TABLE 1.

Current IR detection tool.

Tool name	Abbreviation	Detection method	Advantages	Limitation	References
Intron Retention call	IRcall	Uses ranking strategy to calculate IR score	- Reduce false positive results	- It depends on the quality of the used alignment tool to collect data	Bai et al. (2015)
Intron Retention classifier	IRclassifier	Uses machine learning technology to build up random forest to detect IR events	- High precision
			- Identification of both known and novel IR events
Intron Retention Finder	IRFinder	It detects IR events using IR ratio via measuring the intronic abundance and splicing level	fast and sensitive detection	Possible overlapping between introns and exons from other transcripts	Middleton et al. (2017)
			High Accuracy and precision	Calculating IR based on junction reads not on the expression level of intron
			Free available database for over 2000 IR human samples	Multiple position reads from the genome produces noise in the results
			Efficient detecting of low coverage. IR events
Intron REtention Analysis and Detector	iREAD	Employs the entropy score to determine the distribution of intronic reads across the intron region	Limited exon-intron overlapping during read	It has no differential analysis	Li et al. (2020)
			Analyze both splice junction reads and intron expression level
			Flexible running operating system
			Sensitive
Keep Me Around	KMA	R packaging tool to quantify IR in RNA data	Reduced false positive results by combining replicates	The IR analysis and quantification are performed in different software	Pimentel et al. (2015)
				The common feature of retained intron which is flat distribution is not identified