Table 2.

Summary of the four OpenSpliceAI model architectures, each trained with a distinct flanking sequence length (80, 400, 2,000, and 10,000 nucleotides). The table lists the kernel sizes (W), dilation rates (AR), number of residual and skip blocks, and total cropping length (CL).

Parameter	Flanking = 80	Flanking = 400	Flanking = 2000	Flanking = 10000
Kernel sizes (W)	[11, 11, 11, 11]	[11, 11, 11, 11, 11, 11, 11, 11]	[11, 11, 11, 11, 11, 11, 11, 11, 21, 21, 21, 21]	[11, 11, 11, 11, 11, 11, 11, 11, 21, 21, 21, 21, 41, 41, 41, 41]
Dilated rates (AR)	[1, 1, 1, 1]	[1, 1, 1, 1, 4, 4, ,4 ,4]	[1, 1, 1, 1, 4, 4, ,4 ,4, 10, 10, 10, 10]	[1, 1, 1, 1, 4, 4, ,4 ,4, 10, 10, 10, 10, 25, 25, 25, 25]
Residual blocks	4	8	12	16
Skip connections	1 (inserted after residual block 4)	2 (inserted after residual blocks 4 & 8)	3 (inserted after residual blocks 4, 8, & 12)	4 (inserted after residual blocks 4, 8, 12, & 16)
Cropping length (CL); $(\mathbf{2}\times \Sigma [\mathit{\text{AR}}\times (\mathbf{\text{W}}-\mathbf{1})\left]\right)$	80	400	2,000	10,000