Table 2.

Working principles, aligners, prior separately running tools, and running platform of reviewed tools.

Tool	Working Principle	Running Platform	Prior Separately Running Tools	Aligner
					Alt Event Finder	Combines read alignment tool and transcript isoform reconstruction tool to produce transcript isoform annotations.	Linux	BFAST/TopHat + Cufflinks/Scripture	BFAST/ TopHat
SpliceMap	Pins down one end of the SJ by matching the read spanning a junction longer than its half-length and then uses the matched part as a seeding to locate the other end.	Linux	no need	Bowtie/ Eland/ Seq Map
FineSplice	Uses a semi-supervised anomaly detection approach of a logistic regression model to estimate the SJs from the TopHat2 alignment results.	Linux	TopHat2	TopHat2
RSW	Uses Bowtie to do initial read alignment, splits IUM reads into halves, and then uses Bowtie to align the split read halves back to the reference genome.	Linux	no need	Bowtie
RSR	Similar to RSW, but improved via a linear regression equation in the Generalized Linear Model.	Linux	no need	Bowtie
PASTA	Uses heuristic patterned alignments and a logistic regression statistical model to detect exon-intron junctions.	Linux	no need	Bowtie
rMATS	Utilizes the hierarchical framework to simultaneously calculate variability among replicates and estimate uncertainty of isoform fraction within individual replicates.	Linux	no need	STAR
SOAPsplice	Reports SJ candidates from both spliced and intact alignment and filters out false positives	Linux	no need	BWT
SplicePie	Detects AS events by capturing pre-mRNAs at different splicing stages: pre-, intermediate-, and post-splicing stages.	Linux	GSNAP + SAMtools	GSNAP
SplicingCompass	Calculates geometric angles between the multiple- dimension exon read counts vectors to detect differential AS events.	R package	TopHat + BEDTools	TopHat
TopHat	Mapped regions are computed to yield initial consensus. Initially unmapped reads are indexed and mapped to the potential SJs.	Linux	no need	Bowtie/ Bowtie2
TrueSight	Forms a unified model that uses adaptive training of iterative logistic regression to identify novel SJs and rule out unreliable SJs.	Linux	no need	Bowtie
NSMAP	Identifies the structures of expressed isoforms and estimates the expression levels of known and novel expressed isoforms at the same time	MatLab package	TopHat	TopHat
rSeqDiff	Uses an extended linear Poisson model to identify differential isoform expressions in multiple RNA-seq samples	R package	rSeq	SeqMap/ Eland/ BWA/ Bowtie/ bowtie2
rSeqNP	Executes a non-parametric approach to test the differential expression (DE) and differential splicing (DS) using RNA-seq data.	R package	rSeq/RSEM	SeqMap/ Eland/ BWA/ Bowtie/ bowtie2