Fig. 3

Construction of prognostic signatures based on LASSO COX analysis. Each curve in the figure represents the trajectory of each independent variable coefficient. The vertical axis is the value of the coefficient, the lower horizontal axis is log2-Lambda value, and the upper horizontal axis is the number of non-zero coefficients in the model on each scale. The small serial numbers before each curve in the box were used to mark each variable. Each colored line represents the value taken by a different coefficient in the model. Lambda is the weight given to the regularization term, so as lambda approaches zero, the loss function of the model approaches the OLS loss function. When lambda is very small (leftmost), the LASSO solution should be very close to the OLS solution, and all coefficients are in the model. As lambda grows (from left to right), the regularization term has greater effect, more and more coefficients will be zero valued and fewer variables in the model remain. a alternate acceptor, b alternate donor sites, c alternate promoters, d alternate terminators, e exon skips, f mutually exclusive exons, g retained introns, and h comprehensive signature