SGSeq (version 1.0.6)

predictTxFeatures: Transcript feature prediction from BAM files

Description

Transcript features are predicted for each sample and subsequently merged across samples. Terminal exons that share a splice site with internal exons are filtered. For details see the help pages for predictTxFeaturesPerSample, mergeTxFeatures and filterTerminalExons.

Usage

predictTxFeatures(sample_info, which = NULL, alpha = 2, psi = 0, beta = 0.2, gamma = 0.2, min_junction_count = NULL, min_n_sample = 1, min_overhang = NA, cores_per_sample = 1, BPPARAM = MulticoreParam(1))

Arguments

sample_info

data.frame with sample information. Required columns are “sample_name”, “file_bam”, “paired_end”, “read_length”, “frag_length” and “lib_size”. Alignment information can be obtained with function getBamInfo.

which

GRanges of genomic regions to be considered for feature prediction, passed to ScanBamParam

alpha

Minimum FPKM required for a splice junction to be included. Internally, FPKMs are converted to counts, requiring arguments read_length, frag_length and lib_size. alpha is ignored if argument min_junction_count is specified.

psi

Minimum splice frequency required for a splice junction to be included

beta

Minimum relative coverage required for an internal exon to be included

gamma

Minimum relative coverage required for a terminal exon to be included

min_junction_count

Minimum fragment count required for a splice junction to be included. If specified, argument alpha is ignored.

min_n_sample

Minimum number of samples a feature must be observed in to be included

min_overhang

After merging, all or a subset of terminal exons sharing a splice site with an internal exon are removed. min_overhang specifies the minimum overhang required for a terminal exon to be included. Use NA to remove all terminal exons sharing a splice site with an internal exon. Use NULL to disable filtering (not usually recommended, as this will result in terminal exon predictions for each splice site; disabling filtering is useful if results are subsequently merged with other predictions and filtering is postponed until after the merging step).

cores_per_sample

Number of cores per sample

BPPARAM

BiocParallelParam for processing samples in parallel, defaults to MulticoreParam(1)

Value

A TxFeatures object

Examples

dir <- system.file("extdata", package = "SGSeq")
si$file_bam <- file.path(dir, "bams", si$file_bam)
txf <- predictTxFeatures(si, gr)