search

broadinstitute / cmQTL

High-dimensional phenotyping to define the genetic basis of cellular morphology
BSD 3-Clause "New" or "Revised" License
6 stars 0 forks source link

Predict cell health readouts #53

Closed shntnu closed 4 years ago

shntnu commented 4 years ago

Being addressed here https://github.com/broadinstitute/cmQTL/pull/51

Discussion context: https://github.com/broadinstitute/cmQTL/issues/47#issuecomment-661617001

gwaybio commented 4 years ago

In https://github.com/broadinstitute/cmQTL/issues/47#issuecomment-687181206 @jatinarora-upmc asked:

@gwaygenomics @bethac07 @shntnu just following up on cell health readouts, was it feasible to align the features?

We did not embark down this path yet. Legacy CellProfiler datasets lose value if we are unable to align features between CellProfiler versions. We need to do this in a systematic way so that past and future datasets can be interoperable. The Cell Health dataset and LINCS dataset are two major examples.

The cmQTL dataset is another great example demonstrating the need for interoperable CellProfiler features. Because we are analyzing the CellProfiler features like we would individual genes, we need to make sure that the feature measurements retain biological signal across CellProfiler versions.

Another critical point (the other side of the same coin) is that to increase value of CellProfiler features, we need to make sure that we optimize their internal sensitivity within

  1. Software versions
  2. Wetlab protocols
shntnu commented 4 years ago

Let's reopen if we decide to pursue this again