AnnotatedBCGEData
Abstract
AnnotatedBCGEData is an R/Bioconductor package that provides
access to more than 100 gene-expression datasets generated using
microarray and RNA-Sequencing technologies. We collected these datasets
from public repositories, including Gene Expression Omnibus, ArrayExpress,
and The
Cancer Genome Atlas. Where feasible, we obtained and reprocessed the
raw data using computational pipelines and annotations that are
standardized and more updated than those used to process the data
originally. Additionally, we annotated the sample metadata using
ontology terms from the National
Cancer Institute Thesaurus. This was necessary because different
data submitters used different terminology to describe breast-cancer
patients and samples. Mapping the metadata to ontology terms makes it
possible for users to search the data using these standardized
terms.
Installation instructions
Get the latest stable R release from CRAN.
Then install AnnotatedBCGEData from Bioconductor using the following
code:
if (!requireNamespace("BiocManager", quietly = TRUE)) {
install.packages("BiocManager")
}
BiocManager::install("AnnotatedBCGEData")Or install the development version:
Finally, load the package:
Data retrieval
Attempting to store all of the data within this package would be infeasible due to the data’s size. Thus, we store the data remotely. Once a user has identified dataset(s) to analyze, our package retrieves the data from a public Zenodo repository. (If they prefer, researchers can find details about the data and access the files directly by exploring our Zenodo repository with a Web browser.)
Searching for ontology terms
This package contains functions designed to help users filter the
included datasets based on their associated ontology terms. The first
function that researchers typically use is
searchOntologyTerms. Its purpose is to help users identify
ontology terms mapped to the data. It accepts two parameters: a string
to search by and what should be searched. The second parameter allows
the user to indicate whether the string is expected to match the “Name”
(default), “Definition”, “URI”, or “Code” of the ontology term(s). The
function returns a tibble with information about the matching terms. The
user reviews the matching term(s) and retrieves the corresponding
ontology code(s), which can then be passed to the
searchForDatasetsByField or
searchForDatasetsByValue functions, which return
information about datasets associated with the term(s).
In the following example, we are using the
searchOntologyTerms function to search for ontology terms
that match the string, “Progesterone Receptor Status”. Because we did
not specify a value for the term_type argument, the default
of “Name” is used.
## ℹ Number of records: 3## ℹ Successfully fetched list of published records - page 1 (size = 10)## ✔ Successfully fetched list of published records!## ✔ Successfully fetched published record for concept DOI '10.5281/zenodo.17545644'!## # A tibble: 5 × 4
## Code Preferred_Label Synonyms Definitions
## <chr> <chr> <chr> <chr>
## 1 C16149 Progesterone Receptor Status Progesterone Receptor|Pro… Indicates …
## 2 C147856 Progesterone Receptor Status 0 PgR Status 0|Progesterone… An immunoh…
## 3 C147857 Progesterone Receptor Status 1+ Progesterone Receptor Sta… An immunoh…
## 4 C147858 Progesterone Receptor Status 2+ Progesterone Receptor Sta… An immunoh…
## 5 C147859 Progesterone Receptor Status 3+ Progesterone Receptor Sta… An immunoh…In the following similar example, we are searching for ontology terms that match the shorter (and lowercase) string, “receptor status”.
## # A tibble: 11 × 4
## Code Preferred_Label Synonyms Definitions
## <chr> <chr> <chr> <chr>
## 1 C16150 Estrogen Receptor Status Estrogen Receptor Alpha … "The estro…
## 2 C147852 Estrogen Receptor Status 0 ER Status 0|Estrogen Rec… "An immuno…
## 3 C147853 Estrogen Receptor Status 1+ ER Status 1+|Estrogen Re… "An immuno…
## 4 C147854 Estrogen Receptor Status 2+ ER Status 2+|Estrogen Re… "An immuno…
## 5 C147855 Estrogen Receptor Status 3+ ER Status 3+|Estrogen Re… "An immuno…
## 6 C16149 Progesterone Receptor Status Progesterone Receptor|Pr… "Indicates…
## 7 C147856 Progesterone Receptor Status 0 PgR Status 0|Progesteron… "An immuno…
## 8 C147857 Progesterone Receptor Status 1+ Progesterone Receptor St… "An immuno…
## 9 C147858 Progesterone Receptor Status 2+ Progesterone Receptor St… "An immuno…
## 10 C147859 Progesterone Receptor Status 3+ Progesterone Receptor St… "An immuno…
## 11 C94299 Receptor Status Receptor Status "Refers to…The function returns a tibble containing details about ontology terms related to both progesterone receptor status and estrogen receptor status.
The National Cancer Institute Thesaurus provides synonyms for most
terms. When a user searches by “Name” with the
searchOntologyTerms function, it automatically checks
against synonyms as well. This is helpful when the search text does not
match the “preferred name” of an ontology term but instead matches a
synonym. For example, we could use the following code to search for
ontology terms related to progesterone receptor status using a synonym
that is commonly used in medical contexts.
## # A tibble: 5 × 4
## Code Preferred_Label Synonyms Definitions
## <chr> <chr> <chr> <chr>
## 1 C16149 Progesterone Receptor Status Progesterone Receptor|Pro… Indicates …
## 2 C147856 Progesterone Receptor Status 0 PgR Status 0|Progesterone… An immunoh…
## 3 C147857 Progesterone Receptor Status 1+ Progesterone Receptor Sta… An immunoh…
## 4 C147858 Progesterone Receptor Status 2+ Progesterone Receptor Sta… An immunoh…
## 5 C147859 Progesterone Receptor Status 3+ Progesterone Receptor Sta… An immunoh…Another alternative is to search ontology term definitions. This is helpful when you don’t know the exact term but understand the concept you’re looking for. In this example, we are searching for definitions matching the string, “immunohistochemical”.
## # A tibble: 16 × 4
## Code Preferred_Label Synonyms Definitions
## <chr> <chr> <chr> <chr>
## 1 C147852 Estrogen Receptor Status 0 ER Status… An immunoh…
## 2 C147853 Estrogen Receptor Status 1+ ER Status… An immunoh…
## 3 C147854 Estrogen Receptor Status 2+ ER Status… An immunoh…
## 4 C147855 Estrogen Receptor Status 3+ ER Status… An immunoh…
## 5 C147524 HER2/Neu Expression by Immunohistochemistry 0 HER2/Neu … An immunoh…
## 6 C147525 HER2/Neu Expression by Immunohistochemistry 1+ HER2/Neu … An immunoh…
## 7 C147526 HER2/Neu Expression by Immunohistochemistry 2+ HER2/Neu … An immunoh…
## 8 C147527 HER2/Neu Expression by Immunohistochemistry 3+ HER2/Neu … An immunoh…
## 9 C141481 HER2/Neu Positive by Immunohistochemistry HER2/Neu … An immunoh…
## 10 C185751 HER2/Neu Status by Immunohistochemistry HER2/Neu … The result…
## 11 C84373 IGF1R Immunohistochemical Expression IGF1R Imm… An immunoh…
## 12 C175007 Immunohistochemical Test Result Immunohis… The result…
## 13 C147856 Progesterone Receptor Status 0 PgR Statu… An immunoh…
## 14 C147857 Progesterone Receptor Status 1+ Progester… An immunoh…
## 15 C147858 Progesterone Receptor Status 2+ Progester… An immunoh…
## 16 C147859 Progesterone Receptor Status 3+ Progester… An immunoh…In this case, the function returns a tibble containing details about ontology terms for concepts related to using immunohistochemical methods for profiling. This includes terms related to progesterone receptor status as well as other cell-surface markers.
Identifying datasets associated with specific ontology terms
After we have identified ontology terms for the research topic we
wish to study, we can identify datasets associated with those terms. The
searchForDatasetsByField function accepts codes for
ontology terms. Each of these codes should be a unique identifier
associated with an ontology term (typically found using the
searchOntologyTerms function). The
searchForDatasetsByField function checks the ontology terms
associated with metadata columns (fields) and returns a tibble with the
identifier(s) of the associated dataset(s), the names of columns that
have been mapped to any of the terms, and the associated codes. The
dataset identifiers can then be passed to the getDataset
function to retrieve the data.
Similarly, the searchForDatasetsByValue function accepts
codes for ontology terms that have been mapped to values within a given
column (field). The function searches across the datasets and returns a
tibble with information about the matching datasets. The dataset
identifiers can then be passed to the getDataset function
to retrieve the data.
In the previous examples, we searched for ontology terms related to
progesterone receptor status. One of the matching terms used
“Progesterone Receptor Status” as its name “C16149” as its code. We also
identified “C16150” as the code associated with “Estrogen Receptor
Status.” In the following example, we are using the
searchForDatasetsByField function to search for datasets
with a column (field) that has been mapped to either of these ontology
terms.
## ℹ Number of records: 6## ℹ Successfully fetched list of published records - page 1 (size = 10)## ✔ Successfully fetched list of published records!## ✔ Successfully fetched published record for concept DOI '10.5281/zenodo.17583904'!## Rows: 67866 Columns: 8
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: "\t"
## chr (8): Dataset_ID, Original_Field_Name, Original_Value, Basic_Data_Type_On...
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.## # A tibble: 134 × 3
## Dataset_ID Original_Field_Name Code
## <chr> <chr> <chr>
## 1 GSE61304 er C16150
## 2 GSE20271 er_percent_positive C16150
## 3 GSE5847 ER_status C16150
## 4 ABiM.100 er C16150
## 5 GSE11001 er C16150
## 6 GSE19615 er C16150
## 7 GSE19697 er C16150
## 8 GSE24185 er C16150
## 9 GSE2990 er C16150
## 10 GSE43365 er C16150
## # ℹ 124 more rowsAlternatively (or instead), we can use the
searchForDatasetsByValue function to search for the
datasets with values in any column that have been associated with a
specified ontology term. For example, if we wanted to identify datasets
with samples that had been indicated as positive for the progesterone
receptor, we could use the code “C15496” for the Progesterone Receptor
Positive term.
## Rows: 67866 Columns: 8
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: "\t"
## chr (8): Dataset_ID, Original_Field_Name, Original_Value, Basic_Data_Type_On...
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.## # A tibble: 67,866 × 8
## Dataset_ID Original_Field_Name Original_Value Basic_Data_Type_Ontolog…¹
## <chr> <chr> <chr> <chr>
## 1 GSE62944_Tumor ajcc_staging_edition 4th Discrete Set String Data…
## 2 GSE62944_Tumor ajcc_staging_edition 4th Discrete Set String Data…
## 3 GSE62944_Normal ajcc_staging_edition 5th Discrete Set String Data…
## 4 GSE62944_Normal ajcc_staging_edition 5th Discrete Set String Data…
## 5 GSE62944_Tumor ajcc_staging_edition 5th Discrete Set String Data…
## 6 GSE62944_Tumor ajcc_staging_edition 5th Discrete Set String Data…
## 7 GSE62944_Normal ajcc_staging_edition 6th Discrete Set String Data…
## 8 GSE62944_Normal ajcc_staging_edition 6th Discrete Set String Data…
## 9 GSE62944_Tumor ajcc_staging_edition 6th Discrete Set String Data…
## 10 GSE62944_Tumor ajcc_staging_edition 6th Discrete Set String Data…
## # ℹ 67,856 more rows
## # ℹ abbreviated name: ¹Basic_Data_Type_Ontology_Term
## # ℹ 4 more variables: Field_Name_Ontology_Terms <chr>,
## # Value_Ontology_Terms <chr>, NCIT_field_code <chr>, NCIT_value_code <chr>## # A tibble: 59 × 5
## Dataset_ID Original_Field_Name Field_Code Original_Value Code
## <chr> <chr> <chr> <chr> <chr>
## 1 GSE16391 er_pgr C16150 1 C15496
## 2 GSE16391 er_pgr C16149 1 C15496
## 3 GSE2603 path_pr_status C16149 P C15496
## 4 GSE61304 pgr C16149 1 C15496
## 5 GSE6532_U133A pgr C16149 1 C15496
## 6 GSE6532_U133B pgr C16149 1 C15496
## 7 GSE6532_U133Plus2 pgr C16149 1 C15496
## 8 GSE9195 pgr C16149 1 C15496
## 9 GSE45255 pgr_status C16149 PgR+ C15496
## 10 GSE58984 pgr_status C16149 1 C15496
## # ℹ 49 more rowsObtaining dataset-level metadata
In addition to using the ontology, users can decide which datasets to
analyze based on metadata about the datasets. This is possible using the
getDatasetMetadata function, which accepts a vector of
dataset identifiers. Below is an example.
## ℹ Number of records: 3## ℹ Successfully fetched list of published records - page 1 (size = 10)## ✔ Successfully fetched list of published records!## ✔ Successfully fetched published record for concept DOI '10.5281/zenodo.17780657'!## Rows: 102 Columns: 12
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: "\t"
## chr (12): accession_id, contact_city, contact_country, contact_institute, av...
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.## # A tibble: 3 × 6
## Dataset_ID Source Experiment_Type Title Summary Overall_Design
## <chr> <chr> <chr> <chr> <chr> <chr>
## 1 GSE10797 GEO Expression profiling by array Transc… The mo… Breast tissue…
## 2 GSE41197 GEO Expression profiling by array Differ… We use… RNA from huma…
## 3 GSE59772 GEO Expression profiling by array Analys… Triple… Three FFPE br…Obtaining gene-expression data (as SummarizedExperiment objects)
Installing our package does not download the datasets themselves. The
data files are stored in a Zenodo repository and retrieved on demand. To
obtain data, use the getDataset function, which stores the
data and metadata in a SummarizedExperiment object.
The SummarizedExperiment object for each dataset
includes 3 items. One is the gene-expression data, a numeric matrix with
samples as columns and genes as rows. Ensembl gene identifiers are used
as row names. Each data value is the expression level for its respective
sample and gene. Second is the sample metadata, which often includes
patient demographics and sample details. Each row represents a sample,
and each column represents a different metadata variable. Third is
feature metadata: information about the genes that were profiled for
each sample. Each row represents a gene and provides information like
the Entrez gene ID, gene symbol, and chromosome identifier.
## ℹ Number of records: 2## ℹ Successfully fetched list of published records - page 1 (size = 10)## ✔ Successfully fetched list of published records!## ✔ Successfully fetched published record for concept DOI '10.5281/zenodo.21985420'!## Rows: 108 Columns: 4
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: "\t"
## chr (4): Dataset_ID, Expressions_DOI, Expressions_Filename, Metadata_Filename
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.
## ℹ Number of records: 4
##
## ℹ Successfully fetched list of published records - page 1 (size = 10)
##
## ✔ Successfully fetched list of published records!
##
## ✔ Successfully fetched published record for concept DOI '10.5281/zenodo.17428997'!
##
## ℹ Number of records: 4
##
## ℹ Successfully fetched list of published records - page 1 (size = 25)
##
## ✔ Successfully fetched list of published records!
##
## Either the data was not found in the cache or a newer version of the data was identified. Downloading now.
##
## ℹ Number of records: 4
##
## ℹ Successfully fetched list of published records - page 1 (size = 10)
##
## ✔ Successfully fetched list of published records!
##
## ✔ Successfully fetched published record for concept DOI '10.5281/zenodo.17428997'!
##
## Rows: 9593 Columns: 22
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: "\t"
## chr (5): Dataset_ID, HGNC_Symbol, Ensembl_Gene_ID, Chromosome, Gene_Biotype
## dbl (17): Entrez_Gene_ID, GSM1010328, GSM1010329, GSM1010330, GSM1010331, GS...
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.
## Rows: 16 Columns: 5
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: "\t"
## chr (5): Dataset_ID, Sample_ID, Platform_ID, Patient_ID, disease_state
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.expression_data = assay(GSE41197_SE)
sample_metadata = colData(GSE41197_SE)
feature_data = rowData(GSE41197_SE)The colData and rowData functions of the
SummarizedExperiment class return specialized data
structures. These can be easily converted to data frames using
as.data.frame. However, converting directly to the
tibble type is undesirable because it removes the row
names.
The getDataset function creates a cache (using
BiocFileCache) on the user’s machine. All downloaded files
are saved to this cache so that files do not need to be re-downloaded
when the same dataset is retrieved. By default, the cache location is
set using tempdir(). This directory is removed when the R
session ends. Therefore, the user may specify an alternative cache
directory if they wish to ensure the cached files persist between R
sessions.
In some cases, multiple versions of a given dataset are available on
Zenodo. The getDataset function allows the user to specify
which version of the data to use. By default, the most recent version is
downloaded. The v argument accepts an integer specifying
the version. The following example illustrates retrieving a specific
version of a dataset.
## Rows: 108 Columns: 4
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: "\t"
## chr (4): Dataset_ID, Expressions_DOI, Expressions_Filename, Metadata_Filename
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.
## ℹ Number of records: 4
##
## ℹ Successfully fetched list of published records - page 1 (size = 10)
##
## ✔ Successfully fetched list of published records!
##
## ✔ Successfully fetched published record for concept DOI '10.5281/zenodo.17428997'!
##
## ℹ Number of records: 4
##
## ℹ Successfully fetched list of published records - page 1 (size = 25)
##
## ✔ Successfully fetched list of published records!
##
## Either the data was not found in the cache or a different version was requested. Downloading now.
##
## ℹ Number of records: 4
##
## ℹ Successfully fetched list of published records - page 1 (size = 10)
##
## ✔ Successfully fetched list of published records!
##
## ✔ Successfully fetched published record for concept DOI '10.5281/zenodo.17428997'!
##
## ℹ Number of records: 4
##
## ℹ Successfully fetched list of published records - page 1 (size = 25)
##
## ✔ Successfully fetched list of published records!
##
## ℹ Number of records: 1
##
## ℹ Successfully fetched list of published records - page 1 (size = 10)
##
## ✔ Successfully fetched list of published records!
##
## ✔ Successfully fetched record for DOI '10.5281/zenodo.17429158'!
##
## Rows: 9593 Columns: 22
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: "\t"
## chr (5): Dataset_ID, HGNC_Symbol, Ensembl_Gene_ID, Chromosome, Gene_Biotype
## dbl (17): Entrez_Gene_ID, GSM1010328, GSM1010329, GSM1010330, GSM1010331, GS...
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.
## Rows: 16 Columns: 5
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: "\t"
## chr (5): Dataset_ID, Sample_ID, Platform_ID, Patient_ID, disease_state
##
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.v2_expression_data = assay(GSE41197_v2)
v2_metadata = colData(GSE41197_v2)
v2_feature_data = rowData(GSE41197_v2)If the specified version does not exist or needs to be downloaded, a message will be displayed.
The getDataset function accepts a single dataset
identifier. At the time of package development, the Zenodo API limits
the number of requests per minute to 30. Because of this, we have
included logic in our package to pause and retry if a request to
download a dataset fails. These API limits should not cause problems in
most cases, but users should be aware that this may cause delays when
downloading large amounts of data.
Example 1: Plotting gene-expression values
In this example, we demonstrate simple data exploration. We first access the gene-expression matrix. To make the data easier to plot, we then convert the matrix into a tibble and pivot the data. We use ggplot2 functions to create a box plot showing the gene expression levels across the samples for 10 genes. (The values are centered at zero, which explains why some values are negative.)
# Set the random seed so the graph is reproducible.
set.seed(0)
as_tibble(expression_data, rownames='Ensembl_Gene_ID') %>%
head(n = 10) %>%
pivot_longer(-Ensembl_Gene_ID, names_to = "Sample_ID", values_to = "Expression_Level") %>%
ggplot(aes(x=Ensembl_Gene_ID, y=Expression_Level)) +
geom_boxplot() +
geom_jitter(color = "red", alpha = 0.2) +
labs(x='Ensembl Gene ID', y='Expression level') +
theme_bw() +
theme(axis.text.x = element_text(angle = 45, hjust = 1))
Example 2: Using metadata to create a heatmap
In this example, we download the data and metadata with the eventual goal of creating a heatmap of gene-expression levels. First, we load two datasets, convert them into tibbles, and identify samples from stromal cells. Then we limit the SummarizedExperiment data to those samples.
# Retrieve the data from Zenodo
GSE10797_SE <- getDataset("GSE10797")
GSE59772_SE <- getDataset("GSE59772")
# Retrieve stroma sample IDs as vectors
GSE10797_stroma_sample_ids = colData(GSE10797_SE) %>%
as_tibble(rownames = 'Sample_ID') %>%
filter(tissue_source == 'stromal cells from breast cancer patient') %>%
pull(Sample_ID)
GSE59772_stroma_sample_ids = colData(GSE59772_SE) %>%
as_tibble(rownames = 'Sample_ID') %>%
filter(tissue == 'Stroma') %>%
pull(Sample_ID)
GSE10797_SE = GSE10797_SE[,colnames(GSE10797_SE) %in% GSE10797_stroma_sample_ids]
GSE59772_SE = GSE59772_SE[,colnames(GSE59772_SE) %in% GSE59772_stroma_sample_ids]Now we filter the data to genes that both data sets have in common.
common_genes <- intersect(rownames(GSE10797_SE), rownames(GSE59772_SE))
GSE10797_SE <- GSE10797_SE[common_genes,]
GSE59772_SE <- GSE59772_SE[common_genes,]
# This also has the benefit that the genes will be in the same order in both.Finally, we create a tibble with the data from both sources. After doing so, we calculate the variance across the 5 samples. We then select the 10 genes with the highest variance and create a heatmap of those. Please note that this analysis is simplified. In an actual data analysis, we would need to account for systematic biases due to combining datasets from disparate sources.
combined_data = cbind(assay(GSE10797_SE), assay(GSE59772_SE)) %>%
as_tibble(rownames = 'Gene_ID') %>%
pivot_longer(-Gene_ID, names_to = 'Sample_ID', values_to = 'Expression_Level')
top_variance_genes = group_by(combined_data, Gene_ID) %>%
summarize(Variance = var(Expression_Level)) %>%
arrange(desc(Variance)) %>%
head(n = 10) %>%
pull(Gene_ID)
combined_data = filter(combined_data, Gene_ID %in% top_variance_genes)
ggplot(combined_data, aes(x = Sample_ID, y = Gene_ID, fill = Expression_Level)) +
geom_tile() +
labs(x = 'Sample', y = 'Gene', fill = 'Expression level',
title = 'Genes in stromal samples with highest variance') +
theme_bw() +
theme(axis.text.x = element_text(angle = 45, hjust = 1),
plot.title = element_text(hjust = 0.5, face = "bold"))
Example 3: Using Feature Data
In this final example, we use gene annotations to identify which genes to plot. We filter each dataset to only include lncRNA genes on either chromosome 1 or 2. Because we previously limited the genes to those present in both datasets, we only need to use one of the datasets to identify the relevant genes.
gene_ids = rowData(GSE10797_SE) %>%
as_tibble(rownames = 'Ensembl_Gene_ID') %>%
filter(Chromosome %in% 1:2 & Gene_Biotype == "lncRNA") %>%
pull(Ensembl_Gene_ID)
GSE10797_SE <- GSE10797_SE[gene_ids,]
GSE59772_SE <- GSE59772_SE[gene_ids,]Next, we extract the expression values, put them into tibbles, and pivot them.
gse10797_data = assay(GSE10797_SE) %>%
as_tibble(rownames = 'Gene_ID') %>%
pivot_longer(-Gene_ID, names_to = 'Sample_ID', values_to = 'Expression_Level') %>%
mutate(Dataset_ID = "GSE10797")
gse59772_data = assay(GSE59772_SE) %>%
as_tibble(rownames = 'Gene_ID') %>%
pivot_longer(-Gene_ID, names_to = 'Sample_ID', values_to = 'Expression_Level') %>%
mutate(Dataset_ID = "GSE59772")
combined_data = bind_rows(gse10797_data, gse59772_data)Finally, we combine the data sets using bind_rows() and read it into a ggplot.
ggplot(combined_data, aes(x = Gene_ID, y = Expression_Level, fill = Dataset_ID)) +
geom_boxplot() +
theme_bw() +
labs(x = 'Gene', y='Expression Level',
title = 'lncRNA expression levels on chromosomes 1 & 2') +
coord_flip()
Session Info
## R version 4.6.1 (2026-06-24)
## Platform: x86_64-pc-linux-gnu
## Running under: Ubuntu 26.04.1 LTS
##
## Matrix products: default
## BLAS: /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3
## LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/libopenblasp-r0.3.32.so; LAPACK version 3.12.0
##
## locale:
## [1] LC_CTYPE=en_US.UTF-8 LC_NUMERIC=C
## [3] LC_TIME=en_US.UTF-8 LC_COLLATE=en_US.UTF-8
## [5] LC_MONETARY=en_US.UTF-8 LC_MESSAGES=en_US.UTF-8
## [7] LC_PAPER=en_US.UTF-8 LC_NAME=C
## [9] LC_ADDRESS=C LC_TELEPHONE=C
## [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C
##
## time zone: Etc/UTC
## tzcode source: system (glibc)
##
## attached base packages:
## [1] stats4 stats graphics grDevices utils datasets methods
## [8] base
##
## other attached packages:
## [1] SummarizedExperiment_1.43.0 Biobase_2.73.2
## [3] GenomicRanges_1.65.4 Seqinfo_1.3.2
## [5] IRanges_2.47.5 S4Vectors_0.51.10
## [7] BiocGenerics_0.59.12 generics_0.1.4
## [9] MatrixGenerics_1.25.0 matrixStats_1.5.0
## [11] zen4R_0.10.6 tidyr_1.3.2
## [13] tibble_3.3.1 dplyr_1.2.1
## [15] ggplot2_4.0.3 AnnotatedBCGEData_0.99.13
## [17] BiocStyle_2.41.0
##
## loaded via a namespace (and not attached):
## [1] tidyselect_1.2.1 farver_2.1.2 blob_1.3.0
## [4] filelock_1.0.3 S7_0.2.2 fastmap_1.2.0
## [7] BiocFileCache_3.3.0 XML_3.99-0.25 digest_0.6.39
## [10] lifecycle_1.0.5 RSQLite_3.53.3 magrittr_2.0.5
## [13] compiler_4.6.1 rlang_1.3.0 sass_0.4.10
## [16] tools_4.6.1 utf8_1.2.6 yaml_2.3.12
## [19] knitr_1.52 labeling_0.4.3 S4Arrays_1.13.2
## [22] bit_4.6.0 curl_8.1.0 DelayedArray_0.39.8
## [25] plyr_1.8.9 xml2_1.6.0 RColorBrewer_1.1-3
## [28] abind_1.4-8 withr_3.0.3 purrr_1.2.2
## [31] sys_3.4.3 grid_4.6.1 scales_1.4.0
## [34] cli_3.6.6 crayon_1.5.3 rmarkdown_2.32
## [37] otel_0.2.0 httr_1.4.9 tzdb_0.5.0
## [40] readxl_1.6.0 DBI_1.3.0 cachem_1.1.0
## [43] stringr_1.6.0 parallel_4.6.1 BiocManager_1.30.27
## [46] cellranger_1.1.0 XVector_0.53.0 vctrs_0.7.3
## [49] Matrix_1.7-6 jsonlite_2.0.0 hms_1.1.4
## [52] bit64_4.8.6 maketools_1.3.2 jquerylib_0.1.4
## [55] keyring_1.4.1 glue_1.8.1 stringi_1.8.9
## [58] gtable_0.3.6 pillar_1.11.1 htmltools_0.5.9
## [61] R6_2.6.1 dbplyr_2.6.0 httr2_1.3.0
## [64] vroom_1.7.1 evaluate_1.0.5 lattice_0.23-1
## [67] readr_2.2.0 memoise_2.0.1 bslib_0.12.0
## [70] Rcpp_1.1.2 SparseArray_1.13.4 xfun_0.62
## [73] buildtools_1.0.0 pkgconfig_2.0.3