deplanckelab / asap Goto Github PK

It would be cool to be able to create a new visualization from any metadata?
For e.g., I have a public project, which already has a UMAP. I've re-created the analysis on ASAP (https://asap.epfl.ch/projects/8jaw4y) from the count table, and now, I want to be able to display both my UMAP, and the original UMAP (Metadata X and Y) from the metadata?

In the list of public/private projects, it would be useful to know the pipeline used (bulk or single-cell)

Could we make this list more appealing visually? Similar to https://cellxgene.cziscience.com/datasets maybe?
For e.g. adding horizontal lines, having buttons with different colors/looks/icons?
See the name of the study (in blue, small, like in CellxGENE) when available?

Adding filters, to focus on certain organisms for e.g.? A bit like we did in scFAIR?

Hello

I work on melanoma single cell RNAseq data (5k)

I think that ASAP is the greatest application, but slow analysis is shortcoming on the web

Can I use ASAP on my own computer?

This is slowly becoming the central panel for most projects.
We need to enhance it. As currently it's not easy enough to color by gene, metadta, etc....

Probably we could inspire the view to the one done in CELLxGENE (the best current UI I would say)

When we validate to run a method, it stays on the wheel for quite a while, before the card is generated.

We should update the pipeline to make it even more bulk-specific:

• Remove the HVG step, because it's a single-cell-only step
• Scaling: The scaling method is called "Seurat", but Seurat is a single-cell package. So maybe we should rename the method like "Simple scaling" or something similar? Only for the bulk pipeline ofc.
• In Normalization, can we remove the Log2 method? Also could we add back the DESeq2 method?
• In Gene Filtering, can we remove the HVG, Highly Variable Genes, and M3Drop methods? Again, they are single-cell-specific.
• In every card and method, it's written "cells: 40" "genes:52030", but it's not cells for bulk, it should be "samples".
  • In the cards
  • In the Metadata section. Including when we "Add metadata"
  • In the Visualization (it's called Cell scatter plot)

Some projects get stuck in the "Archived / Unarchiving" state, as reported by different users:

"Projects seem stuck with the “Unarchiving project...” message for hours. On the projects page, from the status, it has the symbol of unarchived but it is written “Unarchiving”.

This happens somewhat randomly so we need to understand the cause for fixing it.

We cannot annotate the Category, cluster 2 in https://asap.epfl.ch/projects/8jaw4y

Dans Visualization / Annotation, la vue avec les marker genes et les annotation devraient être compressée.

Par example dans
https://asap.epfl.ch/projects/8jaw4y

Je vois

On pourrait mettre Odd+ Tin- par example pour les up- et down- regulated?

We need to:

• Test the existing model on the server
• Create the UI to use it (probably in the Visualization pane? Maybe as a separate task?
• Adapt it so that we can train model on new datasets (Admin only) and create models to be used for other species
• Version it so that we can re-train on bigger datasets or new annotations of datasets

Hi,

I do not see any activity on ASAP development for the past 17 months and the asap website is not updated recently as well. Is this tool still maintained by Deplancke group?

Thanks.