Table 1.

A Comparison of read accuracy (ABIP) improvements across three quality control (QC) strategies and their impacts on the De Novo assembly results.

QC method	None	50-bp trimmed at both ends	QV-Based		spike-in trained SVR
# of CCS reads selected	all 9812	all 9812	top 3000	top 5000	top 3000	top 5000
90% percentile of read accuracy	99.44%	99.48%	99.62%	99.56%	99.62%	99.56%
50% percentile of read accuracy	97.48%	97.63%	99.12%	98.61%	99.12%	98.67%
10% percentile of read accuracy	92.98%	93.06%	98.44%	94.56%	98.54%	95.09%
De Novo Assembly: # of Contigs	13 (3 FP*)	10* (0 FP)	11 (1 FP)	12 (2 FP)	10 (0 FP)	10 (0 FP)

Note*: final assembled length is 100 bp shorter.

^*FP denotes False Positive