{FlowRepositoryQuery} is an R package providing comprehensive data on all public experiments from the FlowRepository database, updated as of August 1, 2024. It also includes a dedicated Shiny application for filtering experiments based on their marker panels, making it easier to explore relevant datasets.
You can directly use the web tool through the following Shiny Server link:
To install the latest version of the {FlowRepositoryQuery} package
from GitHub, use the following command:
remotes::install_github("i-cyto/FlowRepositoryQuery", ref="main")Once installed, you can start the Shiny application by running:
FlowRepositoryQuery::run_app()
The current tool aims to make it easier to search FlowRepository's public experiments by panel markers. Filtering the panel column is not easy without a regular expression because the names of the markers are not really standardized by humans.
CD3 is not CD33, even though CD33 contains CD3. You therefore need to search for words rather than strings of characters, as shown in the illustration below.
Top match examples with CD3
You can enter several markers at once, and they will all be present in the resulting selection of experiments.
Results of a search for several markers
PD-L1 has many spellings, as does PD1. Therefore, the search must include wildcards to match separators such as '-', '_' or even '.'. This is the role of '.', which symbolizes the wildcard character.
For example, the query 'PD.L1' will find 'PD-L1', 'PD_L1', but also 'PDXL1'. The latter is not relevant but could be easily removed. The most important thing is probably not to miss any writing. If the possible characters are known, they can be listed, for example, 'PD[-_.]L1' for 'PD-L1', 'PD_L1', or 'PD.L1'.
Postfixing a character or a sequence with '*' indicates that it may or may not be present. For example, 'PD[-_.]*L1' will also match 'PDL1'. You probably know enough to enjoy regular expressions for finding markers.
Examples of PD-L1 matching
FlowRepository is an important resource for the community. It contains a large number of public experiments that could be useful for improving analytical pipelines, as external experiments to validate results, or as a source of hypotheses.
To improve analytical pipelines, I'd like to identify experiments and add tags to announce when an experiment could be a standard for developing or validating one of the steps in the pipeline. Unfortunately, it's very difficult to browse and filter the information presented on the FR public page.
I have therefore analyzed the FR public pages, collected information, and compiled it into a table. This table is available in the form of a Google sheet, which could make it easier to access the information. There is a ‘dictionary’ tab to explain the columns reported and their origin.
There is an interesting column entitled "Design". It summarizes the experimental design of an experiment when it is provided by the researcher. It lists each factor (sample types, time points, tissues, etc.), its values, and the number of SCFs per value. This will help me a lot in selecting a clearly defined design. Unfortunately, I know of experiments whose design has not been annotated in FR, which hinders the reuse of experiments and, perhaps, the development of FR.
The experiment metadata was collected on 2024-08-01. There are 2133 experiments, about 413k FCS files for a total volume of about 5 TB.
If you would like to suggest a feature for the query tool, please open a question on the GitHub repository for this package.
For further information, you can contact Samuel Granjeaud at samuel.granjeaud@inserm.fr.
Source code is available on GitHub.
We thank FlowRepository for making the data available to the community. We thank ISAC for supporting FlowRepository.
You are reading the doc about version : 0.1.0
This README has been compiled on the
Sys.time()
#> [1] "2024-10-21 15:36:23 CEST"Here are the tests results and package coverage:
devtools::check(quiet = TRUE)
#> ℹ Loading FlowRepositoryQuery
#> ── R CMD check results ────────────────────────── FlowRepositoryQuery 0.1.0 ────
#> Duration: 22.8s
#>
#> ❯ checking installed package size ... NOTE
#> installed size is 5.2Mb
#> sub-directories of 1Mb or more:
#> data 4.1Mb
#>
#> 0 errors ✔ | 0 warnings ✔ | 1 note ✖