Table 3.

Positive predictive value pairwise comparison for soil and pediatric data

		Compared to:
		Alignment	Meta-MARC HTS Reads	Meta-MARC Assembly	Resfams
Method	Alignment	–	31.4%	33.8%	30.4%
	Meta-MARC HTS Reads	55.7%	–	64.7%	44.6%
	Meta-MARC Assembly	65.2%	33.8%	–	29.1%
	Resfams	68.3%	50.0%	70.1%	–

The table shows the percent of samples where a given method had a higher number of on-target classifications (regardless of total number classified) compared to another method. For example, Meta-MARC Assembly had a greater number of on-target classifications in 33.8% of the Soil and Pediatric samples compared to Meta-MARC high-throughput sequence reads