shimbalama / omics_dashboard Goto Github PK

just got an error when we were trying to toggle between drugs:
⛑️
Failed component prop type: Invalid component prop figure key props supplied to Graph. Bad object: { "props": { "children": "No data selected" }, "type": "Div", "namespace": "dash_html_components" } Valid keys: [ "data", "layout", "frames" ]
4:48:22 PM
Failed component prop type: Invalid component prop figure key props supplied to Graph.
Bad object: {
"props": {
"children": "No data selected"
},
"type": "Div",
"namespace": "dash_html_components"
}
Valid keys: [
"data",
"layout",
"frames"
]

username password login

Ability to export data (as a .csv)

CTRL is fine
CoVB - change to SARS-CoV-2 with DMSO
CoVA - change to SARS-CoV-2 with INCB054329

Page randomly refreshes/resets while navigating

read me etc

Some datasets ie prot have a lot of missing data, nice to remove it easily

only seems to work from phospho

doc strs etc

fit on mobile

Hi Liam
A couple of suggestions for the RNAseq data:

• For volcano plots (which look awesome!) is it possible to have a numerical value (possibly hover) for the p value slider? it will make it easier to see what the parameters are set to

• 'Effect size' might need to be specified. Is it log2 fold change?

• it would also be really helpful to add descriptors for all experiments/datasets (this was planned I think?)

Thank you!!

test conditions match those in box plot and persist

post init to check that indexes are equal. found examples of missign and duplicates gene names

Requirements:

• take new dataset
• determine type
• create required folder structure
• check col names and rename as required (ie POR)
• data validation/pydantic/pandera
• user friendly (GUI?)

Hi Liam

For the functional data, a few suggestions:

• is it possible to select an experiment to see which treatments/drugs were tested? at the moment I think the gate keeper for drop down menus is 'Drug' (this could also be changed to treatment)

• if possible, would also be good to have units on the x axes (ie concentration in nM uM or mM)

• just got an error when we were trying to toggle between drugs:
  ⛑️
  Failed component prop type: Invalid component prop figure key props supplied to Graph. Bad object: { "props": { "children": "No data selected" }, "type": "Div", "namespace": "dash_html_components" } Valid keys: [ "data", "layout", "frames" ]
  4:48:22 PM
  Failed component prop type: Invalid component prop figure key props supplied to Graph.
  Bad object: {
  "props": {
  "children": "No data selected"
  },
  "type": "Div",
  "namespace": "dash_html_components"
  }
  Valid keys: [
  "data",
  "layout",
  "frames"
  ]

Based on discussion in lab meeting I think this is required

box plot header in phospho can get lost

Hi Liam
Harley and I are looking onto phospho and proteomics datasets and have suggestions. Not sure how practical/easy/impossible to do... happy to chat further!
ISSUES

• p values values show up as zero in proteomics data (need to be removed)

• stats bars at top need to be removed from plots (they are confusing...and all zeros anyway)

• y axes for proteomics should be 'Protein abundance'

• x axes for proteomics should be 'Condition'

• y axes for phosphoproteomics should be 'Phosphopeptide abundance'

• x axes for phosphoproteomics should be 'Phosphopeptide'

• Legend label for phosphoproteomics should be 'Condition'

• we tried titin (TTN) knowing that it has heaps of phosphosites - it melted the dashboard. oops. not sure what to suggest here? maybe a maximum entry cutoff? (the dashboard is not loading now...)

SUGGESTIONS:

• could we toggle between abundance values and log10(abundance) values for both proteomics and phospho?

in the phospho plot it would be good to be able to toggle prot poses on and off like test

remove hard coded div ids

Ie look at gene activity in simultaneous plots

based on lab meeting discussion, only simple datasets with 1 POR with show any stats

we tried titin (TTN) knowing that it has heaps of phosphosites - it melted the dashboard. oops. not sure what to suggest here? maybe a maximum entry cutoff? (the dashboard is not loading now...)

Wont change the plot when choosing a different drug for the functional data

For the quaife ryan bulk 2017 rnaseq it needs to be broken up into the different cell types.

For the mills bulk 2017 rnaseq the 2d cells need to be broken into the 20 day and the 1 y rather than a combo

Cheers

James

Hi - just querying this dataset. We have BENOs and mgBENOs but no storm/control condition.
Thanks.

I found the landing page...but it was buried under scRNA seq... should this be shown upon loading?

load data on navigation not all at start

Note sure what format you need, see email

remove units from all metrics (in file)