User Manual

splam

usage: splam [-h] [-v] [-c] {extract,score,clean} ...

splice junction predictor to improve alignment files (BAM / CRAM)

optional arguments:
-h, --help            show this help message and exit
-v, --version
-c, --citation

Commands:
{extract,score,clean}
   extract             Extracting all splice junctions from an alignment or annotation file
   score               Scoring all splice junctions
   clean               Cleaning up spurious splice alignment

splam extract

usage: splam extract [-h] [-V] [-F TYPES] [-P] [-n] [-f FILE_FORMAT]
                     [-d DATABASE] [-o DIR] [-M DIST] [-g GAP] [--fr] [--rf]
                     INPUT

positional arguments:
INPUT                 target alignment file in BAM format or annotation file in GFF format.

optional arguments:
-h, --help            show this help message and exit
-V, --verbose         running splam in verbose mode.
-F TYPES, --features TYPES
                        text file listing parent feature types from which to extract introns
-P, --paired          bundling alignments in "paired-end" mode.
-n, --write-junctions-only
                        writing out splice junction BED file only without other temporary files.
-f FILE_FORMAT, --file-format FILE_FORMAT
                        the file type for Splam to process: "BAM", "GFF", or "GTF"
-d DATABASE, --database DATABASE
                        path to an existing gffutils annotation database
-o DIR, --outdir DIR  output directory; the default is "tmp_out"
-M DIST, --max-splice DIST
                        maximum splice junction length
-g GAP, --bundle-gap GAP
                        minimum gap between bundles
--fr                    assume a stranded fr-secondstrand library
--rf                    assume a stranded rf-firststrand library

splam score

usage: splam score [-h] [-V] [-o DIR] [-b BATCH] [-d pytorch_dev] -G REF.fasta -m MODEL.pt junction_BED

positional arguments:
junction_BED          target splice junctions in bed files.

optional arguments:
-h, --help            show this help message and exit
-V, --verbose
-o DIR, --outdir DIR  the directory where the output file is written to. Default output filename is "junction_score.bed"
-b BATCH, --batch-size BATCH
                        the number of samples that will be propagated through the network. By default, the batch size is set to 10.
-d pytorch_dev, --device pytorch_dev
                        the computing device that is used to perform computations on tensors and execute operations in the PyTorch framework. By
                        default, this parameter is detected automatically.
-G REF.fasta, --reference-genome REF.fasta
                        the path to the reference genome.
-m MODEL.pt, --model MODEL.pt
                        the path to the Splam model

splam clean

usage: splam clean [-h] [-@ threads] [-t threshold] [-n bad intron num]
                   [-P] -o DIR

optional arguments:
-h, --help            show this help message and exit
-@ threads, --threads threads
                        set the number of sorting, compression, and merging threads; the default is 1
-t threshold, --threshold threshold
                        cutoff for identifying spurious splice junctions; the default is 0.1
-n bad intron num, --bad-intron-num bad intron num
                        number of spurious introns used to classify a transcript as bad; the default is 8
-P, --paired          clean an alignment file in paired-end mode
-o DIR, --outdir DIR  required directory containing the extract and score outputs