Estimation of the methylation pattern distribution from deep sequencing data

BMC Bioinformatics

Table 1 Comparison between estimates \( {\widehat{\theta}}_i \) of the true methylation distribution θ _i calculated with the slow and fast implementations of our algorithm for the dataset of Figure 1

Patterns	θ _i	y _i	y _i / N _read	\( {\widehat{\theta}}_i \) (slow)	\( {\widehat{\theta}}_i \) (fast)
000000	0.50	907	0.4535	0.4813	0.4812
000001	0.00	15	0.0075	0.0013	0.0013
000010	0.00	8	0.0040	0	0
000100	0.00	7	0.0035	0	0
001000	0.04	99	0.0495	0.0466	0.0466
001001	0.00	4	0.0020	0.0016	0.0014
001010	0.00	1	0.0005	0	0
001100	0.00	1	0.0005	0	0
010000	0.00	6	0.0030	0	0
011000	0.00	2	0.0010	0.0007	0.0004
011011	0.00	1	0.0005	0.0005	0.0002
011101	0.00	2	0.0010	0	0
011111	0.03	63	0.0315	0.0308	0.0306
100000	0.03	77	0.0385	0.0328	0.0329
101101	0.00	1	0.0005	0	0
101111	0.00	1	0.0005	0	0
110000	0.00	1	0.0005	0	0.0000
110111	0.00	2	0.0010	0	0
111001	0.00	1	0.0005	0	0
111011	0.00	1	0.0005	0	0
111100	0.00	3	0.0015	0	0
111101	0.20	393	0.1965	0.2001	0.2013
111110	0.00	3	0.0015	0	0
111111	0.20	401	0.2005	0.2044	0.2040

y _i is the number of observed reads for pattern i and N _read is the total number of reads.

ISSN: 1471-2105