Fig. 3

DCATS gives accurate conclusions for two experimental datasets. “dcats_autoRef” indicates using the reference group automatically detected by DCATS as the normalization term. A shown is the proportion of each cell types in the Kang dataset [4]. cM, CD14+CD16- monocytes; ncM, CD14+CD16+ monocytes; DC, dendritic cells; Mkc, megakaryocytes; Th, CD4+ T cells; B, B cells; Tc, CD8+ T cells; NK, natural killer cells. B–D show the proportion of each cell types in the Haber dataset [28]. E, Enterocyte; TA, transit amplifying; TAE, TA.Early; EP, Enterocyte.Progenitor; Gob, Goblet. “P” in the first line represents existing proportions’ difference according to the original papers, and “N” represents no significant proportions’ difference. “P” in the last line represents existing proportions’ difference based on results of milo, and “N” represents no significant proportions’ difference. In the rest four lines, “*” represents p-values from 0.05 to 0.1, “**” represents p-values from 0.01 to 0.05, “***” represents p-values less than 0.01. “n.s.” means not significant