. 2023 Sep 4;95:104769. doi: 10.1016/j.ebiom.2023.104769

Table 1.

Composition of the 5 datasets used in the study.

Dataset	Contributors	cell phenotypes	No. of annotations		Total no. of annotations
Dataset	Contributors	cell phenotypes	Training + Validation	Testing	Training + Validation	Testing
Ovarian T cells	T.G and J.A.L¹⁹	CD4+FOXP3+	292	197	1828 (4 slides)	997 (5 slides)
		CD4+FOXP3-	596	168
		PD1+CD8+	726	347
		PD1-CD8+	139	203
		PD1+CD4+	39	60
		PD1+CD8-CD4-	36	22
LUSC T cells	T.M, A.U.A, and J.L.Q	CD4+FOXP3+	746	228	2407 (2 slides)	1383 (2 slides)
		CD4+FOXP3-	1225	696
		CD8+	204	200
		Haematoxylin-stained	232	259
DCIS FOXP3	F.S and E.S.H²⁰	FOXP3+	1030	576	11,459 (7 slides)	3799 (5 slides)
DCIS FOXP3	F.S and E.S.H²⁰	FOXP3−	10,429	3223	11,459 (7 slides)	3799 (5 slides)
Myeloma	Y.H, C.S.Y.L, D.P, L.L, M.R-J and K.Y²¹	CD8+	866	979	3269 (4 slides)	1588 (2 slides)
		CD4+FOXP3−	2244	493
		CD4+FOXP3+	159	116
IMC CD4-CD8	Damond et al.²²	CD4+	987	828	3954 (80 slides)	1085 (20 slides)
IMC CD4-CD8	Damond et al.²²	CD8+	2967	257	3954 (80 slides)	1085 (20 slides)