EPCO-13. CNMF-SNS, A FRAMEWORK FOR UNSUPERVISED INTEGRATION OF MULTIOMICS DATA, IDENTIFIES INVASION AND RECURRENCE ASSOCIATED PROGRAMS IN GBM
Bibliographic record
Abstract
Abstract Glioblastoma (GBM) is an aggressive cancer of the brain for which novel therapies are urgently needed. Heterogeneity in GBM underlies tumor evolution and resistance to therapy and involves genetic diversity of tumor cells, spatially distinct niches harbouring stem-like cells, and immune cells spanning anti- and pro-tumor states. No single profiling methodology can capture all these facets of tumor biology in the same space and time, necessitating multi-omics data collection and computational methods suited for multi-omic integration across and within cohorts. We therefore set out to develop a generalizable framework for integration that can bridge across cohorts and data types. Using consensus non-negative matrix factorization (cNMF), we discover gene expression programs (GEPs) corresponding to cell types and/or states in an unsupervised fashion. Rather than choosing a single rank (number of programs), we identify both high and low-resolution programs for each dataset, and then integrate programs from all datasets and ranks into a solution network space (SNS) on which graph algorithms identify communities of highly similar GEPs identified in all or some datasets. These communities are characterized by gene set enrichment analyses and association with sample metadata. SNS communities can be used to transfer sample metadata across datasets, enabling multi-omics and multi-cohort integration, even in the absence of shared samples or cells. We showcase the utility of cNMF-SNS in integrating multiomics datasets of GBM from single-cell to bulk RNA-Seq and mass spectrometry proteomics. We further demonstrate the power of cNMF-SNS to define and cross-annotate programs identified in 5 complementary mass spectrometry datasets of GBM, layering on survival, driver gene mutation status, clinical features, and imaging features. We demonstrate robust mapping of biologically and clinically relevant processes, highlighting how programs for tumor cell invasion and therapy resistance can be mined for novel therapeutic targets. cNMF-SNS is available at https://github.com/MorrissyLab/cNMF-SNS.
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 | 0.003 | 0.007 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.002 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
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".