Fig. 6From: NMFP: a non-negative matrix factorization based preselection method to increase accuracy of identifying mRNA isoforms from RNA-seq dataNMFP increases the robustness of SLIDE to the choices of regularization parameter λ, from 0.001 to 0.15 by 0.001. F scores of NMFP + SLIDE (orange) and SLIDE (blue) were calculated at 150 values of λ used in the Lasso regression in SLIDE. NMFP used an NMF rank =3. RNA-seq data were simulated from mouse gene ENSMUSG00000025940, which has 7 subexons and 4 annotated isoformsBack to article page