Algorithm predicts cell changes
Researchers from the Duke-NUS Medical School (Duke-NUS), the University of Bristol, Monash University and RIKEN have designed an algorithm that can predict the factors necessary for human cell conversion. The findings, published in Nature Genetics on 18 January, have important implications for regenerative medicine.
Cell types are not constant, and may be reprogrammed or transformed into another cell type through the addition of a unique set of cellular factors. An example of this is seen in the Nobel Prize-winning work by Professor Shinya Yamanaka, who converted skin fibroblast cells to induce pluripotent stem cells (iPS), which could potentially be further reprogrammed to become any type of cells, for instance, retinal cells.
However, the two-step conversion carries the risk of cancerous mutations, leading to unpredictable behaviour. Moreover, determining the unique set of cellular factors needed for each cell conversion is a long and costly process involving much trial and error.
After five years of research, Duke-NUS Senior Research Fellow Dr Owen Rackham has developed a computational data-driven algorithm named Mogrify. Making use of a database containing gene expression from an estimated 300 different human cell and tissue types, the innovative method is able to predict the optimal set of cellular factors required for any given cell conversion. When put to the test, the algorithm passed with flying colours, accurately forecasting the set of cellular factors necessary for previously documented cell conversions. It also successfully predicted the outcome of two human cell conversions which have not been carried out in the past.
“Mogrify acts like a ‘world atlas’ for the cell and allows us to map out new territories in cell conversions in humans,” explained Dr Rackham, who is from the Systems Genetics of Complex Disease Laboratory (SGCDL) in the Centre for Computational Biology at Duke-NUS. He said that one of the first clinical applications they aim to achieve is to reprogramme patients’ defective cells into functioning healthy cells, without the intermediate iPS step.
Associate Professor Enrico Petretto, co-author of the study and head of SGCDL, said that Mogrify leverages big data and systems biology, and can be expanded into clinical applications. Its robustness and accuracy will continue to improve as more data are input into the framework, he added.
The team at Duke-NUS plans to apply Mogrify to translational medicine. Collaborative efforts between research groups within Duke-NUS have been established to use the scientific method for developing treatments for specific diseases, such as cancer.
