I'm a bioinformatics engineer at a Seattle-based company.
Many of my past projects are on GitLab.
I'm interested in knowledge graphs, knowledge engineering, well-designed systems, FAIR data, functional and pretty data viz.
- Rust
- Clojure
R package for Data Integration & Visual Exploration
I'm a bioinformatics engineer at a Seattle-based company.
Many of my past projects are on GitLab.
I'm interested in knowledge graphs, knowledge engineering, well-designed systems, FAIR data, functional and pretty data viz.
Suggested by John on 6/19 -- should be a common use case.
I ran above command and came across this error:
Warning: Error in value[[3L]]: Couldn't normalize path in addResourcePath
, with arguments: prefix
= 'typedarray-0.1'; directoryPath
= '/home/avu/R/x86_64-pc-linux-gnu-library/3.4/plotly/htmlwidgets/lib/typedarray'
[No stack trace available]
Allow users to define subgroups among cases and generate comparison plots such as bar charts.
Allow customization of colorscales. A functional use would be when someone cares only about the magnitude but not sign of correlation; instead of a Red-Blue scale for [1,-1], use a colorscale with color/intensity based on magnitude.
A plotly issue referenced here: plotly/plotly.js#1516
This is needed for DIVE to become more generic and able to be submitted as a CRAN package.
I ran devtools::install_github("avucoh/DIVE"). All dependencies were successfully installed except data.table (see error below). After manually installing and loading data.table, I re-ran DIVE install and it worked fine.
ERROR: dependency 'data.table' is not available for package 'DIVE'
Configuration methods for generic/custom use
Data objects in data/ still need to be documented.
This could be an alternative view to the matrix, where instead of seeing data in (variable * variable), can have a selection of cases, so (case ID * variable) view.
The curator module allows adding/editing metadata.
I ran shiny::runApp(system.file(package = "DIVE"), display.mode = "normal"), which initially yielded the following error:
Error in loadNamespace(name) : there is no package called โDTโ
I installed and loaded the DT package, re-ran the command and ran into a different error:
Error: Couldn't normalize path in addResourcePath
, with arguments: prefix
= 'typedarray-0.1'; directoryPath
= '/home/avu/R/x86_64-pc-linux-gnu-library/3.4/plotly/htmlwidgets/lib/typedarray'
In the interactiveMatrix module, when variables in selected are saved to applied() reactiveVal, certain selections of a single variable triggers this error.
Histology feature data was removed from the early version of app because it was believed the method to extract features could be improved, and once achieved the "cleaner" data could then be re-imported.
Create new UI for the high-throughput integration module
From https://www.ncbi.nlm.nih.gov/gap/phegeni, download and extract gene lists for relevant phenotypes (T1D, blood glucose, etc.) to be included in the Quick List option.
The instructions, i.e. 'data req's', should state that files uploaded should be in text file format and not excel.
controls for high-throughput data clustering
Tried with data from SRP064464
after adding a dataset from GEO, if one tries to also upload data through the "upload my data" file input method, the available datasets menu selection is updated incorrectly
As one of the data tools, be able to bring interesting features from a high-throughput dataset into the clinical/phenotype collection. Also a way to allow comparison between two different types of high-throughput datasets.
Filter with cell/tissues in a given context, e.g. T cells in pancreas vs T cells in peripheral blood. Also, allow user to use filter to select all cells/tissues within a hierarchy.
Suggested by John on 6/19. Right now the same functionality exists through uploading a list of IDs. However, there are enough people who will only want to look up a handful at a time, and the current method is more for "batch" process.
Stems from rstudio/shiny#927
The numeric input is used for specifying % most variable genes/gene products to display. User can manually enter in a min or max beyond what's valid, which means the selection can be 0, giving an error in the plot display.
Looks like an editable config file might be necessary for setting up the app in custom cases.
Currently, the app directory is in the package, so once the package is installed the app can be served by shiny::runApp(system.file(package = "DIVE")
. The app uses the data that comes with the package; some of this data needs to be able to be updated from the curator module, but package data is usually updated by updating the package and it's not good practice to overwrite the data.
Relevant threads:
https://stackoverflow.com/questions/4018519/update-the-dataset-in-an-installed-package
https://stackoverflow.com/questions/14711277/modifying-r-package-data
So all data in a package needs to be copied to a repo where it data can be read/re-written. The data in package will serve as the last official version.
Allow users to pull certain types of data from GEO to visualize in comparison with their own or nPOD's
When the first high-throughput track is initialized with the default cdata selection "donor.type", the display gives message:
Error in seq_len: argument must be coercible to non-negative integer
[No stack trace available].
However, when an additional phenotype/clinical data is selected from cdata, error goes away.
The Methods table would store Method citation reference, annotations, details, etc., and would be provide relational lookup from the main Variables/Metadata table.
For interactiveMatrix, add row and col annotations for metadata, e.g. like https://ropensci.github.io/iheatmapr/.
Clicking on "get from original source" of the metadata table somehow opens the graph for some links.
When all modules are run in the same app, a navigation menu bug causes "sticky" highlighting of the selected tab page and being unable to go back.
Only one metadata selection is available and filters on both rows and columns (creating a symmetric matrix). This doesn't allow viewing the intersection of different metadata, which might be more useful. Users might also expect having separate metadata filters for rows and columns.
Catalog is viewable by variable, not cases.
Each dataset is shown in its own track/rows, with the left hand side displaying high-throughput data the right hand side displaying clinical/phenotype data. Because the tracks are independent, coloring of the same features are not consistent because colors are randomly assigned.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
๐ Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. ๐๐๐
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google โค๏ธ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.