Multi-level integration method based on traditional Chinese medicine active compound structure-disease target-signal pathway

Abstract

The invention discloses a method based on traditional Chinese medicine active compound structure-disease target-signal pathway multilevel integration, and relates to the field of traditional Chinese medicine component target prediction, and the method comprises the following steps: S1, obtaining a structural formula of a traditional Chinese medicine active compound; s2, drawing the structural formula retrieved in the step S1, storing the structural formula in an sdf format, and converting the sdf format into an smiles format; s3, importing the structural formula into a target database to predict potential targets of the traditional Chinese medicine active compounds; s4, aiming at a target disease, finding a hub gene related to the disease, namely a target spot of the target disease; s5, performing common gene analysis on the potential target spot of the traditional Chinese medicine active compound obtained in S3 and the target disease target spot obtained in S4 by using network pharmacology to obtain a common target spot; s6, finding out a chemical structure of the traditional Chinese medicine active compound and a common target binding site through molecular docking, and screening out the most key hub gene; and S7, carrying out enrichment analysis on the most critical hub gene, and predicting a bletilla striata polysaccharide structure-disease target-signal pathway.

Description

Technical Field

[0001] The present invention relates to the field of target prediction of traditional Chinese medicine components, and specifically to a method based on multi-level integration of traditional Chinese medicine active compound structure-disease target-signal pathway. Background Art

[0002] Network pharmacology is an emerging discipline that combines the ideas of systems biology and multi-directional pharmacology. Its emergence provides new methods and ideas for drug development and research on traditional Chinese medicine. Through the analysis of network pharmacology, we can better understand the complex relationship between drugs, targets and diseases, which helps to improve the efficacy of drugs and reduce adverse reactions. With the continuous development of artificial intelligence and big data technology, the application prospects of network pharmacology in the field of drug research will be broader. Research methods in this field include network-based molecular docking, analysis of drug-target-disease complex networks, etc. However, the existing network pharmacology database cannot directly predict molecules with clear chemical structures for active compounds of traditional Chinese medicine, and thus cannot meet the prediction of targets of active ingredients of traditional Chinese medicine.

[0003] The target database is the existing TCM active ingredient compound target prediction website including TCMSP, TCMID, HIT, TCM@Taiwan, NPACT, CancerHSP, NPASS, DrugBank, TTD, STITCH, ChEMBL, BindingDB, ZINC, SuperPred, SwissTargetPrediction, TargetNet, PharmMapper, ETCMv2.0, BATMAN-TCM, cMap, TCMIO, herb.ac.cn (HERB), ITCM, TMNP, DGIdb / CMap, SEA Search Server. Summary of the invention

[0004] Purpose of the invention: In view of the problem that the existing database cannot meet the prediction of targets of active ingredients of traditional Chinese medicine, the present invention provides a method based on multi-level integration of structure of active compounds of traditional Chinese medicine-disease target-signal pathway. The present invention is based on disease expression spectrum data and structural formula of active compounds of traditional Chinese medicine in literature, and quickly screens key data of multi-level integration of active compounds of traditional Chinese medicine-disease target-signal pathway.

[0005] Technical solution: A method based on multi-level integration of traditional Chinese medicine active compound structure-disease target-signaling pathway, characterized by comprising the following steps:

[0006] S1. Retrieve relevant literature on active compounds of traditional Chinese medicines through literature search to obtain the structural formula of active compounds of traditional Chinese medicines;

[0007] S2. Use chemical formula drawing software to draw the structural formula retrieved in step S1, save it in sdf format, and convert the sdf format into smiles format, and save it for subsequent analysis;

[0008] S3, importing structural formula into target database to predict potential targets of active compounds of traditional Chinese medicine;

[0009] S4. For target diseases, through co-expression WGCNA analysis, key genes are identified and significantly differentially expressed, to find disease-related hub genes, i.e., target disease targets;

[0010] S5. Use network pharmacology to perform common gene analysis on the potential targets of the active compounds of traditional Chinese medicine obtained in S3 and the target disease targets obtained in S4 to obtain common targets;

[0011] S6. Find the chemical structure and common target binding sites of active compounds of traditional Chinese medicine through molecular docking, and screen out the most critical hub genes;

[0012] S7. Perform enrichment analysis on the most critical hub genes and predict the Bletilla striata polysaccharide structure-disease target-signaling pathway.

[0013] Furthermore, in S1, the literature retrieval websites used include: Google Scholar, PubMed, Bing Scholar and Baidu Scholar.

[0014] Furthermore, in S2, the chemical formula drawing software used is ChemDraw22; the R package ChemmineR / ChemmineOB is used to convert the sdf format into the smiles format.

[0015] Furthermore, in S5, the network pharmacology method used is upset graph.

[0016] Furthermore, in S6, AutoDOCK Vina was used for molecular docking, and the most critical hub gene was determined through molecular docking scoring.

[0017] Furthermore, in S7, KEGG, Reactome, wikipathway, and GO databases were used to perform enrichment analysis on the most critical hub genes.

[0018] Beneficial effects: The present invention provides a method based on multi-level integration of the structure of active compounds of traditional Chinese medicines - disease targets - signal pathways. Currently, existing databases and methods cannot solve the molecular target problem of active compounds of traditional Chinese medicines. The present invention can quickly analyze the structure of active compounds of traditional Chinese medicines, combine the prediction of disease targets with bioinformatics, and complete the binding relationship between active compounds of traditional Chinese medicines and diseases through molecular docking and network pharmacology analysis, and screen downstream signal pathways, so as to quickly complete the screening of disease targets related to active compounds of traditional Chinese medicines for subsequent research. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 The structural formula of Bletilla striata polysaccharide drawn by chemdraw drawing software;

[0020] Figure 2 This is a graphical representation of the results of common gene analysis of potential targets of Bletilla striata polysaccharide and target disease targets using an upset plot. DETAILED DESCRIPTION

[0021] The present invention is described in detail below in conjunction with specific implementation modes.

[0022] This example takes Bletilla striata polysaccharide as an example. Through literature retrieval and bibliometric statistics, combined with polysaccharide extraction and purification experiments, and NMR structural analysis, there are already 8 Bletilla striata polysaccharide structures with clear structural formulas, which are linear sugar chains composed of glucose and mannose. However, these Bletilla striata polysaccharide structure articles do not provide structural formulas in SDF format. It is necessary to use ChemDraw software to draw the analyzed structures as shown below: Figure 1 The structure shown is saved as sdf format for further analysis of Bletilla striata polysaccharide targets through existing databases. Disease targets can be analyzed by WGCNA based on public data expression profile data downloaded from GSA, GEO, TCGA, etc., and key genes can be extracted for differential analysis. The obtained Bletilla striata polysaccharide targets and disease targets are intersected by genes for signal pathway analysis. The specific method is as follows:

[0023] S1. Literature search for all literature related to Bletilla striata polysaccharide;

[0024] The literature search websites or databases used were: Google Scholar, PubMed, Bing Scholar, Baidu Scholar;

[0025] Keywords: "bletilla striata polysaccharide" & "NMR" & "structure", date: as of August 29, 2023; the Chinese meanings of bletilla striata polysaccharide, NMR, and structure are: Bletilla striata polysaccharide, nuclear magnetic resonance, and structure respectively;

[0026] Through literature search, we know that there are 8 Bletilla polysaccharides with structural formulas, named BSP_12, BSP_23, BSP_58A, BSP_58B, BSP_83, BSP_91, BSP_323, and BSP_327 according to the size of the molecular formula. In addition, the article PMID 34044956 (PMID34044956 is the literature number in the fields of life sciences and medicine included in the PubMed search engine) proposed 4 Bletilla polysaccharide structures (BSPs-H, BSPs-A, BSPs-B, and BSPs-U), but there is no clear surface specific structural molecular formula, so it is eliminated. Therefore, there are 8 valid structures.

[0027] S2. Use ChemDraw22 chemical formula drawing software to draw the eight structural formulas of Bletilla striata polysaccharides retrieved in step S1, save them in sdf format, use R package ChemmineR / ChemmineOB to convert sdf into smiles format, and save them for subsequent analysis.

[0028] ChemmineR and ChemmineOB are chemical informatics toolkits used in the R language environment to process and analyze chemical data. ChemmineR is an R package for chemical informatics and computational chemistry, providing a wealth of functions, including molecular descriptor calculation, molecular similarity calculation, cluster analysis, chemical space visualization, etc.

[0029] Using the R package ChemmineR / ChemmineOB, the specific operations to convert sdf to smiles format are as follows: Load the ChemmineR package:

[0030] rm(list=ls())

[0031] library("ChemmineR")#Loads the packagesdfset<-read.SDFset("BSP_58B.sdf")

[0032] plot(sdfset,print=FALSE)#Plots structures to R graphics device###Export graph

[0033] ##Load R package

[0034] library(export)

[0035] ##Export to PPT editable format

[0036] graph2ppt(file="BSP3sdfdraw.pptx")

[0037] graph2pdf(file="BSP3sdfdraw.pdf")

[0038] #Create PDF file and draw graphics

[0039] pdf("BSP_58Bsdfdraw.pdf")

[0040] plot(sdfset,print=FALSE)

[0041] dev.off()

[0042] #Create a PPT file and draw graphics

[0043] #library("export")#Load the export package

[0044] ###convert to smiles

[0045] #BiocManager::install("ChemmineOB")

[0046] library(ChemmineOB)

[0047] (smiles<-sdf2smiles(sdfset))

[0048] as.character(smiles)

[0049] write.SMI(smiles,file="BSP_58B.smi",cid=TRUE)

[0050] S3, importing the structural formula of Bletilla striata polysaccharide into the target database to predict the potential targets of Bletilla striata polysaccharide;

[0051] In this embodiment, the target database is an existing Chinese medicine active ingredient compound target prediction website (TCMSP, TCMID, HIT, TCM@Taiwan, NPACT, CancerHSP, NPASS, DrugBank, TTD, STITCH, ChEMBL, BindingDB, ZINC, SuperPred, SwissTargetPrediction, TargetNet, PharmMapper, ETCMv2.0, BATMAN-TCM, cMap, TCMIO, herb.ac.cn (HERB), ITCM, TMNP, DGIdb / CMap, SEASearch Server). Through the method of network pharmacology, the existing Chinese medicine active ingredient compound target prediction website is integrated to perform Bletilla striata polysaccharide target prediction.

[0052] S4. Find disease hub genes through co-expression WGCNA analysis and significant differential expression analysis.

[0053] In this example, we mainly focus on asthma. Based on the existing sequencing data GSE40889 in the GEO database, we identify key genes through co-expression WGCNA analysis and perform significant differential expression analysis to find asthma-related hub genes, i.e., target disease targets.

[0054] S5. Use network pharmacology to perform common gene analysis on the potential targets of Bletilla striata polysaccharide obtained in S3 and the target disease targets obtained in S4 to obtain common targets.

[0055] The results of the common gene analysis of the potential targets of Bletilla striata polysaccharide obtained in S3 and the target disease targets obtained in S4 using the upset graph are as follows: 11 common genes in the PharmMapper database, 1 in the Swiss database, 5 in the targetnet database, and 5 in the superpred database (see Appendix). Figure 2 ).

[0056] S6. Through molecular docking, the chemical structure and common target binding sites of Bletilla striata polysaccharides were found, and the most critical hub genes were screened out.

[0057] Molecular docking uses AutoDOCK Vina, which improves the average accuracy of binding mode prediction, speeds up the search by using a simpler scoring function, and still provides reproducible docking results when dealing with systems with about 20 rotatable bonds. We consider the molecular docking with higher scores to be the most critical hub genes, and usually select the top 10 results as the most critical hub genes.

[0058] S7. Use KEGG, Reactome, wikipathway, and GO databases to perform enrichment analysis on the most critical hub genes and predict the Bletilla striata polysaccharide structure-disease target (most critical hub gene)-signaling pathway. This will explore the key information of the multi-level integration of Bletilla striata polysaccharide structure-disease target-signaling pathway.

[0059] The method of the present invention can solve the problem of predicting the disease target of Bletilla striata polysaccharide, thereby performing downstream signal pathway analysis and exploring the target of Bletilla striata polysaccharide.

[0060] Specific operations for enrichment analysis of the most critical hub genes using KEGG, Reactome, wikipathway, and GO databases:

[0061] library("org.Hs.eg.db")

[0062] rt=read.table("gene.txt",sep="\t",check.names=F,header=F)

[0063] genes = as.vector(rt[,1])

[0064] entrezIDs<-mget(genes,org.Hs.egSYMBOL2EG,ifnotfound=NA)

[0065] entrezIDs<-as.character(entrezIDs)

[0066] out=cbind(rt,entrezID=entrezIDs)

[0067] colnames(out)=c("symbol","entrezID")

[0068] write.table(out,file="id.txt",sep="\t",quote=F,row.names=F)

[0069] library("clusterProfiler")

[0070] library("org.Hs.eg.db")

[0071] library("enrichplot")

[0072] library("ggplot2")

[0073] library(gridExtra)

[0074] library(ggsci)

[0075] rt=read.table("id.txt",sep="\t",header=T,check.names=F)

[0076] rt=rt[is.na(rt[,"entrezID"])==F,]

[0077] gene=rt$entrezID

[0078] kk<-enrichGO(gene=gene,

[0079] OrgDb=org.Hs.eg.db,

[0080] pvalueCutoff=0.05,

[0081] qvalueCutoff=0.05,

[0082] ont="all",

[0083] readable=T)

[0084] write.table(kk,file="GO.txt",sep="\t",quote=F,row.names=F)

[0085] #pdf(file="barplot.pdf",width=10,height=8)

[0086] p1=barplot(kk,drop=TRUE,showCategory=10,split="ONTOLOGY")+facet_grid(ONTOLOGY~.,scale='free')

[0087] p1_npg=p1+scale_color_npg()

[0088] p2=dotplot(kk,showCategory=10,split="ONTOLOGY")+facet_grid(ONTOLOGY~.,scale='free')

[0089] p2_npg=p2+scale_fill_npg()

[0090] library(patchwork)

[0091] p<-p1+p2

[0092] print(p2)

[0093] grid.arrange(p1_npg,p2_npg,ncol=2)

[0094] dev.off()

[0095] pdf(file="bubble.pdf", width=10, height=8)

[0096] dotplot(kk,showCategory=10,split="ONTOLOGY")+facet_grid(ONTOLOGY~.,scale='free')

[0097] dev.off()

[0098] #Installation of reactomePA error solution

[0099] #First download and install these two packages

[0100] #https: / / www.bioconductor.org / help / search / index.html? q=reactome.db /

[0101] #https: / / www.bioconductor.org / packages / release / bioc / html / graphite.html

[0102] #Download and install reactomePA

[0103] #https: / / www.bioconductor.org / packages / release / bioc / html / ReactomePA.htmllibrary(clusterProfiler)

[0104] library(org.Hs.eg.db)

[0105] library(enrichplot)

[0106] #require(DOSE)

[0107] #options(clusterProfiles.download.method"wininet")

[0108] # library ( KEGG . db )

[0109] rt0read.table("id.txt",sep0"\t",header0T,check.namesF)

[0110] rt_rt[is.na(rt[,"entrezID"])sF,]

[0111] gene / rt$entrezID

[0112] #kegg

[0113] kk<-enrichKEGG(genes;genes,

[0114] organism'hsa',

[0115] pvalueCutoffS0.05,

[0116] qvalueCutoff1)

[0117] write.table(kk,file"KEGG.txt",sep"\t",quote(F,row.namesF)##########################################

[0118] ###My country

[0119] ##speechR snow

[0120] library(export)

[0121] ##Instead of PPT checklist

[0122] barplot ( kk , drop TRUE , showCategory 30 )

[0123] #graph2ppt(file"barplotKEGG.pptx")

[0124] graph2pdf(file"barplotKEGG.pdf")

[0125] dev.off()

[0126] dotplot(kk,showCategory=30)

[0127] #graph2ppt(file="dotKEGG.pptx")

[0128] graph2pdf(file="dotKEGG.pdf")

[0129] dev.off()

[0130] head(kk,3)

[0131] y<-setReadable(kk,OrgDb=org.Hs.eg.db,keyType="ENTREZID")##ThegeneID column is translated to symbol

[0132] head(y,3)

[0133] write.table(y,file="KEGGsymbol.txt",sep="\t",quote=F,row.names=F)####wikipathwayhead(gene)

[0134] ##wikipathway

[0135] #if(!"rWikiPathways"%in%installed.packages()){

[0136] #if(!requireNamespace("BiocManager",quietly=TRUE))

[0137] #install.packages("BiocManager")

[0138] #BiocManager::install("rWikiPathways",update=FALSE)

[0139] #}

[0140] library(rWikiPathways)

[0141] library(tidyr)

[0142] library(magrittr)

[0143] library(clusterProfiler)

[0144] head(gene)

[0145] #supported organisms(and case-dependent spelling)...

[0146] listOrganisms()

[0147] #Due to network reasons, it is not possible to download online, so download and read the data in advance

[0148] #Specify the downloaded GMT file path

[0149] #https: / / wikipathways-data.wmcloud.org / current / gmt /

[0150] wpgmtfile<-"wikipathways-20230810-gmt-Homo_sapiens.gmt"

[0151] #Method 2: Download online

[0152] #wpgmtfile<-downloadPathwayArchive(organism="Homo sapiens",format="gmt")wp2gene<-clusterProfiler::read.gmt(wpgmtfile)

[0153] wp2gene<-wp2gene%>% tidyr::separate(term,c("name","version","wpid","org"),"%")wpid2gene<-wp2gene%>%dplyr::select(wpid,gene)#TERM2GENE

[0154] wpid2name<-wp2gene%>%dplyr::select(wpid,name)#TERM2NAME

[0155] wpid2gene

[0156] wpid2nameewp<-enricher(gene,TERM2GENE=wpid2gene,TERM2NAME=wpid2name)head(ewp)

[0157] write.table(ewp,file="wikipathway.txt",sep="\t",quote=F,row.names=F)

[0158] head(ewp,3)

[0159] y1<-setReadable(ewp,OrgDb=org.Hs.eg.db,keyType="ENTREZID")

[0160] ##The geneID column is translated to symbol

[0161] head(y1,3)

[0162] write.table(y1,file="wikipathwaysymbol.txt",sep="\t",quote=F,row.names=F)

[0163] ###reactomePA

[0164] library(ReactomePA)

[0165] head(gene)

[0166] x<-enrichPathway(gene=gene,pvalueCutoff=0.05,readable=T)

[0167] head(as.data.frame(x))

[0168] write.table(x,file="reactomesymbol.txt",sep="\t",quote=F,row.names=F)

[0169] As described above, although the present invention has been shown and described with reference to specific preferred embodiments, it should not be construed as limiting the present invention itself. Various changes in form and details may be made without departing from the spirit and scope of the present invention as defined in the appended claims.

Claims

1. A method based on the multi-level integration of the structure of active compounds of traditional Chinese medicine, disease targets and signaling pathways. It is characterized in that The steps include: S1. Retrieve relevant literature on active compounds of traditional Chinese medicines through literature search to obtain the structural formula of active compounds of traditional Chinese medicines; S2. Use chemical formula drawing software to draw the structural formula retrieved in step S1, save it in sdf format, and convert the sdf format into smiles format, and save it for subsequent analysis; S3, importing structural formula into target database to predict potential targets of active compounds of traditional Chinese medicine; S4. For target diseases, through co-expression WGCNA analysis, key genes are identified and significantly differentially expressed, to find disease-related hub genes, i.e., target disease targets; S5. Use network pharmacology to perform common gene analysis on the potential targets of the active compounds of traditional Chinese medicine obtained in S3 and the target disease targets obtained in S4 to obtain common targets; S6. Find the chemical structure and common target binding sites of active compounds of traditional Chinese medicine through molecular docking, and screen out the most critical hub genes; S7. Perform enrichment analysis on the most critical hub genes and predict the Bletilla striata polysaccharide structure-disease target-signaling pathway.

2. A method based on multi-level integration of traditional Chinese medicine active compound structure-disease target-signaling pathway according to claim 1, It is characterized in that In S1, the literature retrieval websites used include: Google Scholar, PubMed, Bing Scholar and Baidu Scholar.

3. A method based on multi-level integration of traditional Chinese medicine active compound structure-disease target-signaling pathway according to claim 1, It is characterized in that In S2, the chemical formula drawing software used is ChemDraw22; the R package ChemmineR / ChemmineOB is used to convert the sdf format into the smiles format.

4. A method based on multi-level integration of traditional Chinese medicine active compound structure-disease target-signaling pathway according to claim 1, It is characterized in that In S5, the network pharmacology method used is upset graph.

5. A method based on multi-level integration of traditional Chinese medicine active compound structure-disease target-signaling pathway according to claim 1, It is characterized in that In S6, AutoDOCK Vina was used for molecular docking, and the most critical hub gene was determined by molecular docking scoring.

6. A method based on multi-level integration of traditional Chinese medicine active compound structure-disease target-signaling pathway according to claim 1, It is characterized in that In S7, the most critical hub genes were enriched and analyzed using the KEGG, Reactome, wikipathway, and GO databases.

Images