trainBambu: Function to train a model for use on other data

In GoekeLab/bambu: Context-Aware Transcript Quantification from Long Read RNA-Seq data

Function to train a model for use on other data

Description

This function train a model for use on other data

Usage

trainBambu(
  rcFile = NULL,
  min.readCount = 2,
  nrounds = 50,
  NDR.threshold = 0.1,
  verbose = TRUE
)

Arguments

rcFile	summerized experiment object with read classes/ranges produced from bambu(quant = FALSE, discovery = FALSE) or rcOutdir
min.readCount	the minimum number of reads a read class is required to be have to be used for training
nrounds	xgboost hyperparameter - the number of decision trees in the final mode
NDR.threshold	the effective NDR threshold that bambu will try and match on other samples when using this model
verbose	if additional messages should be output Output - A list containing 6 objects which is passed directly into bambu(opt.discovery=list(defaultModels=trainBambu())) transcriptModelME - the model for multi-exon transcripts transcriptModelSE - the model for single-exon transcripts txScoreBaseline - the txScore used for NDR calibration for multi-exon transcripts txScoreBaselineSE - [DEPRECATED] the txScore used for NDR calibration for single-exon transcripts lmNDR = lmNDR - the linear model of the reletionship between txScore and NDR used to calculate the baseline for multi-exon transcripts lmNDR.SE = lmNDR.SE - the linear model of the reletionship between txScore and NDR used to calculate the baseline for single-exon transcripts NDR.threshold - the NDR threshold usd to calculate the txScoreBaseline on the lmNDR (baselineFDR)

Details

Function to train a model for use on other data

Value

It returns a model object to use in bambu

GoekeLab/bambu documentation built on April 6, 2024, 10:36 p.m.