Table 2. Influence of different predictors on the probability of Spiroplasma infection [Pr(Spiro+)] in generalized linear models (GLM).

Generalized linear model (GLM) by watershed, with season as the predictor. The table includes watershed, season of collection [wet, intermediate (Inter), and dry], GLM estimate of Pr(Spiro+), the lower 2.5% limit and the upper 97.5% limit of the estimate, the Tukey’s contrast (TC) used to test for significant differences in Pr(Spiro+) between seasons, and the TC p-value. The only significant differences in Pr(Spiro+) between seasons was between the wet and the intermediate seasons in the Albert Nile and Achwa River watersheds. Additionally, none of the other predictors (GB, mtDNA haplogroup, trypanosome infection status, or sex) had significant effect on Pr(Sprio) (S5 Table). For all p-values, the level of significance was marked *** if < 0.0001, ** if < 0.001, * if < 0.05, and not marked if > 0.05.

GLM
Watershed	Season	Pr(Spiro+)	Lower limit (2.5%)	Upper limit (97.5%)	TC	TC p-value
Albert Nile	Wet	0.16	0.10	0.22	Wet—Inter	1.160e-07 ***
	Inter	0.42	0.37	0.48
Achwa River	Wet	0.08	0.04	0.13	Wet—Inter	5.380e-06 ***
	Inter	0.36	0.27	0.46
Okole River	Wet	0.13	0.05	0.26	Wet—Inter	0.7777
	Inter	0.03	0.02	0.06	Inter—Dry	0.5577
	Dry	0.06	0.03	0.11	Wet—Dry	0.0962
Lake Kyoga	Wet	0.44	0.23	0.67	Wet—Inter	1
	Inter	0.00	0.00	1.00	Inter—Dry	1
	Dry	0.00	0.00	1.00	Wet—Dry	1
Kafu River	Wet	0.00	0.00	1.00	Wet—Dry	1
	Dry	0.04	0.01	0.09