Abstract PO-020: Gene regulatory network connectivity analysis identifies novel candidate effectors of HNSC tumorigenesis
Bibliographic record
Abstract
Abstract The ever-increasing availability of publicly available cancer omics datasets makes it possible to perform in-silico studies aimed at elucidating disease mechanisms of action (MOAs) that may aid in therapeutic development. To this end, we recently developed Structure Learning for Hierarchical Networks (SHiNe) to model and distinguish direct and indirect signaling interactions from gene expression data to identify key signaling “hubs” and “bottlenecks”, associated crosstalk, and downstream targets that we used to identify potential effectors of head and neck squamous carcinoma (HNSC). SHiNe is an advanced gene regulatory network reconstruction approach optimized for learning multiple Markov networks in the “large p, small n” settings (i.e., large number of genes, small number of samples) typical of omics data. We applied SHiNe to the analysis of TCGA-HNSC RNA-seq data to learn HPV-negative and subtype- specific networks for previously validated molecular subtypes (Atypical, Basal, Classical, and Mesenchymal). SHiNe learned networks were sparser, with higher clustering coefficients, and more significantly enriched for protein-protein interactions than randomly simulated networks. Further annotation of the learned networks with multiple layers of omics information revealed several highly eigen-central genes in the HPV-negative network to be characterized by additional omics features. Examples included EP300, ranked 2nd by eigen-centrality with a single somatic mutation (SSM) rate of 7.87% (versus 4.59% in all other TCGA tumors), MYH9 (rank=12, SSM=6.1%), RPS6KA4 (rank=39, CNV Gain=16.5%), GNAI3 (rank=45, CNV Loss=10.55%), and CFL1 (rank=51, CNV Gain=20.72%). Community detection in the HPV(-) network structure using the Walktrap algorithm yielded 33 signaling communities, with the highest centrality community encompassing most of the central genes within the network (e.g., KLHL11, EP300, TAOK1, ADAM17, APAF1) and including genes strongly enriched for DNA repair and maintenance pathways. Co-localization analysis of omics features in the HPV-negative network identified within the most central community revealed a cluster of genes with copy number variation gain ≥ 25%, including ABCC5, ACTL6A, AP2M1, ATP11B, MAP3K13, NCBP2, PAK2, PIK3CA, PLD1, PRKCI, SKIL, TBL1XR1, and ZMAT3, with most of these genes located on the q arm of chromosome 3. While amplification of 3q26-29 is a known characteristic of HNSCC associated with poorer patient outcome, the MOA remains unclear, and thus our network analysis offers an opportunity to identify meaningful signaling interactions for therapeutic intervention of 3q26-29 mutated tumors. Indeed, we have identified candidate associations with the 3q26-29 amplicon genes that we are currently validating. We are also developing a graphical interface for the interactive inspection of the inferred networks. Taken together, these results highlight the potential for network-based analysis to support the study and identification of novel candidate regulators of head and neck tumorigenesis. Citation Format: Anthony Federico, Eric Reed, Maria Kukuruzinska, Xaralabos Varelas, Stefano Monti. Gene regulatory network connectivity analysis identifies novel candidate effectors of HNSC tumorigenesis [abstract]. In: Proceedings of the AACR-AHNS Head and Neck Cancer Conference: Innovating through Basic, Clinical, and Translational Research; 2023 Jul 7-8; Montreal, QC, Canada. Philadelphia (PA): AACR; Clin Cancer Res 2023;29(18_Suppl):Abstract nr PO-020.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.001 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.003 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".