Most proteins exert their functions in complex with other interactors. Single mutations can exhibit a profound impact on perturbing protein interactions, leading to human disease. However, predicting the effect of single mutations on protein interactions remains a major computational challenge. Deep learning, particularly protein language models or transformers, has become an effective tool in bioinformatics for protein structure prediction. However, the functional divergence of mutations makes it difficult to predict their interaction perturbation profiles. To address this fundamental challenge, we present eSIG-Net (edgetic mutation Sequence-based Interaction Grammar Network), a novel sequence-based "Interaction Language Model" for predicting protein interaction alterations caused by single mutations. eSIG-Net combines various protein sequence embeddings, introduces a mutation-encoding module with syntax and evolutionary insights, and employs contrastive learning to evaluate mutation-induced interaction changes. eSIG-Net significantly outperforms current state-of-the-art sequence-based and structure-based prediction methods at predicting mutational impact on protein interactions. We highlight examples where eSIG-Net nominates causal variants with high confidence and elucidates their functional role under relevant biological contexts. Together, eSIG-Net is a first-in-kind "interaction language model" that can accurately predict interaction-specific rewiring by single mutations with only sequence information, and exhibits generalizability across biological contexts.

Supplementary Figure S5. Overall survival in patients with or without baseline liver metastases.

Supplementary Figure S2. Progression-free survival in the randomized cohorts.

Supplementary Figure S3. Overall survival in the randomized cohorts.

<p>Supplementary Figures 1-10. Figure-S1. Immune score prediction of response to PD-1 immune checkpoint blockade; Figure-S2. Immune mutation burden in cancer; Figure-S3. Transcriptome alteration of immunome in cancer; Figure-S4. Enrichment of differentially expressed genes (DEGs) in immunological gene signatures; Figure-S5. Immune infiltration score in cancer; Figure-S6. Candidate immune responder genes in cancer; Figure-S7. The ROC curves for different signatures in two datasets; Figure-S8. Barplots showing the AUCs of different signatures across 17 datasets; Figure-S9. CNV alteration and QTL of immune-related genes; Figure-S10. QTLs of immune-related genes across cancer types.</p>

Supplementary Figure 3. Effects of MLN8237 on markers of apoptosis and senescence. A-B) Immunofluorescence analysis of p53 wildtype CAL-51 TNBC cell line stably transfected with a p73 shRNA construct (p73-55).

Supplemental Table 1: Molecular and clinical characteristics of the 10 CRC explants used in the PDX model experiments. Seven explants have a KRAS mutation, 4 have a PIK3CA mutation, and none have a BRAF mutation. Over half of the explants came from patients who had previously received treatment for mCRC. Supplemental Table 2: Using DCE-MRI measurements, treatment with cabozantinib demonstrated significant decrease in vascularity on day 28 compared to baseline and day 7. Regorafenib treatment did not demonstrate a decrease in vascularity. Supplemental Table 3: [18F] FDG-PET imaging revealed significantly reduced FDG avidity on days 7 and days 28 of cabozantinib treatment compared to baseline, whereas FDG avidity did not diminish in the regorafenib treated explants. Supplemental Table 4: To better understand the potential metabolic effects of cabozantinib, we analysed signalling pathways involved in metabolism using the KEGG database. Receptor tyrosine kinase (RTK) assay revealed a decrease in many of the pathways responsible for cell metabolism after 3 days of cabozantinib treatment in three of the sensitive CRC explants. Supplemental Figure 1: Cabozantinib and regorafenib treatment in a Humanized HGF mouse model. Cabozantinib had greater antitumor effects when compared to regorafenib in the hHGF scid mice in CRC203 and CRC020. Similar results were observed when compared to athymic nude mice for CRC203 (REG TGII: 39.3, Cabo TGII: 14.93) and CRC020 (REG TGII: 17.5, Cabo TGII: -5.68). Supplemental Figure 2: Changes in TIE2 and VEGFR2 expression were variable and did not demonstrate a sustained decrease after treatment with regorafenib. Supplemental Figure 3: After 7 days of regorafenib treatment, there was no significant reduction in activation of p-AXL, AKT, and pS6 compared to control, and there was a variable trend in p-MET and p-RET levels as measured by RTK assay. Supplemental Figure 4: Regorafenib treatment led to decreased ATG3, LC3A/B, and beclin-1 in the CRC098 explant; however, levels of these enzymes increased in the CRC162 explant. Supplemental Figure 5: In vitro analysis of caspase 3/7 expression in cell lines HCT116 (a, c) and HT29 (b, d) after cabozantinib and crizotinib treatment showed significant increase in autophagy index relative to control*** and to regorafenib###. Caspase 3/7 expression was measured via fluorescence using the CYTO-ID® Autophagy Detection Kit. Crizotinib combined with SBI-020695, an anti-ULK-1 agent, did not result in increased caspase 3/7 expression, whereas caspase 3/7 expression significantly increased with cabozantinib plus SBI-020695.

Supplementary Table 2 - PDF file 62K, Suppl Table 2 IC50s for single agents and combination

Survival curves for RX-5902 treatment and Abraxane and Vehicle control in vivo

Supplementary Figure 3 - PDF file 91K, Combination Fraction Inhibited Graphs

<p>Table S3. Critical regions of immunological genes that are associated with immune response in cancer.</p>

Supplementary Figure Legends - PDF file 41K, Supplementary Figure Legends

CCR Translation for This Article from Development of an Integrated Genomic Classifier for a Novel Agent in Colorectal Cancer: Approach to Individualized Therapy in Early Development

Supplementary Figures S1-S4; Supplementary Tables S1-S5.

Supplementary Figure Legends

Supplementary Figure 1 - PDF file 110K, Single Agent Proliferation Curves

Supplementary Data from Preclinical Development of the Class-I–Selective Histone Deacetylase Inhibitor OKI-179 for the Treatment of Solid Tumors

PDF file - 79K, FISH for AURKA

Supplementary Table 3: Quantification of total cell counts and average cellular size under conditions depicted in Fig. 4A.

Supplementary Methods - PDF file 124K, Supplementary Methods