-protein interactions, predicts expression noise and plasticity in yeast

Table 2 Multiple regression results predicting plasticity

Factor		Subtract Factor from Best Model Predicting Plasticity	Factor in Isolation
Y2H PPI	R²	0.0191¹	0.0194
	p	***	***
ACMS PPI	R²	ns	0.0040
	p	-	*
TATA Box (+/−)	R²	0.0242²	0.0445
	p	***	***
Self Interaction (+/−)	R²	0.0087³	0.0015
	p	**	ns
Gene Essentiality (+/−)	R²	ns	0.0005
	p	-	ns
Noise	R²	0.05344⁴	0.0781
	p	***	***
Noise if TATA(+)	R²	0.0315	0.0449
	slope	167.24	0.9949
	p	***	***
Noise if TATA(−)	R²	0.0224	0.0438
	slope	70.444	−0.9858
	p	***	***
TATA × Noise	R²	0.0085	-
	p	**	-
Y2H if Self	R²	0.0123	0.0004
	slope	−0.3145	−0.0221
	p	***	ns
Y2H if Non-self	R²	0.0070	0.0154
	slope	−0.0839	−0.1113
	p	**	***
Y2H × Self Interaction	R²	0.0059	-
	p	**	-

¹ removed Y2H if Self and Y2H if Non-self.
² also removed Noise if TATA(+), Noise if TATA(−) and restored Noise.
³ also removed Y2H if Self and Y2H if Non-self and restored Y2H PPI.
⁴ removed Noise if TATA(+) and Noise if TATA(−).
After extensive model building, we found that high plasticity is predicted by low stickiness (low Y2H degree), presence of a TATA box, ability to bind itself, and high noise. The TATA × noise and self-interaction × Y2H interaction terms are also statistically significant. To provide greater insight, we transformed the interaction terms as described in the Table1 legend. R² values are shown for each predicting factor in isolation (last column), as well as, more importantly, for the reduction in the total coefficient of determination R² when the factor is removed from the best model. Sometimes, as indicated in the footnotes, this involved removing multiple terms and reversing the interaction factor transformation to get a biologically interpretable result. “ns” indicates p > 0.05, * p < 0.05, ** p < 0.01, *** p < 0.001.

ISSN: 1752-0509