A classification approach for genotyping viral sequences based on multidimensional scaling and linear discriminant analysis

Table 4 Accuracy and coverage of each subtype decision step for HIV-1 nucleotide gene segments

Sequence set*	Description	No. of sequences	Accuracy (%)	Coverage (%)
(1)	Subtypes given by LANL	162,669		100
(2)	[Nested analysis] Outlierness < 2.0 & Pval > 0.99 among (1)	130,721		80.4
	Correctly classified among (2)	129,302	98.91
(3)	(1)-(2)	31,948		19.6
(4)	Outlierness < 2.0 & Subtype(major) = subtype(nested) among (3)	22,599		14.1
	Correctly classified among (4)	21,429	94.82
(5)	(3)-(4)	9,349		5.7
(6)	[Major analysis] Outlierness < 1.0 & Pval > 0.99 among (5)	1,075		0.7
	Correctly classified among (6)	781	72.65
(7)	(5)-(6)	8,274		5.1
(8)	Subtype assigned (2)+(4)+(6)	154,395		94.9
	Correctly classified among (8)	151,512	98.13
(9)	(1)-(8)	8,274		5.1
	Pval < 0.6 among (9)	292		0.2
	Outlierness > 10.0 among (9)	756		0.5

*The sequence sets (2), (4), and (6) correspond to the decision steps (i), (ii), and (iii) in the main text of the "A proposed process for subtype decision" section of Results and Discussion, respectively

ISSN: 1471-2105