Table 1.

Features extracted from mappability-profile processed coverage profiles for the machine-learning process. The median variable importance is summarized from all models trained in experiments described in the Results section. The top 5 features of each method are highlighted in bold font

Categories	Features	Median Variable Importance
		IIMI-RF	IIMI-X
Basic overall summaries of the virus genome	average coverage	0.1003	0.0324
	maximum coverage	0.1189	0.0294
	length of genome	0.0917	0.0716
Nucleotide composition percentages	% of A	0.0482	0.0361
	% of C	0.0417	0.0203
	% of T	0.0557	0.0285
	GC content	0.0559	0.0359
Percentage of coverage exceeds a certain threshold ( reads) across the whole virus genome		0.1894	0.6699
		0.0967	0.0081
		0.062	0.0012
		0.0484	0.0179
		0.0367	0.0064
		0.0295	0.0009
		0.0363	0.0057
		0.0475	0.0034
		0.0657	0.0323