空间维度上的非整倍体预测

Cancer emergence and progression are often driven by genomic instability and disruption of the tissue microenvironment. Insights on the prognostic values of genomic features have been reported in cohort of cancer patients, suggesting Copy Number Alterations (CNAs) signature as highly relevant in multiple cancer types [1,2]. Moreover, insights on the transcriptional state of the tumor microenvironment suggest the recurrence of spatially defined gene modules with functional significance among cancer states [3].Building on the above insights, we aim to study the topology of the spatial distribution of CNAs and its influence on the transcriptional state of the tumor microenvironment. To this end, we describe a workflow that enables the prediction of CNAs in spatially resolved transcriptomics data, and the construction of a topological signature from the CNAs prediction map of the tumor sample. The topological signature is then used as a topological descriptor of the tumor microenvironment.We ported to Python the Numbat algorithm [4] and expanded it with a custom implementation of personalized PageRank algorithm [5] on the diffusion map [6] of the CNAs feature, enhancing the spatial signal of the predicted CNAs. Superlevel set filtration is then applied to the CNAs predictions, obtaining a persistent homology signature of the CNAs spatial distribution [7].Experiments suggest that our method for including the spatial context in CNAs predictions produces simpler data topology (spatial smoothing) and, similarly to a recent report [8], that the resulting persistent homology signature of spatial CNAs can be used to discriminate between states of the tumor microenvironment. References: 1. Smith JC, SheltzerJM. Genome-wide identification and analysis of prognostic features in human cancers. Cell Rep. 2022. 2. Steele, C.D., Abbasi, A., Islam, S.M.A. et al. Signatures of copy number alterations in human cancer. Nature 606, 984-991 (2022). 3. Barkley, D., Moncada, R., Pour, M. et al.Cancer cell states recur across tumor types and form specific interactions with the tumor microenvironment. Nat Genet54, 1192-1201 (2022). 4. Gao, T., Soldatov, R., Sarkar, H. et al. Haplotype-aware analysis of somatic copy number variations from single-cell transcriptomes. Nat Biotechnol 41, 417-426 (2023). 5. Page, L. and Brin, S. and Motwani, R. and Winograd, T. (1999). The PageRank Citation Ranking: Bringing Order to the Web. Technical Report. Stanford InfoLab. 6. R.R. Coifman, S. Lafon, F. Warner, & S.W. Zucker. Geometric diffusions as a tool for harmonic analysis and structure definition of data: Diffusion maps, PNAS. U.S.A. (2005). 7. Carlsson, G. E. (2009). Topology and data. Bulletin of the American Mathematical Society, 46(2), 255-308. 8. I. H.R. Yoon, R. Jenkins, C. Swanton, H. M. Byrne, E. Sahai. Deciphering the diversity and sequence of extracellular matrix and cellular spatial patterns in lung adenocarcinoma using topological data analysis. bioRxiv 2024.01.05.574362