SkeletalVis: Exploration and Visualisation of Skeletal Transcriptomics Data

Description

Allows search and visualisation of a collection of uniformly processed skeletal transcriptomic datasets. Includes methods to identify datasets where genes of interest are differentially expressed and find datasets with a similar gene expression pattern to a query dataset Soul J, Hardingham TE, Boot-Handford RP, Schwartz JM (2019) <doi.org/10.1093/bioinformatics/bty947>.

Author(s)

Maintainer: Jamie Soul jamie.soul@liverpool.ac.uk [copyright holder]

See Also

Useful links:

• https://github.com/soulj/SkeletalVis

• Report bugs at https://github.com/soulj/SkeletalVis/issues

Browse the skeletalvis database metadata

Description

Opens an interactive table showing the experiments available for analysis. Row can be selected to return the datasetID for the comparison of interest within that experiment.

Usage

browse_skeletalvis(skeletalvis)

Arguments

Value

The datasetID of the selected gene expression profile from the interactive table.

Examples

if(interactive()){
skeletalvis <- load_skeletalvis(demo=TRUE)
selected_id <- browse_skeletalvis(skeletalvis)
}

Get cosine similarity for a query dataset against the skeletalvis database

Description

Computes the cosine similarity of the log2 fold changes of a given query dataset against the SkeletalVis fold changes, returning a data frame of experiment similarities.

Usage

experiment_similarity(skeletalvis, dataset, add_meta_data = TRUE)

Arguments

Value

A data frame containing cosine similarity values, IDs, and z-scores.

Examples

skeletalvis <- load_skeletalvis(demo=TRUE)

# Create a query dataset (this should be a data frame with the first column as gene IDs)
query_dataset <- data.frame(ID = c("SOX9", "ACAN"), fold_change = c(2.5, -1.8))

# Get cosine similarities
similarity_results <- experiment_similarity(
  skeletalvis = skeletalvis,
  dataset = query_dataset,
)

# View results
head(similarity_results)

Retrieve metadata for the SkeletalVis database

Description

'get_comparisons'Retieves the metadata file containing accession and comparison information for the skeletalvis database.

Usage

get_comparisons(skeletalvis, accession = NULL)

Arguments

Value

A dataframe containing metadata.

Examples

# Get the path to the directory with the accessions.txt file
skeletalvis_path <- load_skeletalvis(demo=TRUE)

# Retrieve the metadata
metadata <- get_comparisons(skeletalvis_path)

# View the metadata
head(metadata)

metadata <- get_comparisons(skeletalvis_path, "GSE85761")

Retrieve the experiment table for the SkeletalVis database

Description

Retrieves the experiment table file containing experiment accessions and descriptions from the skeletalvis database.

Usage

get_exp_table(skeletalvis)

Arguments

Value

A dataframe containing the experiment information

Examples

# Get the path to the directory with the accessions.txt file
skeletalvis_path <- load_skeletalvis(demo=TRUE)

# Retrieve the metadata
exptable <- get_exp_table(skeletalvis_path)

# View the metadata
head(exptable)

Get the gene differential expression results for an experiment of interest

Description

This function loads the differential expression data (log2 foldchanges and FDR values) for a given dataset id

Usage

get_experiment(skeletalvis, dataset_id)

Arguments

Value

A data frame containing differential expression results for the specified dataset ID.

Examples

skeletalvis <- load_skeletalvis(demo=TRUE)
experiment_results <- get_experiment(skeletalvis, "GSE12860_6")

Get gene differential expression results for genes of interest

Description

Extracts the fold change and p-values across the SkeletalVis database for the specified genes.

Usage

get_gene_fold_changes(
  skeletalvis,
  gene_symbols,
  return_fdr = TRUE,
  add_meta_data = TRUE
)

Arguments

Value

A tibble containing gene expression results for the specified genes.

Examples

skeletalvis <- load_skeletalvis(demo=TRUE)

gene_results <- get_gene_fold_changes(skeletalvis, c("SOX9","ACAN"))

head(gene_results)

Load SkeletalVis data

Description

Checks the default SkeletalVis data location for the differential expression data and metadata files. If any files are missing, it lists them and ask permission to download the missing files.

Usage

load_skeletalvis(verbose = TRUE, ask = TRUE, demo = FALSE)

Arguments

Value

The path to the SkeletalVis data folder.

Examples

# Load the demo SkeletalVis data.
skeletalvis <- load_skeletalvis(demo = TRUE)

# Download full dataset (only if running interactively)
if (interactive()) {
  skeletalvis <- load_skeletalvis()
}

Plot ranked cosine similarity scores with top n labels

Description

Takes a similarity table with columns for ID, cosine, and zscore values, and generates a plot showing the rank of the z-scores on a log scale. Labels for the top n IDs with the highest z-scores are displayed on the plot.

Usage

plot_similarity(similarity_table, top_n = 10)

Arguments

Value

A ggplot object displaying the ranked z-scores on a log scale, with labels for the top n items by zscore.

Examples

data(query)
skeletalvis <- load_skeletalvis(demo=TRUE)
similarity_results <- experiment_similarity(skeletalvis, query)

# Plot similarity table with labels for top 5 items
plot_similarity(similarity_results, top_n = 5)

Example gene expression dataset

Description

An example gene expression dataset comparing damaged versus intact osteoarthritic cartilage.

Usage

query

Format

## 'query' A data frame with 42,459 rows and 2 columns:

ID

Human Gene Symbol

'E-MTAB-4304_1'

log2 fold change

Source

<https://www.ebi.ac.uk/biostudies/arrayexpress/studies/E-MTAB-4304>

Examples

head(query)

Search for experiments in SkeletalVis

Description

Search for experiments matching a search term in the SkeletalVis database. The search can be done across the entire database or specific metadata columns.

Usage

search_skeletalvis(skeletalvis, search_term, columns = NULL)

Arguments

Value

A data frame containing experiments that match the search term in the specified columns or the whole table.

Examples

skeletalvis <- load_skeletalvis(demo=TRUE)
# Search across all columns for rows containing "SOX9"
result <- search_skeletalvis(skeletalvis, "SOX9")

# Search only in the 'Perturbation' and 'Description' columns
result <- search_skeletalvis(skeletalvis, "SOX9", columns = c("Perturbation", "Description"))

View curated osteoarthritis associated genes

Description

Retrieves the OATargets table of literature curated genes associated with osteoarthritis joint damage in animals.

Usage

view_curated_oagenes(skeletalvis)

Arguments

Value

A data frame with the following columns:

PMID

The PubMed ID of the paper describing the observation.

Gene

The human gene symbol of the gene modulated in the study.

Effect on gene product

The effect of the gene modulation on the gene product.

Model

The type of OA model where the gene has been modulated.

Susceptibility observed

Increase (detrimental) or decrease (protective) in OA severity upon gene modulation.

Inferred gene effect

Based on the modulation effect, whether the gene is protective or detrimental in OA.

Delivery

The tissue specificity of the gene modulation.

Species

The species under study.

pub_date

The date of publication (epub).

LastAuthor

The last author of the publication.

Type

Genetic or exogenous modulation.

Intervention

For exogenous modulations, the method of gene modulation.

simpleModel

The broad type of OA model where the gene has been modulated.

effectConsensus

Consensus inferred gene effect across observations; 'ambiguous' if discrepancies exist.

NumStudies

The number of publications studying each gene.

Examples

skeletalvis <- load_skeletalvis(demo=TRUE)
oatargets <- view_curated_oagenes(skeletalvis)

View the OATarget gene network

Description

This function visualises the protein-protein interaction network surrounding a specified gene, highlighting those genes that have been experimentally pertubated and the resulting effect of OA.

Usage

view_network(skeletalvis, query, hide_unannotated = TRUE, physics = TRUE)

Arguments

Details

The function extracts the subnetwork surrounding the specified gene. Nodes are coloured based on their effect ("Protective", "Detrimental", "Ambiguous", "No effect", or "Not measured"). If 'hide_unannotated = TRUE', nodes not yet studied are removed unless they are the target gene.

Value

A visNetwork interactive network'

Examples

# Specify the path to the skeletalvis directory
skeletalvis <- load_skeletalvis(demo=TRUE)

# Visualize the network for a specific gene
vis <- view_network(skeletalvis, "COL2A1")

# Show unannotated nodes
vis_filtered <- view_network(skeletalvis, "COL2A1", hide_unannotated = FALSE)

View prioritised osteoarthritis associated genes

Description

Retrieves the machine learning model prioritised table of osteoarthritis joint damage in animals. Genes are ranked by their probability to cause an OA damage phenotype when experimentally perturbed. The predicted direction of effect and druggability of that gene is also provided.

Usage

view_prioritised_oagenes(skeletalvis)

Arguments

Value

A data frame with the following columns:

Gene

The human gene symbol.

Rank

The rank of the priortisation scores from the machine learning model

PredictedEffect

The predicted effect this gene

Category_sm

The small molecular tractibility of this gene

Category_ab

The antibody tractibility of this gene

Examples

skeletalvis <- load_skeletalvis(demo=TRUE)
oatargets <- view_prioritised_oagenes(skeletalvis)

Volcano Plot of Gene Expression Data

Description

Creates a volcano plot showing the log2 fold change and FDR values with optional labelling of points.

Usage

volcano_plot(
  data,
  number_points = 5,
  selected_points = NULL,
  interactive = FALSE,
  logFC_threshold = log2(1.5),
  FDR_threshold = 0.05,
  point_size = 2,
  lab_size = 4
)

Arguments

Value

A ggplot object representing the volcano plot.

Examples

skeletalvis <- load_skeletalvis(demo=TRUE)

experiment_results <- get_experiment(skeletalvis, "GSE155118_1")

volcano_plot(experiment_results)