Back

Accurate identification of cancer cells in complex pre-clinical models using deep-learning: a transfection free approach.

Cortesi, M.; Liu, D.; Powell, E.; Barlow, E.; Warton, K.; Ford, C. E.

2023-10-18 bioengineering
10.1101/2023.10.15.562411 bioRxiv
Show abstract

3D co-cultures are key tools for in vitro biomedical research as they recapitulate more closely the in vivo environment, while allowing control of the density and type of cells included in the analysis, as well as the experimental conditions in which they are maintained. More widespread application of these models is hampered however by the limited technologies available for their analysis. The separation of the contribution of the different cell types, in particular, is a fundamental challenge. In this work, we present ORACLE, a deep neural network trained to distinguish between ovarian cancer and healthy cells based on the shape of their nucleus. The extensive validation that we have conducted includes multiple cell lines and patient derived cultures to characterise the effect of all the major potential confounding factors. High accuracy and reliability were maintained throughout the analysis demonstrating ORACLE effectiveness with this detection and classification task. ORACLE is freely available (https://github.com/MarilisaCortesi/ORACLE/tree/main) and can be used to recognise both ovarian cancer cell lines and primary patient-derived cells. This feature sets ORACLE apart from currently available analysis methods and opens the possibility of analysing in vitro co-cultures comprised solely of patient-derived cells.

Matching journals

The top 10 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.