An accurate and interpretable model for siRNA efficacy prediction

Table 6 Performance on independent datasets.

Dataset	Sparse	Spectral	Composite	BIOPREDsi
Harborth (44)	0.43	0.49	0.48	0.45
Reynolds (240)	0.54	0.49	0.55	0.55
Vickers (76)	0.58	0.54	0.49	0.57
Shabalina (653)	0.48	0.45	0.48	-

This table shows the performance of the LASSO model with different features (sparse-19, spectrum-19, composite-19/19) on four independent datasets, as well as the performance of the BIOPREDsi algorithm on the the first three datasets (from Huesken et al. [33]). The performance is measured as the Pearson correlation coefficient between predicted and true efficacy).