Repository navigation
Modify quant - #266
Modify quant#266
Conversation
…genes and update annotation and quantification accordingly
jonathangoeke
left a comment
jonathangoeke
left a comment
There was a problem hiding this comment.
quantificaiton always returns 0 for this branch for all samples, not sure why (but is an outdated branch in any case). Main comments:
- change equiRCs and txIds and readClassIds to integer and integerlist
- IncompatibleCounts could be incompatibleCounts
@cying111
| return(se)}) | ||
| } | ||
| # combine incompatible counts | ||
| IncompatibleCounts = Reduce(merge_wrapper, lapply(countsSe, FUN = function(se){metadata(se)$IncompatibleCounts})) |
There was a problem hiding this comment.
to be consistent, we should name this incompatibleCounts (camelCase)
| min.primarySecondaryDist = min.primarySecondaryDist, | ||
| min.primarySecondaryDistStartEnd = min.primarySecondaryDistStartEnd, | ||
| verbose = verbose) | ||
| readClassMod <- modifyIncompatibleAssignment(readClassDist) |
There was a problem hiding this comment.
this function only works with the distTable, maybe better like this?
readClassDist$distTable <- modifyIncompatibleAssignment(readClassDist$distTable )
then we don't have 2 copies of the readClass object?
| min.primarySecondaryDistStartEnd = min.primarySecondaryDistStartEnd, | ||
| verbose = verbose) | ||
| readClassMod <- modifyIncompatibleAssignment(readClassDist) | ||
| annotationsUpdated <- updateAnnotations(readClassMod, annotations, verbose) |
There was a problem hiding this comment.
this function assigns read classes evenly to genes. Ideally this is done based on the overall gene count (after quantification?) can probably be ignored though as this is a very special case of unassigned reads overlapping multiple genes?
| distTable[order(readClassId,GENEID,tx_id), | ||
| read_class_id := paste(unique(tx_id), collapse = "."), | ||
| by = list(readClassId, GENEID)] |
There was a problem hiding this comment.
this is ideally done using txid (integer) and equiRCByTxId (integerlist) which is much faster and memory efficient
There was a problem hiding this comment.
(which would the definition of equiRCs throughout quantification)
There was a problem hiding this comment.
for the unidentified this could be tracked with an additional column of logicals
This PR includes changes: