Fig. 6: Enhancing ICI prediction by incorporating tumor subclonal architecture. | Nature Communications

Fig. 6: Enhancing ICI prediction by incorporating tumor subclonal architecture.

From: NeoPrecis: enhancing immunotherapy response prediction through integration of qualified immunogenicity and clonality-aware neoantigen landscapes

Fig. 6: Enhancing ICI prediction by incorporating tumor subclonal architecture.The alt text for this image may have been generated using AI.

a Computation of NeoPrecis-LandscapeClone. Immunogenicity predictions from NeoPrecis-Immuno or NeoPrecis-Integrated are summed within each mutation cluster. These cluster sums are then combined using a weighted average, where the prevalence of each cluster serves as the weight. This approach generates a tumor-centric immunogenicity score by integrating multiple mutation-centric predictions. b Distribution of mutation number, Gini index (sGini), and prevalence Gini index (pGini). ICI response (positive or negative) is annotated with the mark shape. c AUROC comparison of TMB, CSiN, ioTNL, NP-LandscapeSum, NP-LandscapeCCF, and NP-LandscapeClone across two patient groups, heterogeneous (low sGini–low pGini) and homogeneous (high sGini–high pGini), in melanoma and NSCLC. NP denotes NeoPrecis. d Density plot estimated using kernel density estimation (KDE) showing the distribution of %ClonalMuts, the ratio of clonal mutations (CCF ≥ 0.85) to total mutations, across different smoking statuses. Values are clipped between 0 and 1. e Density plot estimated using KDE shows the distribution of %Binding-I, the ratio of MHC-I binding mutations (PHBR ≤ 2) to total mutations, across different smoking statuses. Values are clipped between 0 and 1.

Back to article page