Augmenting cancer cell proteomics with cellular images - A semantic approach to understand focal adhesion

J Biomed Inform. 2019 Dec:100:103320. doi: 10.1016/j.jbi.2019.103320. Epub 2019 Oct 24.

Abstract

If monolayers of cancer cells are exposed to microgravity, some of the cells cease adhering to the bottom of a culture flask and join three-dimensional aggregates floating in the culture medium. Searching reasons for this change in phenotype, we performed proteome analyses and learnt that accumulation and posttranslational modification of proteins involved in cell-matrix and cell-cell adhesion are affected. To further investigate these proteins, we developed a methodology to find histological images about focal adhesion complex (FA) proteins. Selecting proteins expressed by human FTC-133 and MCF-7 cancer cells and known to be incorporated in FA, we transformed the experimental data to RDF to establish a core semantic knowledgebase. Applying iterative SPARQL queries to Linked Open Databases, we augmented these data with additional functional, transformation- and aggregation-related relationships. Using reasoning, we retrieved publications with images about the spatial arrangement of proteins incorporated in FA. Contextualizing those images enabled us to gain insights about FA of cells changing their site of growth, and to independently validate our experimental results. This new way to link experimental proteome data to biomedical knowledge from various sources via searching images may generally be applied in science when images are a tool of knowledge dissemination.

Keywords: Cellular Imaging; Graphical SPARQL; Linked open database; Semantic Knowledgebase; Tissue engineering.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Focal Adhesions*
  • Humans
  • Knowledge Bases
  • MCF-7 Cells
  • Neoplasm Proteins / metabolism*
  • Neoplasms / pathology*
  • Proteomics*
  • Semantics*

Substances

  • Neoplasm Proteins