Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 62 of the 62 references it cites were found in a registry or an index.

GO-PCA: An Unsupervised Method to Explore Gene Expression Data Using Prior Knowledge

Open access
Wagner, Florian · 2015 · PLOS ONE

Published version of a preprint

Abstract

METHOD: Genome-wide expression profiling is a widely used approach for characterizing heterogeneous populations of cells, tissues, biopsies, or other biological specimen. The exploratory analysis of such data typically relies on generic unsupervised methods, e.g. principal component analysis (PCA) or hierarchical clustering. However, generic methods fail to exploit prior knowledge about the molecular functions of genes. Here, I introduce GO-PCA, an unsupervised method that combines PCA with nonparametric GO enrichment analysis, in order to systematically search for sets of genes that are both strongly correlated and closely functionally related. These gene sets are then used to automatically generate expression signatures with functional labels, which collectively aim to provide a readily interpretable representation of biologically relevant similarities and differences. The robustness of the results obtained can be assessed by bootstrapping. RESULTS: I first applied GO-PCA to datasets containing diverse hematopoietic cell types from human and mouse, respectively. In both cases, GO-PCA generated a small number of signatures that represented the majority of lineages present, and whose labels reflected their respective biological characteristics. I then applied GO-PCA to human glioblastoma (GBM) data, and recovered signatures associated with four out of five previously defined GBM subtypes. My results demonstrate that GO-PCA is a powerful and versatile exploratory method that reduces an expression matrix containing thousands of genes to a much smaller set of interpretable signatures. In this way, GO-PCA aims to facilitate hypothesis generation, design of further analyses, and functional comparisons across datasets.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1371/journal.pone.0143196

Cites

62 of 62 references verified0 citation errors · 0 not checkable

Statistics

Figures

Fig 1 GO-PCA schematic. The figure provides a simplified illustration of the key idea behind the method using PC 1 of the DMAP dataset and the GO term “bicarbonate transport” (BT; GO:0015701) as an example. All genes annotated with the GO term are highlighted in purple. a After performing PCA on the gene expression matrix, genes are ranked according to their PC loadings. b The ranked list of genes is tested for GO enrichment using the XL-mHG test. In this example, the highest enrichment score (red arrow) is associated with the first five BT genes (HBA1, HBB, CA1, CA2, and RHAG; p = 8.3 × 10−8). These five genes are then used to generate a signature labeled with “bicarbonate transport” (red star in Fig 2). See Methods for details.
Fig 1 GO-PCA schematic. The figure provides a simplified illustration of the key idea behind the method using PC 1 of the DMAP dataset and the GO term “bicarbonate transport” (BT; GO:0015701) as an example. All genes annotated with the GO term are highlighted in purple. a After performing PCA on the gene expression matrix, genes are ranked according to their PC loadings. b The ranked list of genes is tested for GO enrichment using the XL-mHG test. In this example, the highest enrichment score (red arrow) is associated with the first five BT genes (HBA1, HBB, CA1, CA2, and RHAG; p = 8.3 × 10−8). These five genes are then used to generate a signature labeled with “bicarbonate transport” (red star in Fig 2). See Methods for details.CC BY · PubMed Central
Fig 2 Validation of GO-PCA using a dataset of 211 human transcriptomes representing diverse hematopoietic cell types (DMAP; [26]). Shown is a heat map of the signature matrix generated by GO-PCA, with signatures (rows) ordered using hierarchical clustering with correlation distance and average linkage, and samples (columns) ordered according to their known lineage identities. Arrows and boxes indicate specific associations between signatures and lineages discussed in the text. Signature labels (left) indicate the name of the GO term the signature is derived from. “BP”, “MF”, and “CC” refer to “biological process”, “molecular function”, and “cellular component”, the three main branches of the gene ontology. The three numbers in parentheses indicate, respectively, 1) the principal component that the GO term was found to be enriched in (with a negative sign indicating enrichment among the genes with the lowest loadings), 2) the number of genes in the signature, and 3) the total number of genes in the analysis that were annotated with the GO term. Signature expression levels are calculated as the unweighted average over the standardized expression levels of each gene in the signature (see Methods for details).
Fig 2 Validation of GO-PCA using a dataset of 211 human transcriptomes representing diverse hematopoietic cell types (DMAP; [26]). Shown is a heat map of the signature matrix generated by GO-PCA, with signatures (rows) ordered using hierarchical clustering with correlation distance and average linkage, and samples (columns) ordered according to their known lineage identities. Arrows and boxes indicate specific associations between signatures and lineages discussed in the text. Signature labels (left) indicate the name of the GO term the signature is derived from. “BP”, “MF”, and “CC” refer to “biological process”, “molecular function”, and “cellular component”, the three main branches of the gene ontology. The three numbers in parentheses indicate, respectively, 1) the principal component that the GO term was found to be enriched in (with a negative sign indicating enrichment among the genes with the lowest loadings), 2) the number of genes in the signature, and 3) the total number of genes in the analysis that were annotated with the GO term. Signature expression levels are calculated as the unweighted average over the standardized expression levels of each gene in the signature (see Methods for details).CC BY · PubMed Central
Fig 3 Analysis of the robustness of DMAP GO-PCA signatures using bootstrapping. a Detection rates of the 50 DMAP signatures generated by GO-PCA (see Fig 2). GO-PCA was applied to 50 bootstrap samples with the same size as the original dataset (n = 211). In each analysis, a signature from the original analysis was counted as detected if there existed a signature based on the exact same GO term (yellow bars), or based on either the same GO term or any “related” GO term (blue bars). Related GO terms were defined as all ancestor and descendent terms in the Gene Ontology. This figure panel was generated using the GO-PCA script gopca_plot_bootstrap_signature_recovery.py. b Number of principal components tested and signatures generated for bootstrap samples with different sample sizes. For each size, GO-PCA was applied to 50 bootstrap samples of that size. Shown are the median values for each measure, and error bars indicate the inter-quartile range. This figure panel was generated using the GO-PCA script gopca_plot_bootstrap_numbers.py. c Bootstrap detection rates of all DMAP GO-PCA signatures, as a function of the number of PCs included in the analysis (left), and as a function of the size of the bootstrap samples (right). This figure panel was generated using the GO-PCA scripts gopca_plot_bootstrap_pc_matrix.py and gopca_plot_bootstrap_sample_size_matrix.py.
Fig 3 Analysis of the robustness of DMAP GO-PCA signatures using bootstrapping. a Detection rates of the 50 DMAP signatures generated by GO-PCA (see Fig 2). GO-PCA was applied to 50 bootstrap samples with the same size as the original dataset (n = 211). In each analysis, a signature from the original analysis was counted as detected if there existed a signature based on the exact same GO term (yellow bars), or based on either the same GO term or any “related” GO term (blue bars). Related GO terms were defined as all ancestor and descendent terms in the Gene Ontology. This figure panel was generated using the GO-PCA script gopca_plot_bootstrap_signature_recovery.py. b Number of principal components tested and signatures generated for bootstrap samples with different sample sizes. For each size, GO-PCA was applied to 50 bootstrap samples of that size. Shown are the median values for each measure, and error bars indicate the inter-quartile range. This figure panel was generated using the GO-PCA script gopca_plot_bootstrap_numbers.py. c Bootstrap detection rates of all DMAP GO-PCA signatures, as a function of the number of PCs included in the analysis (left), and as a function of the size of the bootstrap samples (right). This figure panel was generated using the GO-PCA scripts gopca_plot_bootstrap_pc_matrix.py and gopca_plot_bootstrap_sample_size_matrix.py.CC BY · PubMed Central
Fig 4 Application of GO-PCA to 214 mouse transcriptomes representing diverse immune and stromal cell types (IGP1). [36] Signatures are labeled and ordered as in Fig 2. Arrows and boxes indicate associations discussed in the main text. Samples are ordered according to their known lineage identities.
Fig 4 Application of GO-PCA to 214 mouse transcriptomes representing diverse immune and stromal cell types (IGP1). [36] Signatures are labeled and ordered as in Fig 2. Arrows and boxes indicate associations discussed in the main text. Samples are ordered according to their known lineage identities.CC BY · PubMed Central
Fig 5 Application of GO-PCA to 479 primary glioblastomas (GBM). Shown is the signature matrix generated by GO-PCA as a heat map. Signatures are labeled and ordered as in S2 Fig, and samples are ordered using hierarchical clustering with correlation distance and average linkage. Colored bars at the side indicate the four main groups of signatures discussed in the text (blue = neuronal; gray = proliferative; yellow = immunological; red = extracellular matrix (ECM)-related). This figure (without colored bars) was generated using the GO-PCA script gopca_plot_signature_matrix.py.
Fig 5 Application of GO-PCA to 479 primary glioblastomas (GBM). Shown is the signature matrix generated by GO-PCA as a heat map. Signatures are labeled and ordered as in S2 Fig, and samples are ordered using hierarchical clustering with correlation distance and average linkage. Colored bars at the side indicate the four main groups of signatures discussed in the text (blue = neuronal; gray = proliferative; yellow = immunological; red = extracellular matrix (ECM)-related). This figure (without colored bars) was generated using the GO-PCA script gopca_plot_signature_matrix.py.CC BY · PubMed Central
Fig 6 Associations between signatures generated by GO-PCA analysis of the GBM dataset, and five previously defined GBM subtypes [40]. Signatures are shown in the same order as in Fig 5. Left panel: Heat map showing median signature expression levels for each signature in each GBM subtype. Right panel: Heat map showing significance of association, as determined by two-sided Mann-Whitney U tests (see Methods). Only associations with p-values significant at the α = 0.05 significance level after Bonferroni correction are shown.
Fig 6 Associations between signatures generated by GO-PCA analysis of the GBM dataset, and five previously defined GBM subtypes [40]. Signatures are shown in the same order as in Fig 5. Left panel: Heat map showing median signature expression levels for each signature in each GBM subtype. Right panel: Heat map showing significance of association, as determined by two-sided Mann-Whitney U tests (see Methods). Only associations with p-values significant at the α = 0.05 significance level after Bonferroni correction are shown.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

Show all 60 references

Cited by

See all 35

Metadata

Authors
Wagner, F.
Affiliations
Duke University (US)
Corresponding author
Wagner, Florian
Venue
PLOS ONE
Published
2015-11-17
Volume
10
Issue
11
Pages
e0143196
ISSN
1932-6203 (ISSN-L)
OpenAlex
W2174288164
Cited by
35
References
60
Citations by year
2026: 1 · 2025: 2 · 2024: 4 · 2023: 3 · 2022: 4 · 2021: 7 · 2020: 2 · 2019: 2 · 2018: 4 · 2017: 2 · 2016: 3 · 2015: 1
FWCI
1.45
Type
journal article
Language
en
Open access
Locations
  • PLoS ONE (publishedVersion, cc-by)
  • bioRxiv (Cold Spring Harbor Laboratory) (acceptedVersion, cc-by-nc-nd)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • RePEc: Research Papers in Economics (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • Figshare (submittedVersion, cc-by)
Topics
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Gene expression and cancer classification
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Bioinformatics and Genomic Networks
  • Life Sciences › Biochemistry, Genetics and Molecular Biology › Molecular Biology › Genomics and Phylogenetic Studies
Keywords
gene expression signatures, gene ontology enrichment, unsupervised dimensionality reduction, principal component analysis, hematopoietic lineage, glioblastoma subtypes, bootstrapping, hypothesis generation
MeSH terms
Algorithms, Animals, Glioblastoma, Hematopoietic Stem Cells, Humans, Neoplasm Staging, Software, Cluster Analysis, Genome, Cell Lineage, Principal Component Analysis, Databases, Genetic, Mice, Molecular Sequence Annotation, Transcriptome

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: GO-PCA: An Unsupervised Method to Explore Gene Expression Data Using Prior Knowledge. https://app.cochat.ai/papers/10.1371/journal.pone.0143196
The paper
Wagner, F. (2015). GO-PCA: An Unsupervised Method to Explore Gene Expression Data Using Prior Knowledge. PLOS ONE. https://doi.org/10.1371/journal.pone.0143196

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.