Molecular Vision: Cevik, Mol Vis 2023; 29:217-233. Table 3

Table 3. Structure analysis and overall pathogenicity assessments of the missense ABCA4 variants found in the study.

Domain	ABCA4 variants	Clinical significance	Functional in vitro studies	Pathogenicity	RMSD	In silico	Protein structural changes
Domain	ABCA4 variants	Clinical significance	Functional in vitro studies	prediction	RMSD	∆∆G exp. (kcal/mol)	Protein structural changes
ECD1	c.161G>A:p(C54Y)	P/LP	-	P	1.263	41.7	Broken disulfide bond, clashes
	c .634C>T:p(R212C)	P/LP	↓ Expression^{20, 62}, ↓ΑΤP binding²⁰, Basal&Stim. ΑΤPase⁶²	P	0.971	1.26	Broken H bond
	c .1622T>C:p(L541P)	CI	Mislocalization, ¯ Basal&Stim. ATPase^17,20	P	1.271	7.65	Buried Pro introduced, broken helix
NBD1	c.3113C>T:p(A1038V)	P/LP	↓ Basal&Stim. ΑΤPase¹⁷, Normal subcellular localization²³	B	1.248	6.6	Slight conformation change
	c .3292C>T:p(R1098C)	P/LP	-	P	1.217	2.49	Lost salt bridge with Asp-2242
RD1	c.3758C>T:p(T1253M)	CI	-	P	1.392	1.54	Premature B-sheet, loss of H-bond
TMD2	c.4139C>T:p(P1380L)	P/LP	↓ Basal&Stim. ATPase ¹⁹	P	1.264	12.2	Clashes, loss of the Pro-induced kink
ECD2	c.4793C>A:p(A1598D)	P/LP	↓ Expression, ↓ Basal&Stim. ATPase ⁶²	Pǂ	1.197	13.56	Clashes, buried hydrophilic introduced
NBD2	c.5882G>A:p(G1961E)	CI	↓ Basal&Stim. ATPase ^17,20	P	0.985	31.68	Clashes, buried hydrophilic introduced
RD2	c.6698A>T:p(E2233V)	Novel	-	P	1.35	0.71	Loss of H-bond: RD2-NBD1 interaction *

Reference genome assembly, GRCh38:Chr1:83457325–104273917, Reference Transcript: NM_000350.3. Informatics-based pathogenicity prediction is based on CADD, PolyPhen-2, REVEL, and Consurf amino acid conservation scores [45]. ǂ When there is no absolute consensus, given based on the majority's prediction. (See supp. Table 2). *The novel variant c.6698A>T:p(E2233V) was also predicted to impact splicing with significant alteration of auxiliary sequences ESE / ESS motifs ratio (−10; HSF-Pro [46]). The wild-type (WT) and variant protein models were predicted in the AlphaFold2 [35] through the ColabFold AlphaFold2_advanced notebook running in Google Colaboratory [39]. The in silico stability change prediction is based on the mean of ∆∆Gs calculated from all ABCA4 experimental structures (exp.) using the FoldX plugin for Yasara [40,41]. ∆∆G is the measure of the free energy change due to an amino acid substitution. ∆∆G ≥1, the substitution is defined as destabilizing. For ∆∆G values based on each protein structure, and their correlations, see supplementary files: Figure S3 and Table S3. All other structural analyses were performed in PyMOL2. Root-mean square displacement (RMSD; all-atom) is calculated to determine the structural alignment scores. VUS: Variants of Uncertain significance, CI: Conflicting interpretation, NBD: nucleotide-binding domain, ECD: Exocytoplasmic Domain, TMD: Transmembrane Domain, RD: Regulatory Domain.