library(meshes)
data(geneList, package="DOSE")
de <- names(geneList)[1:100]
x <- enrichMeSH(de, MeSHDb = db, database='gendoo', category = 'C')14 MeSH enrichment analysis
The meshes package (Yu 2018) supports enrichment analysis (over-representation analysis and gene set enrichment analysis) of gene list or whole expression profile using MeSH annotation. Data source from gendoo, gene2pubmed and RBBH are all supported. User can select interesting category to test. All 16 categories of MeSH are supported. The analysis supports about 200 species (see also Chapter 23 for more details).
14.1 Chapter overview
| Aspect | MeSH enrichment |
|---|---|
| Questions | Which MeSH concepts/categories associate with a gene set or are enriched across a ranked profile? |
| Input | A gene vector or ranked vector, species-specific MeSHDb, and a source such as gendoo, gene2pubmed, or RBBH. |
| Methods | ORA and GSEA with selectable MeSH category and data source. |
| Main functions | enrichMeSH(), gseMeSH(); semantic comparisons use the meshes similarity functions. |
| Output | enrichResult / gseaResult objects for downstream plotting and comparison. |
| Main limitations | Results depend on database preparation, source, species coverage, and ID mapping; the same concepts may differ across sources. |
14.2 MeSH over-representation analysis
First, we need to load/fetch species-specific MeSH annotation database, please refer to Chapter 23.
In this example, we use data source from gendoo and C (Diseases) category.
head(x) ID Description GeneRatio BgRatio RichFactor
D000705 D000705 Anaphase 26/100 232/26076 0.11206897
D019926 D019926 Cyclin B 30/100 428/26076 0.07009346
D018386 D018386 Kinetochores 24/100 224/26076 0.10714286
D020090 D020090 Chromosome Segregation 26/100 362/26076 0.07182320
D011123 D011123 Polyploidy 16/100 169/26076 0.09467456
D016547 D016547 Kinesins 22/100 476/26076 0.04621849
FoldEnrichment oddsRatio zScore pvalue p.adjust
D000705 29.22310 43.95303 26.79137 4.495523e-31 2.036472e-27
D019926 18.27757 27.54271 22.36165 1.061989e-29 2.405404e-26
D018386 27.93857 40.69895 25.12335 3.337217e-28 5.039198e-25
D020090 18.72862 26.81145 21.07514 5.663108e-26 6.413470e-23
D011123 24.68734 32.14815 19.16801 3.951760e-18 3.580295e-15
D016547 12.05193 15.85576 15.09898 6.849607e-18 5.171453e-15
qvalue
D000705 1.360487e-27
D019926 1.606957e-26
D018386 3.366491e-25
D020090 4.284588e-23
D011123 2.391855e-15
D016547 3.454845e-15
geneID
D000705 4085/991/54821/3832/6790/22974/983/9133/332/4751/81930/55165/10232/11065/2305/1111/6241/9055/220134/9493/7153/9212/7272/10112/1062/10403
D019926 4085/991/3832/6790/5080/22974/983/83461/9133/332/10232/11065/55872/8208/2305/3868/1111/23397/9055/9493/7153/890/4605/64151/7272/10112/1062/9787/81620/10403
D018386 4085/991/55839/54821/3832/79019/6790/22974/983/332/4751/81930/11065/55143/1111/9055/220134/7153/9212/7272/1062/9787/10460/10403
D020090 4085/991/55839/51203/54821/79019/6790/22974/983/9133/332/4751/81930/11065/55143/2305/1111/9493/7153/64151/9212/7272/1062/9787/10460/10403
D011123 4085/991/6790/9133/332/10232/55143/2305/1111/9493/7153/890/4605/9212/1062/81620
D016547 4085/3832/6790/22974/3833/983/332/81930/146909/55143/2305/1111/23397/9055/9493/7153/64151/9212/10112/1062/9787/10403
Count
D000705 26
D019926 30
D018386 24
D020090 26
D011123 16
D016547 2214.3 MeSH gene set enrichment analysis
In the following example, we use data source from gene2pubmed and test category G (Phenomena and Processes).
y <- gseMeSH(geneList, MeSHDb = db, database = 'gene2pubmed', category = "G")head(y) ID Description setSize enrichmentScore NES
D000705 D000705 Anaphase 208 0.6734042 3.001720
D018386 D018386 Kinetochores 194 0.6293151 2.839466
D051738 D051738 Origin Recognition Complex 65 0.7215645 2.687916
D019926 D019926 Cyclin B 399 0.5417028 2.610667
D049468 D049468 Prometaphase 26 0.8614331 2.608087
D011123 D011123 Polyploidy 160 0.5966375 2.595049
pvalue p.adjust qvalue rank leading_edge
D000705 1e-10 1.916604e-08 8.869359e-09 1071 tags=35%, list=9%, signal=33%
D018386 1e-10 1.916604e-08 8.869359e-09 759 tags=26%, list=6%, signal=25%
D051738 1e-10 1.916604e-08 8.869359e-09 1077 tags=45%, list=9%, signal=41%
D019926 1e-10 1.916604e-08 8.869359e-09 1499 tags=30%, list=12%, signal=27%
D049468 1e-10 1.916604e-08 8.869359e-09 370 tags=54%, list=3%, signal=52%
D011123 1e-10 1.916604e-08 8.869359e-09 1077 tags=31%, list=9%, signal=29%
core_enrichment
D000705 991/2305/9493/1062/9133/10403/7153/6241/55165/11065/220134/22974/4751/983/54821/10232/4085/81930/332/3832/7272/9212/1111/9055/10112/6790/891/24137/9232/1164/11004/990/5347/29127/701/11130/57405/1894/9700/5888/56992/4998/10733/29899/699/4609/1063/5111/5688/5709/26271/55055/51053/641/5698/1719/3925/5693/8317/5713/3930/5721/5691/10051/5685/8568/4172/23481/5690/5684/5885/5686/5695
D018386 55143/991/1062/10403/7153/9787/11065/220134/22974/10460/4751/79019/55839/983/54821/4085/81930/332/3832/7272/9212/1111/9055/6790/891/11004/5347/29127/701/11130/79682/57405/10615/1894/2491/9700/5888/23594/54801/29899/1058/11135/699/6491/1063/55055/8317/4112/10036/79980/9735
D051738 8318/55388/890/81620/1111/4174/4171/990/5347/898/23594/4998/4175/4173/10926/6502/4609/5111/84823/51053/1869/1719/8317/5427/4176/10036/1019/4172/11200
D019926 991/2305/9493/1062/3868/4605/9133/10403/7153/23397/9787/11065/55872/83461/22974/890/983/10232/4085/5080/81620/332/3832/7272/64151/8208/1111/9055/10112/6790/891/24137/9232/4001/4171/1164/11004/993/990/5347/701/1894/9700/5888/7083/898/56992/4998/4288/10733/1163/9134/4173/6502/6772/994/9918/699/4609/3945/1063/5111/5688/84823/5709/26271/51053/1869/330/1029/5698/4904/4067/5693/5902/7032/430/5713/10036/9585/1302/5721/2810/5691/5499/578/9088/1116/995/10051/867/5685/1019/2597/8568/2023/5690/5684/5885/5686/5695/11200/10263/10213/10059/4839/3195/6873/7534/1476/5588/10592/63967/7164/999/1020/5699/899/5714/1994
D049468 55143/991/1062/10403/4085/1111/5347/29127/701/79682/1894/9700/699/4609
D011123 55143/991/2305/9493/1062/4605/9133/7153/890/10232/4085/81620/332/9212/1111/6790/891/9232/990/5347/701/9700/5888/898/23594/4998/6502/2537/6772/1058/699/4609/5111/10397/26271/51053/1869/1719/4830/3925/2821/9585/2810/2120/353/58/2956/1019/2539/11200
log2err
D000705 NaN
D018386 NaN
D051738 NaN
D019926 NaN
D049468 NaN
D011123 NaNUsers can use visualization methods implemented in enrichplot to visualize these enrichment results. With these visualization methods, see also Chapter 26, it’s much easier to interpret enriched results.
14.4 Next steps
- For database construction and similarity, see MeSH semantic similarity.
- For plotting, use the visualization map.
- For reusable resource packaging, see GSON.
References
Yu, Guangchuang. 2018. “Using Meshes for MeSH Term Enrichment and Semantic Analyses.” Bioinformatics 34 (21): 3766–67. https://doi.org/10.1093/bioinformatics/bty410.