Reducing the haystack to find the needle: improved protein identification after fast elimination of non-interpretable peptide MS/MS spectra and noise reduction

Mujezinovic, Nedim; Schneider, Georg; Wildpaner, Michael; Mechtler, Karl; Eisenhaber, Frank

doi:10.1186/1471-2164-11-S1-S13

Table 2 Performance of the MSCleaner version 2.0 over a large test set.

From: Reducing the haystack to find the needle: improved protein identification after fast elimination of non-interpretable peptide MS/MS spectra and noise reduction

A1	A2	A3	A4	A5	A6	A7	A8	A9	A10	A11	A12	A13	A14	A15	A16
alphaAmyl_col1	10108	633	24	11.30	667	24	31.65	60.07	667	24	15.13	51.09	667	24	18.07
alphaAmyl_col2	10184	698	35	9.82	780	35	34.20	50.22	780	35	19.05	20.25	780	35	22.76
AmylGlu_col1	10030	736	28	13.26	761	28	28.40	79.24	761	28	8.66	73.58	761	28	10.63
AmylGlu_col2	9870	801	36	13.31	860	37	29.50	72.62	860	37	11.70	63.95	860	37	14.29
apo_col1	10032	2606	63	11.72	2814	63	30.76	63.10	2814	63	13.93	54.49	2814	63	16.78
apo_col2	10090	2571	60	12.13	2761	60	32.95	53.12	2761	60	17.53	44.32	2761	60	21.03
betaGal_col1	10324	1459	56	7.17	1567	57	34.98	48.06	1567	57	22.05	40.53	1567	57	24.60
betaGal_col2	10368	1309	51	8.12	1508	56	36.71	42.90	1454	55	24.76	33.10	1454	55	28.61
CarAnly_col1	9946	586	49	12.35	616	49	26.35	90.31	573	49	3.65	84.94	607	49	5.48
CarAnly_col2	9534	582	52	13.40	616	52	26.27	86.07	616	52	5.08	78.44	616	52	7.66
Cat_col1	10098	1798	61	11.13	1886	61	30.88	67.26	1879	61	13.13	57.89	1879	61	16.50
Cat_col2	10034	1567	65	11.78	1693	65	31.90	59.50	1693	65	15.91	48.55	1693	65	19.56
phosB_col1	10118	2780	59	10.30	3079	61	35.13	63.49	3014	60	14.26	54.46	3047	61	17.25
phosB_col2	10096	2655	61	10.52	3116	65	32.58	53.96	3084	65	17.58	44.31	3116	65	21.16
GluDey_col1	10006	892	36	11.29	986	36	27.30	79.55	986	36	7.75	73.42	986	36	9.71
GluDey_col2	9886	850	34	11.81	962	34	28.73	72.51	962	34	10.13	62.25	962	34	13.51
GluTra_col1	10022	351	25	10.36	389	25	28.61	71.64	348	25	10.25	62.78	389	25	14.30
GluTra_col2	10156	341	33	9.18	384	33	31.31	61.15	384	33	14.25	49.59	384	33	28.11
Immo_col1	10330	506	35	9.27	565	35	36.20	42.30	565	35	24.95	34.44	565	35	27.66
Immo_col2	10334	356	66	8.61	500	66	38.05	37.06	500	66	27.31	28.47	500	66	30.31
LacDe_col1	10286	1549	58	10.36	1694	58	35.36	53.20	1694	58	20.03	44.86	1694	58	23.15
LacDe_col2	10250	1346	54	9.07	1483	54	36.48	40.16	1483	54	25.60	31.67	1483	54	28.31
LactoPee_col1	10242	1613	45	13.16	1764	45	34.78	62.12	1756	45	15.91	52.37	1764	45	19.53
LactoPee_col2	10402	1679	43	9.09	1890	44	35.18	51.70	1890	44	20.31	41.76	1890	44	23.85
Myo_col1	9958	561	66	11.67	594	66	27.26	85.42	594	66	5.46	79.25	594	66	7.45
Myo_col2	9744	530	66	12.15	584	66	28.01	80.83	584	66	6.95	70.92	584	66	10.35

A1 name of test set (.mgf file; see Methods),
A2 total number of spectra (.dta files),
A3 MASCOT score of top protein hit with the original .mgf file
(without application of MS Cleaner),
A4 sequence coverage (in %) without application of MS Cleaner,
A5 fraction of non-interpretable "bad" spectra found with sequence ladder
length n= 4 among all peaks (intensity threshold s= 100%)
A6 MASCOT score of the top protein hit for this search,
A7 sequence coverage (in % of the whole protein length) for this search,
A8 MS Cleaner processing time (in min) on a PC with a single Pentium IV (to
achieve exact time consumption values, we did not use the cluster version and
stopped the "soft frequency recognition option")
A9 fraction of non-interpretable "bad" spectra found with sequence ladder
length n= 4 among the s= 20% most intense peaks
A10 MASCOT score of the top protein hit for this search,
A11 sequence coverage (in % of the whole protein length) for this search,
A12 MS Cleaner processing time (in min),
A13 fraction of non-interpretable "bad" spectra found with sequence ladder
length n= 4 among the s= 25% most intense peaks (in % of A2; i.e.,
of all spectra)
A14 MASCOT score of the top protein hit for this search,
A15 sequence coverage (in % of the whole protein length) for this MASCOT
search,
A16 MS Cleaner processing time on the same machine as described in the legend of
Table 1 (in min).
The sequence ladder criterion (minimal ladder length 4 with varying peak intensity thresholds) and the noise suppression algorithms of MS Cleaner 2.0 have been applied over a large set of tandem MS results. For each of the test proteins, two independent sample preparations and dataset recordings (marked with appendices _col1 and _col2 in the dataset name) were carried out: α-amylase, amylogucosidase, apo-transferrin, β-galactidase, carbonic anhydrase, catalase, phosphorylase B, glutamic dehydrogenase, glutathione transferase, immunoglobulin γ, lactic dehydrogenase, lactoperoxidase, myoglobin). For these datasets, the MASCOT interpretation was carried out on a cluster in parallel with other jobs; therefore, no computation time is provided.

Back to article page

ISSN: 1471-2164

Contact us

Submission enquiries: bmcgenomics@biomedcentral.com
General enquiries: ORSupport@springernature.com

BMC Genomics

Contact us