Extended Data Fig. 9: Evaluation of clone fitness model predictions.
From: Neoantigen quality predicts immunoediting in survivors of pancreatic cancer
The log-likelihood score (Supplementary Methods, eq. (31)) is shown for the STS and LTS cohorts to estimate the statistical information gain of fitness models and the amount of evidence of the selective pressures captured by each of the models. The orange bars show the aggregated log-likelihood scores, \(\Delta {{\mathscr{L}}}^{\text{STS}}\left(F,{F}_{N}\right)\) and \(\Delta {{\mathscr{L}}}^{\text{LTS}}\left(F,{F}_{N}\right),\) of the two-component fitness model, \(F\), with parameters \({\sigma }_{I},{\sigma }_{P}\) optimized for each recurrent tumour sample, as compared to the null model, \({F}_{N}\), standing for neutral clone evolution, with zero fitness and parameters \({\sigma }_{I}=0,{\sigma }_{P}=0\). The red bars present the corresponding aggregated log-likelihood scores \(\Delta {{\mathscr{L}}}^{\text{STS}}\left({F}_{P},{F}_{N}\right)\) and \(\Delta {{\mathscr{L}}}^{\text{LTS}}\left({F}_{P},{F}_{N}\right)\) for the driver-gene only fitness model, \({F}_{P}\), which accounts for positive selection on driver genes but disregards the effect of immune selection, with parameter \({\sigma }_{I}=0,\) and \({\sigma }_{P}\) optimized for each recurrent tumour sample. Finally, the blue bars present the corresponding aggregated log-likelihood scores \(\Delta {{\mathscr{L}}}^{\text{STS}}\left({F}_{I},{F}_{N}\right)\) and \(\Delta {{\mathscr{L}}}^{\text{LTS}}\left({F}_{I},{F}_{N}\right)\) for the immune-only fitness model, \({F}_{I}\), with parameter \({\sigma }_{P}=0,\) and \({\sigma }_{I}\) optimized for each recurrent tumour sample.
