Table 3.

Summary of features investigated in this study

Feature	Type	Description
Distance to nearest splice site	SNP-based	Distance between a given variant and the nearest 5′ or 3′ splice site in the target exon.
ESR change	SNP-based	Change in the frequency of ESR elements subsequent to a single base substitution. This includes:
		ESE to neutral (ESE loss)
		ESE to ESE (no change)
		Neutral to ESE (ESE gain)
		ESE to ESS (ESE loss and ESS gain)
		Neutral to neutral (no change)
		ESS to ESS
		Neutral to ESS (ESS gain)
		ESS to neutral (ESS loss)
		ESS to ESE (ESS loss and ESE gain)
In ESE	SNP-based	Frequency of ESE binding sites (in the wild-type) that overlap with the location of the variant
In ESS	SNP-based	Frequency of ESS binding sites (in the wild-type) that overlap with the variant
ESR hexamer score (ESR-HS)	SNP-based	Hexamer scoring function to express the relationship between disease and neutral variants and their differential distributions with respect to loss or gain of an ESE or ESS
Spectrum kernel	SNP-based	Frequency of 3-mers and 4-mers over an 11 bp window (wild type and mutant)
Change in natural splice site strength	SNP-based	MaxEnt splice site score of natural splice site in mutant allele minus MaxEnt splice site score of wild-type allele
Maximum cryptic splice site	SNP-based	Maximum cryptic splice site (5′ and 3′) score (outside of the natural splice site) found overlapping the variant on the mutant allele
Evolutionarily conserved element	SNP-based	PhastCons conserved element probability for substitution site, based on multiple alignments of 46 placental mammals
Base-wise evolutionary conservation	SNP-based	PhyloP base-wise sequence conservation score at site of single base substitution based on multiple sequence alignment of 46 placental mammals
Natural wild-type splice site strength	Exon-based	MaxEntScan score of the natural 5′ and 3′ splice site of the wild-type target exon
Flanking intron size	Exon-based	Length in base-pairs of the upstream and downstream introns flanking the target exon
Intronic ESS density	Exon-based	Intronic ESS density was calculated for 100 bp upstream and 100 bp downstream of the target exon
Exonic ESS density	Exon-based	ESS density was calculated across the first 50 bp and the last 50 bp of the target exon. If the length of the exon was less than 100 bp, then the full length of the exon was used to calculate the ESS density
Exonic ESE density	Exon-based	Same as above but for ESEs
Internal coding exon	Exon-based	{true, false}, Is the target exon an internal coding exon (that is, the target exon is not the first or last coding exon)
Exonic GC content	Exon-based	Percentage of nucleotides that are either guanine or cytosine in the target exon
Exon size	Exon-based	Size of the target exon
Constitutive exon	Exon-based	Is the target exon constitutively spliced
Exon number	Gene-based	Number of exons in the transcript
Transcript number	Gene-based	Number of different reported isoforms that the target gene encodes