MassComp, a lossless compressor for mass spectrometry data

BMC Bioinformatics

Table 1 Results for FPC, gzip and MassComp when tested on the m/z-intensity pairs of some of the considered files, both in single and double precision

	mzXML	Pairs	FPC	C.R.	gzip	C.R.	MassComp	C.R.	Gain gzip	Gain FPC
MSV000080896 (single-precision)
9122	25	18	17.71	1.61	12	33.33	9.4	47.78	21.67	46.92
9123	25	18	17.68	1.78	12	33.33	9.4	47.78	21.67	46.83
9124	25	18	17.97	0.17	13	27.78	9.5	47.22	26.92	47.13
MSV000080905 (single-precision)
54	7.1	4	3.84	4.00	3.7	7.50	2.8	30.00	24.32	27.08
05	7.7	4.4	3.71	15.68	4.1	6.82	3.1	29.55	24.39	16.44
07	9	4.7	4.48	4.68	4.3	8.51	3.2	31.91	25.58	28.57
MSV000081123 (single-precision)
neg42	11	5.1	4.91	3.73	4.5	11.76	3.5	31.37	22.22	28.72
pos75	6.7	3.6	3.43	4.72	3.2	11.11	2.5	30.56	21.88	27.11
tag71	3.9	1.7	1.51	11.18	1.5	11.76	1.1	35.29	26.67	27.15
Raw Data (double-precision)
110	69	50	41.56	16.88	21	58	11	78	47.62	73.53
121	2,500	1,900	1,257	33.84	986	48.11	856	54.95	13.18	31.90
ADH	1,900	1,500	1,082	27.85	320	78.67	154	89.73	51.88	85.77
Average single-precision
				5.28		16.88		36.83	23.92	32.89
Average double-precision
				26.19		61.59		74.23	37.56	63.73
Average all
				10.51		28.06		46.18	27.33	40.60

All sizes are expressed in MBs. Column mzXML denotes the size of the mzXML file, whereas Pairs contains the size of only the numerical data (including possible padding bits). Best results are highlighted in bold. C.R. denotes compression ratio, computed as 100−compressed size∗100/size of pairs. The gain of MassComp is computed as 1−MassComp size/other size.

ISSN: 1471-2105