supplementary figures - cornell...

Supplementary Figures

Supplementary Figure 1. Comparative evaluation of different XL-MS acquisition strategies at the

redundant CSM level at 1% FDR. Significance was determined by a two-sided Z-test. (analysis at unique

CSM level is shown in Fig. 1).

Supplementary Figure 2. Quality assessment of cross-links from multiple XL-MS acquisition strategies

utilizing false search space (left panel) and precision (right panel). CSMs were identified using XlinkX at

1%FDR with additional filtering cut-offs for ‘XlinkX Score’.

Supplementary Figure 3. Dissection of CSMs from the ensemble i.e., CID-MS2-MS3-ETD-MS2

approach at 1% FDR in terms of quality of the two inherent approaches (i.e., CID-MS2-MS3 and CID-

MS2-ETD-MS2). Venn Diagrams show the overlap of CSMs at 1% FDR, along with further filtering at

different ‘XlinkX Score’ cut-offs. Left panel shows CSM overlap between the two approaches (fraction

of mis-identifications is shown in the parentheses) and the right panel shows overlap for the inter-protein

cross-links (precision is given in parentheses). XlinkX was used to process the raw files for six E. coli CID-

MS2-MS3-ETD-MS2 XL-MS fractions from Liu et al9.

Supplementary Figure 4. Venn diagrams showing the overlap of CSMs from MaXLinker at 1% FDR with

CSMs identified by XlinkX at 1% FDR with further filtering at different ‘XlinkX Score’ cut-offs for

XlinkX (using the same set of raw files for six E. coli MS2-MS3 XL-MS fractions from Liu et al9). Left

panel shows results for analysis using false search space (fraction of mis-identifications given in parentheses

below the number of CSMs) and the right panel represents precision-based analysis (precision given in

parentheses below the number of interprotein CSMs).

Supplementary Figure 5. Illustration of MaXLinker’s ‘MS2 Rescue’ module, in a case where the precursor

mass validation step fails along with uninformative MS3 spectra for either of the peptides

Supplementary Figure 6. Illustration of MaXLinker’s ‘MS2 Rescue’ module, in a case where the

precursor mass validation step fails along with all four MS3 spectra potentially representing different

charge states of one peptide.

P = 1.1 x 10-9

Fraction of m

is-identifications(%)

All

CSM

s

500

10

30

00

1000

1500

2000

20P = 7.9 x 10-9

P = 2.2 x 10-4

MS3

-Onl

y

CID-

MS2

-ETD

-MS2

CID-

MS2

-MS3

CID-

MS2

-MS3

-ETD

-MS2

Supplementary Figure 1

∆ XlinkX Score ≥ 10

∆ XlinkX Score ≥ 20

∆ XlinkX Score ≥ 40

∆ XlinkX Score ≥ 50

Fraction of m

is-identifications(%)

Fraction of m

is-identifications(%)

Fraction of m

is-identifications(%)

Fraction of m

is-identifications(%)

Fraction of m

is-identifications(%)

10

30

0

20

40

10

30

0

20

40

10

30

0

20

40

10

30

0

20

40

10

30

0

20

40

Uni

que

CSM

s

200

0

400

600

200

0

400

200

0

400

200

0

400

200

0

400

200

0

400

600

Precision (%)

25

75

0

50

100

200

0

400

Precision (%)

25

75

0

50

100

200

0

400

Precision (%)

25

75

0

50

100

200

0

400

Precision (%)

25

75

0

50

100

200

0

400

Precision (%)

25

75

0

50

100

MS3

-Onl

yCI

D-M

S2-E

TD-M

S2CI

D-M

S2-M

S3CI

D-M

S2-M

S3-E

TD-M

S2

∆ XlinkX Score ≥ 30

Uni

que

CSM

sU

niqu

e CS

Ms

Uni

que

CSM

sU

niqu

e CS

Ms

Uni

que

CSM

sU

niqu

e CS

Ms

Uni

que

CSM

sU

niqu

e CS

Ms

Uni

que

CSM

s

MS3

-Onl

yCI

D-M

S2-E

TD-M

S2

CID-

MS2

-MS3

CID-

MS2

-MS3

-ETD

-MS2

Supplementary Figure 2

∆ XlinkX Score ≥ 10

All

∆ XlinkX Score ≥ 20

∆ XlinkX Score ≥ 30

∆ XlinkX Score ≥ 40

∆ XlinkX Score ≥ 50

MS2-MS2 MS2-MS3

386(7.5%)

142(37.3%)

157(65.0%)

MS2-MS2 MS2-MS3

303(2.3%)

124(19.4%)

61(36.1%)

MS2-MS2 MS2-MS3

244(2.1%)

149(10.7%)

25(8.0%)

188(0.0%)

180(7.8%)

18(5.6%)

MS2-MS2 MS2-MS3

137(0.0%)

208(6.3%)

12(0.0%)

MS2-MS2 MS2-MS3

101(0.0%)

221(5.4%)

14(0.0%)

MS2-MS2 MS2-MS3 MS2-MS2 MS2-MS3

MS2-MS2 MS2-MS3

MS2-MS2 MS2-MS3

MS2-MS2 MS2-MS3

MS2-MS2 MS2-MS3

57(50.9%)

43(14.0%)

72(4.2%)

31(48.4%)

2(100%)

14(92.9%)

MS2-MS2 MS2-MS3

29(48.3%)

2(100%)

16(87.5%)

27(37.0%)

6(83.3%)

23(82.6%)

28(28.6%)

5(60.0%)

31(74.2%)

26(19.2%)

25(12.0%)

43(65.1%)

Supplementary Figure 3

∆ XlinkX Score ≥ 10

∆ XlinkX Score ≥ 20

∆ XlinkX Score ≥ 30

∆ XlinkX Score ≥ 40

∆ XlinkX Score ≥ 50

37(83.7%)

46(17.4%)

38(50.0%)

36(86.1%)

30(23.3%)

39(48.7%)

36(86.1%)

24(25.0%)

39(48.7%)

309(0.7%)

148(18.9%)

84(1.2%)

296(0.3%)

110(12.7%)

97(2.1%)

288(0.4%)

94(10.6%)

105(1.9%)

280(0.4%)

78(7.7%)

113(1.8%)

275(0.4%)

66(6.1%)

118(1.7%)

35(85.7%)

22(27.3%)

40(50.0%)

34(85.3%)

20(25.0%)

41(51.2%)

MaXLinker

MaXLinker

MaXLinker

MaXLinker

MaXLinker

MaXLinker

MaXLinker

MaXLinker

MaXLinker

MaXLinkerXlinkX v2.0

XlinkX v2.0

XlinkX v2.0

XlinkX v2.0

XlinkX v2.0 XlinkX v2.0

XlinkX v2.0

XlinkX v2.0

XlinkX v2.0

XlinkX v2.0

Supplementary Figure 4

Peptide A

Peptide B Uninformative MS3 spectrum Uninformative MS3 spectrum

AShortScan: 21166m/z: 574.31909z: 2+m: 1147.63091

ALongScan: 21167m/z: 590.30536z: 2+m: 1179.60344

BShortScan: 21170m/z: 565.81036z: 2+m: 1130.61345

BLongScan: 21171m/z: 581.79999z: 2+m: 1162.59270

precursor mass - (AShort + BLong + H2O - H+) = 479.12078precursor mass - (ALong+ BShort + H2O - H+) = 479.12750

Precursorm/z: 702.34265

z: 4+m: 2806.34713

QYLT[K]ELAK

EFADNLDSDF[K]VR

Peptide B

BShortm/z: 805.37466z: 2+m: 1609.74204

Derived BLongm/z: 821.36093z: 2+m: 1641.71458

MS2 rescue m

odule

Derived BShort = precursor mass - (ALong+ H2O - H+)Derived BLong = precursor mass - (AShort+ H2O - H+)

PSM search on MS2 spectrum oncewith each derived (‘Long’ and ‘Short’ )

mass as precursor

Validate derived masses in the MS2 spectrum

High con�dencePSMs

MaXLinker’s further validation �lters

Intraprotein cross-link in E. coli’s 30S ribosomal protein S3

(uniprot: P0A7V3)

Supplementary Figure 5

Peptide A

Peptide B

AShortScan: 12705 m/z: 639. 36523 z: 2+m: 1277. 72319

ALongScan: 12704m/z: 655.35364z: 2+m: 1309.70000

BLongScan: 12706m/z: 437. 23773 z: 3+m: 1309. 69864

BShortScan: 12707m/z: 426. 58060z: 3+m: 1277. 72724

precursor mass - (AShort + BLong + H2O - H+) = 351. 22760 precursor mass - (ALong+ BShort + H2O - H+) = 351.23301

Precursorm/z: 564.05511

z: 4+m: 2253.19697

VNAL[K]EQA

LTAH[K]NAVTLR

Peptide A

Derived AShortm/z: 926.49669 z: 1+m: 926. 49669

Derived ALongm/z: 958.46808z: 1+m: 958.46808

MS2 rescue m

odule

Derived AShort = precursor mass - (BLong+ H2O - H+)Derived ALong = precursor mass - (BShort+ H2O - H+)

PSM search on MS2 spectrum oncewith each derived (‘Long’ and ‘Short’ )

mass as precursor

High con�dencePSMs

MaXLinker’s further validation �lters

Intraprotein cross-link in E. coli’s Histidinol dehydrogenase

(uniprot: P06988)

Validate derived masses in the MS2 spectrum

Supplementary Figure 6