Characterizing and measuring bias in sequence data

Table 2 Data sets and their relative coverage on bias motifs

					Relative coverage
Data set					GC extremes			Special motifs
Sample	#	Library method	Sequencing platform	Coverage (x)	GC ≤ 10%	GC ≥ 75%	GC ≥ 85%	(AT)¹⁵	G\|C ≥ 80%	Bad promoters
P. falciparum	1	Fisher et al.^a with Kapa reagents	Illumina MiSeq	150	0.58	-	-	0.43	-	-
3D7	2	Ion Torrent standard	Ion Torrent PGM	103	0.39	-	-	0.11	-	-
	3	Pacific Biosciences standard	Pacific Biosciences RS	104	0.89	-	-	0.85	-	-
E. coli	4	Fisher et al.^a with Kapa reagents	Illumina MiSeq	380	-	0.82	-	-	-	-
K12 MG1655	5	Ion Torrent standard	Ion Torrent PGM	311	-	0.31	-	-	-	-
	6	Pacific Biosciences standard	Pacific Biosciences RS	115	-	0.97	-	-	-	-
R. sphaeroides	7	Fisher et al.^a with Kapa reagents	Illumina MiSeq	388	-	0.94	0.60	-	-	-
2.4.1	8	Ion Torrent standard	Ion Torrent PGM	302	-	0.39	0.10	-	-	-
	9	Pacific Biosciences standard	Pacific Biosciences RS	142	-	0.97	0.87	-	-	-
Human	10	Aird et al. with Phusion	Illumina HiSeq v2	028	0.58	0.27	0.071	0.38	0.19	0.027
NA12878	11	Aird et al. with Phusion+betaine	Illumina HiSeq v2	048	0.44	0.44	0.28	0.26	0.20	0.14
	12	Aird et al. with AccuPrime	Illumina HiSeq v2	075	0.42	0.42	0.23	0.23	0.38	0.16
	13	Fisher et al.^a	Illumina HiSeq v3	070	0.29	1.1	0.56	0.23	0.44	0.39
	14	Fisher et al.^a with Kapa reagents	Illumina HiSeq v3	120	0.41	0.88	0.48	0.25	0.65	0.36
	14'	Fisher et al.^a with Kapa reagents	Illumina HiSeq v3	000.5	0.41 ± 0.0032	0.88 ± 0.0047	0.48 ± 0.0067	0.25 ± 0.0042	0.65 ± 0.012	0.37 ± 0.022
	15	Ion Torrent standard	Ion Torrent PGM	001.1	0.27	0.36	0.068	0.19	0.26	0.046
	16	Complete Genomics standard	Complete Genomics	079	0.24	0.53	0.18	0.28	0.61	0.092

^aLow-input variation of Fisher et al. [31] (see Materials and methods). Data sets from samples, library construction methods and sequencing platforms are shown, along with their total coverage of the genome, and relative coverage, for each of five bias motifs and a set of 'bad promoters' (see text). Entries are blank if the samples' genome had no instances of the given motif. Data set 14' is the summary of ten random subsamplings from data set 14, with coverage reduced to 0.5×, and we show the mean and standard deviations for the relative coverage measurements from it (see text).

ISSN: 1474-760X