MassComp, a lossless compressor for mass spectrometry data

BMC Bioinformatics

Table 2 Results for gzip and MassComp when tested on whole mzXML

	Num. Files	Average	mzXML	gzip	C.R. [std]	MassComp	C.R. [std]	Gain gzip
MSV000080896	9	34	299	168	43.81 [0.39]	113	62.21 [0.94]	32.74
MSV000080905	514	8	4,100	2,500	39.02 [1.62]	1,800	56.10 [0.81]	28.00
Plate1	123	7.96	979	591	39.63 [1.33]	455	53.52 [0.74]	23.01
Plate2	132	8.33	1,100	633	42.45 [1.59]	486	55.82 [0.73]	23.22
Plate3	135	8.15	1,100	690	37.27 [1.61]	528	52.00 [0.84]	23.48
Plate4	124	8.06	999	608	39.14 [1.79]	467	53.25 [0.88]	23.19
MSV000081123	54	8	406	218	46.31 [3.01]	165	59.36 [1.90]	24.31
Average					43.05 [2.95]		59.22 [2.22]	28.35

Results show the total size of the compressed files when considering all mzXML files within each MassIVE experiment, as well as the compression ratio (denoted by C.R.). We also included the standard deviation, denoted by std. The compression ratio and the gain of MassComp with respect to gzip is computed as in Table 1. Results for experiment MSV000080905 are split into different Plates, as provided in the MassIVE repository, in addition to the overall performance. All sizes are expressed in MBs.

ISSN: 1471-2105