tracking_function

[iziziz]

may be it wouldn't be very difficult to add specific modification for shared/overlap/find.clonotype function which will allow you to "track" repertoire through out several other repertoires? The difference from find.clonotype(d=c(a,b,c), target=a$CDR3.nucleotide.sequence, ...) is that the track.clonotype function will have two important possibilities: a - it will allow you to search for set of target vectors (several target sets of clonotypes) in several sets of repertoires b - it will be able to customize the output and get several columns from repertoire-to-track, not only CDR3.nuc.seq and V.seg

?

[vadimnazarov]

a) What do you mean by "search for a set of target vectors"? You can either merge this sets and use find.clonotypes or apply find.clonotypes to each clonotype vector.

b) The function need to get several columns like in find.clonotypes or something else?

[iziziz]

a) I mean, for example, the situation if I'm trying to track clonotypes in several timepoints. I have repertoire-to-track (CDR3.nuc.seq, V.seg, J.seg) and repertoires where I'll try to find clonotypes from repertoire-to-track. And further and harder :-) I have several (n) such sets: (repertoire-to-track + repertoires-to-scan)*n. Cause of the data might be quite big and I don't want to think - I would like to have special function for the task. Shared.repertoire is unuseful for this cause making "sharing" instead of "tracking". Now I'm using slightly modified find.clonotype for the task. But seems it will be user-friendly step to add smth like such modification as separate function.

b) guess like in find.clonotypes... I failed to imagine smth more than... But I'll try again )

[vadimnazarov]

Could you provide me an artificial example? How do you see this? Like "trackClonotypes(list_of_dfs, .track = c("Nuc seq", "V gene"), .return = c("reads, umis, J genes")) returns a list for each data frame in list_of_dfs with a data frame similar to find.clonotypes."