SvAnna: efficient and accurate pathogenicity prediction of coding and regulatory structural variants in long-read genome sequencing

Table 1 Summary of rules for calculating sequence deleteriousness score δ(g)

SV class	*t ⊂ SV*	t ⇌ SV	SV ⊂ e
SV class	*t ⊂ SV*	t ⇌ SV	Coding or splice	UTR	Intronic	Promoter
DEL	1	1	{0.8, 1}	0 ≤ δ(g)≤1	0	0.4
DUP	1^a	0	{0.8, 1}	0 ≤ δ(g)≤1	0	0.4
INV	0	1	1	0 ≤ δ(g)≤1	0	0.4
INS	–	–	{0.2, 0.9}	0 ≤ δ(g)≤1	0	0.4
BND	–	–	1	1	1	0.4

Higher scores indicate a greater degree of predicted deleterious effect on transcript function. t ⊂ SV: The SV fully contains the transcript in question. t ⇌ SV: Partial overlap of the transcript and the SV. SV ⊂ e The SV is completely contained within the indicated sequence element. {0.8, 1} and {0.2, 0.9} indicate scores for {in-frame, frameshift} variants
^aDuplication of the entire gene is assigned a score of 1, triplication is assigned a score of 2, and so on

ISSN: 1756-994X