Ion-pair selection method for pseudotargeted metabolomics based on

Aug 27, 2018 - It is well known that sequential windowed acquisition of all theoretical Fragment ion (SWATH) MS mode can acquire MS2 information in a ...
3 downloads 0 Views 1MB Size
Subscriber access provided by UNIV OF DURHAM

Article

Ion-pair selection method for pseudotargeted metabolomics based on SWATH MS acquisition and its application in differential metabolite discovery of Type 2 diabetes Lichao Wang, Benzhe Su, Zhongda Zeng, Chao Li, Xinjie Zhao, Wangjie Lv, Qiuhui Xuan, Yang Ouyang, Lina Zhou, Peiyuan Yin, Xiaojun Peng, Xin Lu, Xiaohui Lin, and Guowang Xu Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b02377 • Publication Date (Web): 27 Aug 2018 Downloaded from http://pubs.acs.org on August 28, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 25 1

Ion-pair selection method for pseudotargeted metabolomics based on SWATH MS acquisition and its application in differential metabolite discovery of Type 2 diabetes

Lichao Wang1,3,4, Benzhe Su2, Zhongda Zeng1, Chao Li2, Xinjie Zhao1,4, Wangjie Lv1,4, Qiuhui Xuan1,4, Yang Ouyang1,4, Lina Zhou1,4, Peiyuan Yin1,4, Xiaojun Peng3, Xin Lu1,4, Xiaohui Lin2, Guowang Xu1,4*

1

CAS Key Laboratory of Separation Science for Analytical Chemistry, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, 457 Zhongshan Road, Dalian 116023, China.

2

School of Computer Science & Technology, Dalian University of Technology, 116024, Dalian, China.

3

State Key Laboratory of Fine Chemicals, Dalian University of Technology, Dalian 116023, China.

4

University of Chinese Academy of Sciences, Beijing 100049, China.

* Address correspondence to: Prof. Dr. Guowang Xu, CAS Key Laboratory of Separation Science for Analytical Chemistry, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China. Tel. / Fax: 0086-411-84379530. E-mail: [email protected].

ACS Paragon Plus Environment

Page 3 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry 2

ABSTRACT Pseudotargeted metabolomics method integrates advantages of non-targeted and targeted analysis because it can acquire data of metabolites in the multi-reaction monitoring (MRM) mode of mass spectrometry (MS) without needing standards. The key is the ion-pair information collection from samples to be analyzed. It is well known that sequential windowed acquisition of all theoretical Fragment ion (SWATH) MS mode can acquire MS2 information in a maximum extent. To expediently acquire as many ion-pairs as possible with optimal collision energy (CE), an ion-pair selection approach based on SWATH MS acquisition with variable isolation windows was developed in this study. Initially, non-targeted acquisition of all metabolites information in plasma Standard Reference Material (SRM 1950) was performed by ultra-performance

liquid

chromatography

(UHPLC)-quadrupole

time-of-flight

(Q-TOF) MS platform with three CEs. With the help of software tool, the ion-pairs of unique metabolites were gained. Then they were validated in scheduled MRM coupled with UHPLC. After removing false positive, the ion-pairs with an optimal CE was integrated. A total of 1373 unique metabolite ion-pairs were obtained at positive ion mode. And repeatability of the established pseudotargeted approach was evaluated by intraday and interday precision. The results demonstrated the method was stable, reliable and suitable for metabolomics study. As an application example, alterations of serum metabolites in Type 2 diabetes were investigated by using the established method. This work provides a pseudotargeted ion-pair selection method based on SWATH MS acquisition with the characters of increased metabolite coverage, suitable CE and convenient processing.

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 25 3

INTRODUCTION Metabolomics is focused on endogenous small molecular compounds in biological system, and it can systematically reflect the alterations when conditions of body have changed, including environmental stimuli, genome mutation, pathological stimuli1-3. Metabolomics analysis approaches mainly consists of non-targeted method and targeted method4-6. Nowadays, the pseudotargeted metabolomics method based on mass spectrometry (MS) has been proposed to combine the superiorities of targeted and non-targeted methods7. It had been applied to many metabolomics studies on diseases, such as hepatocellular carcinoma8,9, bladder cancer10, oral squamous cell carcinoma11, and showed special advantages in biomarker discovery and pathology explanation8-11 by using gas chromatography-MS and ultra-high performance liquid chromatography (UHPLC)-MS. Pseudotargeted method acquires ion-pairs from real samples and its aim is to detect as many metabolites as possible in MRM mode

7,8,11

. Its establishment contains two

main steps: 1) MS1 and MS2 information collection of metabolites in the non-targeted mode; 2) defining ion-pairs of individual metabolites for further MRM use. Some researchers also reported ion-pair acquisition methods based on stepped targeted MS/MS methods12,13. Globally optimized targeted (GOT)-MS12 used LC-QQQ to acquire precursor ions and product ions on both reversed-phase (RP) and HILIC columns. Precursor ions were detected by performing selected ion monitoring (SIM) incremental scans with an m/z step 0.5 Da. However, in one LC injection only an m/z range of 30 could be measured (60 SIMs in one cycle) in SIM mode. After that, product ions were profiled by product ion scans. Data independent targeted quantitative metabolomics (DITQM) study13 applied a sequentially stepped targeted MS/MS (sst-MS/MS) method to acquire ion-pairs with an m/z step 1 Da in

ACS Paragon Plus Environment

Page 5 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry 4

data-independent mode. But there was no information about the retention times of precursor ions, and limited MS scan rates could not complete the issue of fragmenting all the compounds in one scan cycle. Thus in one LC injection, approximately 30 precursor ions could obtain their product ions. In above two methods, many LC injections were needed to acquire ion-pairs. Ion-pair acquisition based on information dependent acquisition (IDA) MS mode is more convenient, which can simultaneously obtain the MS spectra of precursor ions and corresponding product ions. To improve the efficiency of the process of selection ion-pairs from tedious IDA raw data, a software was developed in previous study14. However, IDA MS mode is in favor of the MS2 spectra of high intensity ions. Hence, establishment of a novel pseudotargeted ion-pair selection method with a larger metabolite coverage and convenient data processing has important meanings in metabolomics study. Sequential Windowed Acquisition of All Theoretical Fragment Ion (SWATH) is a novel MS technology to gain fragmental information of all precursor ions in theory15. The technology is gradually applied to different fields of metabolomics16-19. For example, Arnhard et al.17 carried out a systematic toxicological research based on LC-MS/MS with SWATH, and the results showed the SWATH acquisition could well detect and identify trace amounts of midazolam and morphine in the presence of interferences. Ma et al.18 reported an polyphenol metabolomics study in six commercial red wines based on SWATH MS. They revealed the links between the polyphenol contents and taste of red wines. Zhu et al. compared the identification abilities of IDA, SWATH and MSAll techniques in a total of 227 drug-related materials20. The result showed SWATH could obtain better quality of MS2 spectra than MSAll. Though its MS2 spectrum quality had a little decrease in comparison with IDA mode, the MS2 acquisition hit rate was obviously improved. Roemmelt et al.

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 25 5

reported that the detection rate of co-elution compounds in SWATH was better than IDA, and IDA failed to trigger about 10% of compounds in clinical and forensic toxicology studies21. SWATH MS technology could markedly improve the ability of MS2 spectra acquisition. When variable Q1 isolation windows are used in SWATH MS mode, it is a significant approach to improve the selectivity and reduce cross fragment ion interference, which has obvious benefits to identify the compounds22. “MS-DIAL” and “MetDIA” are two main softwares used to deal with the intricate SWATH data for metabolomics studies. “MS-DIAL” was reported to deal with the acquired data by mathematical deconvolution23. “MetDIA” was also performed for the acquired data by peak-peak correlation24. Considering the capability of SWATH-MS in abundant fragmental information acquisition, it is of great promise to use it for pseudotargeted method establishment of metabolomics25. In present study, we developed an ion-pair selecting method for pseudotargeted analysis based on UHPLC-SWATH MS with variable Q1 isolation windows and different CEs. Two small in-house software tools were used to simplify the data processing and improve work efficiency. To show the usefulness, the proposed method was ultimately applied to investigate serum metabolites changes related to Type 2 diabetes.

EXPERIMENTAL SECTION Chemicals and Reagents. HPLC grade acetonitrile and methanol were obtained from Merck (Darmstadt, Germany) and Formic acid was purchased from Sigma-Aldrich (St. Louis, USA). Ultrapure water was prepared by Milli-Q system (Millipore, Billerica, MA). Carnitine C2:0 - d3, Carnitine C8:0 - d3, Carnitine C16:0 - d3, valine – d8 (Val – d8), phenyl

ACS Paragon Plus Environment

Page 7 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry 6

alanine - d5 (Phe – d5), cholic acid - d4 (CA - d4) and chenodeoxycholic acid - d4 (CDCA - d4) supplied

by

Sigma-Aldrich

(St.

Louis,

MO,

USA)

and

lyso-phosphatidylcholine 19:0 (LPC 19:0) obtained from Avanti Polar Lipids (Alabaster, USA) were used as the internal standards,they were dissolved in methanol, and the concentrations are given in supplementary Table S1.

Samples Collection and Preparation The Standard Reference Material (SRM 1950 plasma) was purchased from National Institute of Standards and Technology (NIST) and further used to develop pseudotargeted metabolomics method. 100 µL SRM 1950 plasma was mixed with 400 µL ice-cold methanol in 1.5 mL of Eppendorf tube. And it was thoroughly vortexed for 30 s to remove protein and to extract the metabolites. After settled at 4 °C for 10 min, its centrifugation was achieved at 14000 rpm for 15 min at 4 °C. 400 µL supernatant was transferred to an Eppendorf tubes and lyophilized in CentriVap Centrifugal Vacuum Concentrators (Labconco, MO). 50 µL water/acetonitrile 4:1 (v/v), as the reconstitution solution, was added to the freeze-dried residue. After vortexed for 30 s and centrifuged at 14000 rpm for 10 min at 4 °C, the supernatant was applied to analyze metabolites by LC-MS. For metabolomics application study, 30 serum samples of diabetes patients were collected from Soochow University. 30 gender- and age-matched healthy samples were collected from Affiliated Zhongshan Hospital of Dalian University. An equal volume of serum samples from all the diabetes patients and control samples were mixed together for preparation of pooled quality control (QC) sample. The pretreatment process was the same as described for SRM 1950 plasma. One QC was inserted at every 10 samples in the analytical sequence. The QC sample was used to

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 25 7

evaluate the reliability of the whole experiment including sample preparation and UHPLC-QQQ/MS sequence run.

Data Acquisition with UHPLC-SWATH MS For non-targeted metabolomics analysis, an ultra high–performance liquid chromatography (UHPLC) system (Waters, Milford, MA, USA) coupled to a hybrid Q-TOF mass spectrometer (TripleTOF 5600+, AB SCIEX, Framingham, USA) was employed. 5 µL reconstitution sample was injected for chromatography separation on a RP Waters Acquity BEH C8 column (100 mm × 2.1 mm, 1.7 µm) at 50 °C and the flow rate was 0.35 mL/min. The mobile phase A was containing 0.1% (v/v) formic acid in ultrapure water, and the phase B was containing 0.1% (v/v) formic acid in acetonitrile, respectively. The binary solvent gradient conditions were listed as follows: first 5% B was kept for 1.0 min, then linearly increased to 100% B at 23 min, held 100% B for 4 min, returned back to 5% B in 0.1min, followed by equilibration with 5% B in next 2.9 min. The scan range of precursor ion m/z was set as 100-1250 Da in full scan mode. Then the distributions of precursor ions were used to calculate SWATH MS acquisition windows by SWATH Variable Window Calculator_V1.0. Eventually, SWATH MS acquisition was set as follows: the scan ranges of precursor ion m/z and product ion m/z were respectively set as 100-505, 503-751, 749-1250 Da and 80-505, 80-751, 80-1250 Da, respectively. The SWATH windows were shown in supplementary Table S2. In each SWATH MS duty cycle, MS1 accumulation time was 150 ms and MS2 accumulation time was 30 ms. The voltage of spray was set at 5500 V. The temperature of electrospray ion source was set as 500 °C. The CE voltage was in series set at 15 V, 30 V and 45 V in the positive electric spray ion (ESI) mode.

ACS Paragon Plus Environment

Page 9 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry 8

MRM Analysis by UHPLC-Q-trap/MS An UHPLC system (Waters, Milford, MA, USA) coupled to a Q-Trap spectrometer (Q-Trap 5500, AB SCIEX, Framingham, USA) was used to validate ion-pairs and to establish

pseudotargeted

metabolomics

method.

The

same

mobile

phase,

chromatographic column, separation gradient and LC injection volume were applied to the UHPLC-Q-trap/MS. The system was operated in dynamic scheduled MRM mode at ESI+ mode. The MS parameters were listed as follows: temperature of electrospray ion source was 500 °C; the voltage of spray was 5500 V; CE voltage was set at 15 V, 30 V, 45 V; target scan time was 0.8 s; MRM detection window was 50 s; during time was 25 min.

Data processing and statistical analysis Peak detection and alignment of the metabolites acquired by the UHPLC-SWATH MS were performed by using MarkerView software (AB SCIEX, Framingham, USA). And

MS-DIAL

software

was

used

to

obtain

information

(http://prime.psc.riken.jp/Metabolomics_Software/MS-DIAL/index2.html).

of

MS2

Ion-pair

selection was completed by homemade software on Matlab, and ion-pair integration was finished by homemade C-package, they can be downloaded free of charge (http://app.ifc.dicp.ac.cn/Confirmation/Authentication.html). UHPLC-MRM MS data were disposed in Analyst 1.6 software (AB SCIEX, Framingham, USA). For diabetes study, quantitative analysis was performed by MultiQuant 3.0.3 (AB SCIEX, Framingham, USA). The peak areas were normalized by suitable IS with the minimum relative standard deviation (RSD) for QC samples. Then the metabolites with RSD below 30% were used for further analysis. Partial least squares discriminant

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 25 9

analysis (PLS-DA) was used to handle the data in SIMCA-P software (version 11.0; Umetrics). Nonparametric Mann-Whitney U test performed in SPSS 18 (SPSS, Chicago, IL) was applied to evaluate the statistic alterations of metabolites’ level among two groups with p value set at 0.05.

RESULTS AND DISCUSSION Overview of a UHPLC-SWATH MS based pseudotargeted metabolomics method MRM MS is the gold standard for quantification analysis because of its wide linear dynamic range, good repeatability and high sensitivity13. Targeted metabolomics pays attention to a small number of metabolites. Pseudotargeted metabolomics method aimed to detect as many metabolites as possible in real samples based on UHPLC-MRM MS7,8. SWATH has the capability to gain fragmental information of all precursor ions15. In a recent paper on urinary SWATHtoMRM method 25, constant Q1 isolation window (25 Da) was used in SWATH MS and ion-pairs were selected only in one CE value. And variable Q1 isolation windows in SWATH showed obvious advantages to simplify the complexity of MS2 spectra when compared to constant Q1 isolation window22. It is helpful for improving the selectivity and reducing cross fragment ion interference. Until now, there is no method for pseudotargeted metabolomics study based on SWATH MS technology with variable Q1 isolation windows to select ion-pair with a suitable CE from bio-samples. And the method will also have characteristic of wide coverage and being easy to be achieved. The workflow scheme establishing a pseudotargeted ion-pair selecting method based on UHPLC-Q-TOF MS with SWATH acquisition is depicted in Figure 1. First of all, a representative biological sample was used to acquire the retention time, MS1 and MS2 information of metabolites by using SWATH MS based non-targeted method in

ACS Paragon Plus Environment

Page 11 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry 10

different collision energy (CE). The ion-pairs in different CEs were selected based on a software tool, named “MRM Ion-pairs Calculator Tool” running in Matlab environment (Version 7.14.0.739, R2012a, 64-bit). Then, the selected ion-pairs were validated to remove the false positive by a UHPLC coupled with QQQ mass spectrometer system in scheduled MRM method. Moreover, the remaining ion-pairs in different CEs were further integrated to select optimal CE. Based on the optimal ion-pair list, the pseudotargeted method can be established. The repeatability of the method was evaluated by intraday and interday precision. For diabetes study, the same steps were applied to acquire ion-pairs based on serum QC sample and the pseudotargeted method with these ion-pairs was used to investigate the differences between Type 2 diabetes and healthy controls.

Figure 1 | Overview of pseudotargeted method based on UHPLC-SWATH MS.

Development

of

a

pseudotargeted

metabolomics

method

based

on

UHPLC-Q-TOF MS with SWATH acquisition. The process to establish pseudotargeted metabolomics method contains four key steps,

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 25 11

Step 1: Metabolite data acquisition by UHPLC-Q-TOF MS with SWATH mode A non-targeted method based on UHPLC-Q-TOF MS with SWATH acquisition was used to gain as much information of MS/MS ions as possible in raw data collection of a SRM 1950 plasma sample. Non-targeted data contained retention time, MS1 and MS/MS information in CE=15 V, CE=30 V, CE=45 V, as shown in Figure 2. MS1 data were extracted by MarkerViewTM. And MS/MS data were extracted by MS-DIAL for deconvolution of SWATH MS data. A total of 4288 MS1 ions were extracted and about 6000 deconvolution MS/MS spectra were obtained in each CE. And these data were further used to ion-pair selection.

Figure 2 | (A) Steps to acquire possible ion-pair with different CEs; (B) validation of ion-pair and selection of MRM ion-pair with the optimal CE.

Step 2: Ion-pair selection in each CE The data of MS1 and MS/MS were imported into in-house software tool “MRM Ion-pairs Calculator Tool” used to select respective characteristic ion-pair in different CEs (15 V, 30 V and 45 V). As described in Figure 2A, the tool mainly included four sub-steps. 1) Noise of MS2 spectra was filtered to reduce noise signal interference and

ACS Paragon Plus Environment

Page 13 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry 12

to improve accuracy of selected ion pairs. The noise level was set as 200 by statistical analysis of signal distributions; 2) in “acquire links between precursor ions and product ions” step, the precursor ion and its product ions in each CE were matched by the retention time difference and m/z difference. Because the -CH2- group is the minimum m/z loss, the product ions were retained only with the mass difference between MS1 and MS2 greater than 13.9 Da. In this study, the retention time and m/z differences were respectively set as 0.2 min and 0.01 Da; 3) a metabolite may include a number of different ions, such as adduct ions, isotopic ions, neutral loss ions, etc. So a step of ion fusion was added to obtain “one feature for one peak” data. The precursor ions from different adduct ions (e.g., M+Na+, M+NH4+, M+K+), isotopic ions, neutral loss (e.g., H2O, NH3) and oligomers (e.g., 2M+H+, 3M+H+) would be deleted; 4) the product ion with the highest intensity in each CE was chosen as a characteristic product ion of the precursor ion. If no characteristic product ion was found, the precursor ion was set as the same mass as its product ion and the ion-pair was regarded as “sham” ion-pair. 4288 original MS1 ions and approximately 18000 deconvolution MS/MS spectra were obtained from the SRM 1950 plasma sample. After 4 sub-steps, finally, three ion-pair lists of 1668 unique metabolites with different CEs were saved.

Step 3: Ion-pair validation by scheduled MRM The selected ion-pairs in each CE were validated to filtrate false positive by UHPLC coupled with QQQ MS in scheduled MRM mode. The MS-DIAL is a novel tool to solve the precursor – fragment ion problem in untargeted SWATH acquisition4. The chromatographic behaviors of precursor ions and fragment ions are used to do the MS2Dec deconvolution. The method is similar to the DIA - Umpire2, Group – DIA3, which can be well used to determine possible precursor – and fragment – ion signals

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 25 13

in proteomics. The MS-DIAL has been applied to many studies based on SWATH acquisition and proved to gain good results5-7. Considering the situation of co-elution is pervasive and complicated, it is difficult to ensure no mistake exists during the MS2Dec deconvolution by MS-DIAL. So a step of ion-pair validation was added in this experiment. The details of the process are depicted in Figure 2B. From the extracted ion chromatography, ion-pairs which were not well detected were considered as false positive ions and were to be deleted from the lists of different CEs.

Step 4: Integration of final ion-pairs for pseudotargeted method The remaining ion-pairs are to be further integrated by an in-house coded algorithms to select optimal ion-pairs and CE. The ion-pair integration tool contained two main criteria: 1) Compare the intensities of characteristic product ions of the same metabolite in each CE and select the most intensive characteristic ion-pair as optimal characteristic ion-pair. 2) Reserve the characteristic ion-pair preferentially and delete the “sham” ion-pair, if “sham” ion-pair and characteristic ion-pair of the same metabolites simultaneously existed.

Figure 3 | Pseudotargeted ion-pair chromatogram of UHPLC-MRM analysis at positive ion mode.

ACS Paragon Plus Environment

Page 15 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry 14

Finally, 1373 ion-pairs of plasma metabolites were included in positive ion mode and they can well be detected in our pseudotargeted method. The extracted ion chromatogram of the SRM 1950 plasma sample is displayed in Figure 3. 84.9% of 1373 ion-pairs had characteristic product ions, and 207 of them had “sham” ion-pair (supplementary Table S3). The possible reasons of “sham” ion-pair perhaps are that some metabolites are not easy to be fragmented even when CE = 45 V was used. Because of the different peak detection algorithms used in MarkerViewTM software and MS-DIAL software, recognized MS1 peak lists perhaps were different. As a result, some MS1 peaks from MarkerViewTM could not well match their deconvolution MS/MS data. Moreover, although we collected the MS/MS spectra, the intensities of product ions from some low-intensity metabolites were very close to the noise level to be removed.

Advantages of the established pseudotargeted method based on SWATH Similarly, a pseudotargeted method based on the information from UHPLC-Q-TOF with IDA mode was also established. The same plasma sample, pretreatment steps and separation condition were used. And the ion-pair distribution of detected metabolites based on IDA and SWATH is exhibited in the space of precursor ion m/z value versus retention time (Figure 4A). Only 913 ion-pairs had characteristic product ions based on IDA mode, while 1166 ion-pairs existed based on SWATH mode. It means that additional 253 metabolites could have characteristic ion-pairs if UHPLC-Q-TOF with SWATH acquisition was used to construct pseudotargeted metabolomics method. The increased characteristic ion-pairs benefited by the ability of rich fragmental information acquisition by SWATH, especially for low abundant

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 25 15

metabolites. And the result also revealed SWATH could obviously increase the number of characteristic ion-pairs, especially at the time of co-elution occurrence. The advantages of using characteristic ion-pairs in the MRM mode is very obvious, especially for quantitative analysis of metabolites. An example is shown in supplementary Figure S1, characteristic product ion of the metabolite with precursor ion m/z 630.36 was not acquired in IDA mode. The sham ion-pair was used to detect

Figure 4 | (A) Ion-pair distribution of detected metabolites based on IDA (left) and SWATH (right) modes; (B) Disturbance of deconvolution in co-eluted peaks at single CE and its elimination by different CEs.

ACS Paragon Plus Environment

Page 17 of 25 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry 16

this metabolite, but another metabolite with precursor ion m/z 630.74 had a similar retention time. Though these two metabolites could be partly separated, it was still difficult to identify which peak belonged to m/z 630.36 and the peak area could not be well determined. On the other hand, based on SWATH MS it would be easily solved because these two metabolites had their own characteristic ion-pairs. 630.36/147.05 (the blue peak) represented the metabolite with m/z 630.36 and 630.74/166.09 (the red peak) represented the metabolite with m/z 630.74. So the interference of co-eluted metabolites with a similar precursor ion m/z can be well solved by using special product ion. And the accuracy of quantitative results could be obviously improved. Additionally, the use of different CEs in SWATH was an efficient way not only to optimize CE for each ion-pair but also to make up the difficulties in deconvolution for co-eluted metabolites. As shown in Figure 4B, the compound with m/z 483.3382 (the black) and the compound with m/z 488.2958 (the blue) were eluted at 9.86 min. And two metabolites were fragmented in the same SWATH window (477.3-490.3). What’s more, their peak shapes were also similar. It is very difficult to automatically distinguish which was the real product ion of each metabolite by deconvolution. However, different metabolites had different physico-chemical properties and compositions. When the different CEs were used to fragment the co-eluted metabolites, the product ions in MS2 spectra would be changed. Although the ion-pairs of m/z 483.3382 and 488.2958 were respectively regarded as 483.34/466.32, 488.30/466.32 at CE 15 V and 483.34/413.28, 488.30/413.28 at CE 45 V, only 483.34/466.32 at CE 15 V and 488.30/413.28 at CE 45 V could be detected in MRM mode. The m/z 483.34 could be fragmented to m/z 466.32 at CE 15 V, but m/z 413.28 could not be produced from it at CE 45V. Similarly, m/z 488.30 could be fragmented

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 25 17

to m/z 413.28 at CE 45V, but it could not be fragmented to m/z 466.32. It means the use of different CEs in SWATH mode could be a useful way to acquire the right ion-pair for each metabolite, especially in the situation of co-elution.

Repeatability of the established pseudotargeted method The intraday and interday precisions were evaluated according to the calculated coefficient of variation (CV). For the measurement of intraday precision and stability, ten LC injection replicates were performed in one day. The intraday RSD distribution revealed a total of 82.4% metabolites with a RSD less than 15%, which accounted for more than 97.7% of total peak areas (Figure S2A). For the measurement of interday precision and stability, the experiment with seven replicates in continuous three days was carried out. And the results showed that 25.1% metabolites, accounting for only 8.8% of total peak areas, had RSD values larger than 15% (Figure S2B). These data demonstrated this established method was stable and reliable and is suitable for metabolomics study.

Application of established pseudotargeted method in metabolomics study on Type 2 diabetes According to the established pseudotargeted method, a Type 2 diabetes metabolomics study was performed. The serum samples were collected from 30 Type 2 diabetes mellitus patients (male/female: 15/15) and 30 healthy controls (male/female: 15/15). The basic clinical information is listed in supplementary Table S4. The ages and the concentrations of total cholesterol, triacylglycerols, low-density lipoprotein showed no significant differences between diabetes group and control group. However, higher level of fasting glucose (p < 0.001) and lower high-density lipoprotein level (p
1 were chosen for further univariate statistical analysis through the SPSS for investigating significantly changed compounds between patients and healthy controls. In the end, 162 metabolites showed a significant alteration between two groups (p