Great article! It sounds like we still have a long way to go to simulate a cell, particularly because of the importance of its environmental context. It makes me wonder though, is it more practical or advantageous to simulate a larger block, e.g., tissues, rather than individual cells, given the complexity and interactions within biological systems? It probably depends what we are trying to ultimately predict.
The answer to this is probably yes, but it depends on what you want to know. Physiology for bioengineers simulates organs by things like fluid flow and mass balance which can be done pretty mechanistically with differential equations. But this gets harder if you are trying to develop drugs because those typically operate at the molecular level. If you wanted to use organs for ML training, the same sizes become considerably difficult to acquire. For larger order systems there's lots of room for ML to do classification, but it's more challenging to think about simulation.
Great article! It sounds like we still have a long way to go to simulate a cell, particularly because of the importance of its environmental context. It makes me wonder though, is it more practical or advantageous to simulate a larger block, e.g., tissues, rather than individual cells, given the complexity and interactions within biological systems? It probably depends what we are trying to ultimately predict.
The answer to this is probably yes, but it depends on what you want to know. Physiology for bioengineers simulates organs by things like fluid flow and mass balance which can be done pretty mechanistically with differential equations. But this gets harder if you are trying to develop drugs because those typically operate at the molecular level. If you wanted to use organs for ML training, the same sizes become considerably difficult to acquire. For larger order systems there's lots of room for ML to do classification, but it's more challenging to think about simulation.