Table 1.

Factorial analysis of deviance of lethality phenotypes for 55 wild-type strains in 29 perturbations of germline-expressed genes

DOI: http://dx.doi.org/10.7554/eLife.09178.005

	DF	Deviance	Resid. DF	Resid. Dev	F	p-value
NULL	–	–	17,855	2,201,873	–	–
Strain	54	338,618	17,801	1,863,255	334.697	<10−15
Targeted gene	28	1,152,310	17,773	710,945	2196.584	<10−15
Adults per well	1	35,318	17,772	675,627	1885.113	<10−15
Date	1	2406	17,771	673,221	128.416	<10−15
Strain × gene	1512	349,415	16,259	323,806	12.334	<10−15
Strain × adults per well	54	6715	16,205	317,091	6.637	<10−15
Gene × adults per well	28	7358	16,177	309,732	14.026	<10−15

The table rows report information associated with each term in our statistical model (see ‘Materials and methods’), which represent distinct sources for the variation we observed in embryonic lethality. All terms were highly significant, including the strain-by-gene interaction, which represents variation attributable to cryptic genetic modifiers that act gene-specifically. This term and the strain term, which represents variation attributable to informational modifiers affecting germline RNAi, explain similar amounts of variation, and together account for 31% of the total deviance.