Parametric modeling of cellular state transitions as measured with flow cytometry

BMC Bioinformatics

Table 1 Details of competing models for Clb5 data

t	Criterion	NMIX	TMIX	SNMIX	STMIX	STNMIX
0	${\hat{ℓ}}_{max}$	2539.78	2647.35	2682.44	2771.12	2759.40
	BIC	-5033.60	-5239.55	-5300.53	-5468.69*	-5445.25
	D _n	0.0413	0.0262	0.0292	0.0164^†	0.0185
10	${\hat{ℓ}}_{max}$	1201.11	1224.87	1357.82	1405.31	1406.80
	BIC	-2356.27	-2394.60	-2651.32	-2737.09	-2740.08*
	D _n	0.0424	0.0284	0.0312	0.0214	0.0190^†
20	${\hat{ℓ}}_{max}$	-5463.40	-5462.64	-4869.75	-4792.67	-4791.62
	BIC	10972.72	10980.37	9803.79	9658.80	9656.69*
	D _n	0.0758	0.0715	0.0251^†	0.0264	0.0266
25	${\hat{ℓ}}_{max}$	-7040.90	-6981.07	-6992.27	-6918.17	-6916.52
	BIC	14127.73	14017.26	14048.84	13909.82	13906.53*
	D _n	0.0147	0.0145	0.0155	0.0077	0.0075^†
30	${\hat{ℓ}}_{max}$	-7251.45	-7226.16	-7228.28	-7203.05	-7201.55
	BIC	14548.76	14507.34	14520.76	14479.46	14476.46*
	D _n	0.0218	0.0175	0.0143	0.0129	0.0110^†
35	${\hat{ℓ}}_{max}$	-6413.58	-6412.58	-6374.92	-6320.38	-6334.20
	BIC	12872.96	12880.12	12813.96	12714.04*	12741.69
	D _n	0.0196	0.0230	0.0136	0.0117^†	0.0152
40	${\hat{ℓ}}_{max}$	-4626.43	-4625.80	-4546.10	-4429.22	-4461.12
	BIC	9298.55	9306.44	9156.18	8931.56*	8995.37
	D _n	0.0338	0.0306	0.0170	0.0123^†	0.0184
50	${\hat{ℓ}}_{max}$	-1286.86	-1121.26	-1286.53	-1093.80	-1086.34
	BIC	2619.40	2297.35	2637.03	2260.70	2245.79*
	D _n	0.0222	0.0145	0.0218	0.0139	0.0132^†
60	${\hat{ℓ}}_{max}$	-2016.29	-1596.75	-1835.23	-1573.97	-1568.82
	BIC	4078.18	3248.21	3734.30	3220.89	3210.59*
	D _n	0.0540	0.0203	0.0339	0.0172	0.0131^†
75	$\hat{ℓ_{max}}$	-8146.57	-7810.70	-7772.60	-7770.74	-7769.87
	BIC	16393.86	15731.29	15682.55*	15688.00	15686.25
	D _n	0.0219	0.0119	0.0079^†	0.0101	0.1104

Details of competing models for Clb5Δ data. Here, the notations stand for log-likelihood maxima ${\hat{ℓ}}_{max}$ max, BIC values, and distances D_n based on Kolmogorov-Smirnov (K-S) test statistic. The abbreviation of models are the normal mixtures (NMIX), the t mixtures (TMIX), the skew-normal mixtures (SNMIX), the skew-t mixtures (STMIX) and the skew-t-normal mixtures (STNMIX), respectively. According to BIC and D_n, the optimal selection of the STNMIX model with equal dfs is evident for most points. The smallest values of BIC and D_n are indicated by * and ^†, respectively.

ISSN: 1471-2105