. 2023 May 9;75(7):1088–1097. doi: 10.1002/art.42464

Table 3.

Results of the tensor composition analysis of differentially methylated positions

CpG	chr:bp (hg19)	GREAT annotation^*	Model^†	Cell type	Δ Methylation	Joint model P ^‡	Marginal conditional model P
cg13249593	chr17:39705234	KRT19 (–20,674); KRT9 (+23,076)	T₀: R–NR	CD4T CD8T	0.226–0.310	1 × 2.4⁻⁸ 1 × 9.5⁻³ ^§	1 × 4.6⁻⁶ 1 × 1.8⁻⁵
				NK	–0.489		1 × 2.3⁻⁶

CpG

chr:bp (hg19)

GREAT annotation^*

Model^†

Cell type

Δ Methylation

Joint model P ^‡

Marginal conditional model P

cg13249593

chr17:39705234

KRT19 (–20,674);

KRT9 (+23,076)

T₀: R–NR

CD4T

CD8T

0.226–0.310

1 × 2.4⁻⁸

1 × 9.5⁻³ ^§

1 × 4.6⁻⁶

1 × 1.8⁻⁵

–0.489

1 × 2.3⁻⁶

Distance from CpG to transcription start site of nearest upstream and downstream gene annotated using rGREAT software. See Table 2 for definitions.

^†

Model comparing difference in DNA methylation at baseline before treatment with methotrexate between rheumatoid arthritis patients identified as treatment responders compared to nonresponders.

^‡

Joint model in tensor composition analysis tested for evidence of differential DNA methylation within any cell type at each CpG.

^§

P for false discovery rate.