## 1. Metabolome

The raw data from mass spectrometer was imported into commercial software Progenesis QI (version 2.2, hereinafter referred to as QI) for peak picking (https://www.nonlinear.com/progenesis/qi/), to obtain information of metabolites such as mass over charge, retention time and ion area. The QI workflow consists of the following steps: peak alignment, peak picking, and peak identification.

The metabolite identification was performed by Progenesis QI by searching against HMDB (v5.0), METLIN (v3.7.1) and KEGG (v96.0) databases. 

Pre-processing of peak data was performed using metaX (https://www.bioconductor.org/packages/3.2/bioc/html/metaX.html), the steps include: 

- Filtering out low quality ions (first removed ions in QC sample that contain over 50% missing value, then removed ions in actual samples that contain over 80% missing value)

- Using k-nearest neighbor (KNN) method for filling the missing values

- Using probabilistic quotient normalization (PQN) method for data normalization

- Using QC-RSC (Quality control-based robust LOESS signal correction) method to alleviate the effects of peak area attenuation

- Filtering out ions in all QC samples which are RSD > 30% (the ions with RSD > 30% are fluctuate greatly in the experiment and will not be included in downstream statistical analysis)

Taken the analysis of positive ion mode as example:

```R

library(metaX)

para <- new("metaXpara")

pfile <- "m_pos.csv" ## Output from QI, raw peak file with metabolite information

sfile <- "s_pos.list" ## Output from QI, sample list file

idres <- "i_pos.csv" ## Output from QI, ion intensity file

para@outdir <- "metaX_result_pos"

para@prefix <- "pos"

para@sampleListFile <- sfile

para@ratioPairs <- "COPD:Healthy"

para <- importDataFromQI(para, file=pfile)

plsdaPara <- new("plsDAPara")

plsdaPara@scale = "pareto"

plsdaPara@cpu = 4

plsdaPara@kfold = 3

#plsdaPara@do = FALSE

res <- doQCRLSC(para, cpu=1)

missValueImputeMethod(para)<-"KNN"

p <- metaXpipe(para, plsdaPara=plsdaPara, missValueRatioQC=0.5, missValueRatioSample=0.8, cvFilter=0.3, idres=idres, qcsc=0, scale="pareto", remveOutlier=FALSE, nor.method="pqn", t=1, nor.order = 1, pclean = FALSE, doROC=FALSE)

save(p, file="pos.rda")

sessionInfo()

```

The processed metabolome data are uploaded as metabolome.txt

The detailed information for each metabolite, including KEGG/HMDB/METLIN/PubChem/ChEBI IDs, SMILES structure, class and pathway is uploaded as compound_information.txt

## 2. Sputum and serum proteome

A panel of 280 proteins were measured using custom Quantibody Human Antibody Array (test procedure no. SOP-TF-QAH-001, SOP-TF-QAH-003 microarray) from RayBiotech (https://www.raybiotech.com/inflammation-protein-arrays/).

The processed sputum and serum proteome data are uploaded as sputum_proteome.txt and serum_proteome.txt

The detailed information of the 280 proteins is uploaded as protein_information.txt