Run clusterExpressionPatterns based on conditions?

HectorRDB / condiments

Trajectory inference across multiple conditions with condiments: differential topology, progression, differentiation, and expression

https://hectorrdb.github.io/condiments/

Other

16 stars 1 forks source link

Run clusterExpressionPatterns based on conditions? #12

Closed ghost closed 2 years ago

ghost commented 2 years ago

Hi Hector,

As far as I know, clusterExpressionPatterns clusters genes based on different lineages. Is there a way to run it based on consitions? I have a dataset with one lineage and two conditions. This is maybe more about tradeSeq but since I'm trying to do TI across conditions, I wonder if you know how to do it.

Thanks!

HectorRDB commented 2 years ago

Tagging @koenvandenberge here to check but:

Indeed, tradeSeq has no implementation of clusterExperiment for conditions, however
Lineages and conditions are equivalent under the hood so you use fitGam, inverting the conditions and lineages and then using clusterExperiment. Hope that helps, Hector

koenvandenberge commented 2 years ago

Yes, I agree. clusterExpressionPatterns is just a wrapper around(a) first calling predictSmooth to get the fitted values; (b) then clustering the genes' fitted values using clusterExperiment. You can still use predictSmooth to get the fitted values across multiple conditions, and then cluster genes using your preferred clustering algorithm.

In our workshop, we do something similar, clustering the genes using hierarchical clustering. However, note that there we only use the fitted values from a single condition to cluster the genes.

HectorRDB commented 2 years ago

Closing due to lack of activity. Feel free to reopen if issue persists.