search

selkamand / sigstash

Easily load published mutational signatures into R
Other
0 stars 0 forks source link

Add a unique code linked directly to the matrix itself (sigminer and cosmic versions) #40

Closed selkamand closed 2 months ago

selkamand commented 2 months ago

So when we pass around signature collections to different tools, it would be nice to have functions to automatically map which collection we're using based on the md5 of the serialised collection.

The digest R package will let us do this pretty easily. digest::digest(algo = "md5")

We should expose a function:

sig_identify_collection(signatures)

Where we input the collection and return a unique name / description of the signature collection (including whether its the sigminer or sigverse/cosmic version)

selkamand commented 2 months ago

We should also explore adding the signature dataset name / format as an attribute to the tibbles we load e.g. 

attr(signatures, "dataset_name") <- dataset_name
attr(signatures, "format") <- "sigminer" / "sigstash" / "tidy"

Which we can look up with attributes() when logging