Table 1. Performance comparison in predicting knockout effects.

Method	Global Acc.	Global Coverage	Mating Acc.	Mating Coverage
Sign-linear	80.2%	76.4%	93.3%	92.2%
Sign-clustering	88.3%	73.8%	96%	94%
SPINE node variant	72.5%	2.6%	89.3%	89.3%
SPINE edge variant	NA	NA	99%	98%
Yeang et al. [5]	NA	NA	97.1%	97.1%

Shown are coverage and accuracy levels in predicting knockout effects using the entire knockout data (left) or focusing on the mating network (right). The results for the sign-linear and sign-clustering algorithms are presented for the most permissive decision cutoff (50%).