Genuinely curious, couldn't we work with smaller models and adapt/evolve them as we learn more? A bit like physics trying to find the unified theory?
back
1 comments
The problem is one of going from a system of two or three variables to one of two or three thousand; this cannot happen via an evolutionary progression, it requires wholly new kinds of thinking (and math).
Which are the main candidates for new paradigms in systemic biology ?
I'm not sure they exist yet; we're still at the stage where we are trying to appreciate the complexity of the systems we are studying. The first step is to catalog. For example we've only recently begun to appreciate the extent of genetic heterogeneity in cancers (and the therapeutic significance of this variation). Now that we have this sense people are beginning to think about new experimental methods for collecting this kind of information; we are still a few years (maybe a decade) out from good models of how to analyze it.
I can't naively hope that very high performance (exascale) computing helps finding the right theories here. But I do hope that people don't waste time digging in the wrong direction. Facing complexity is better than not being aware of it.
ps: I hope though, in the case of non mental disorders, that finer and cheaper technology allow for preemptive discovery, continuous monitoring and non invasive therapies.
Computing will help. But we also need new fundamental technologies that get us information that old methods from molecular biology miss. DNA sequencing isn't everything--we need to be able to look at ALL of the biomolecules within an organism and their dynamics.
I know. I was talking to someone in the genetic field and he agreed that the legacy view on genetics is out of touch. So far mainstream media still report this kind of sequencing wows but I hope research is aiming at a deeper yet leaner model.
The mathematician John Baez is exploring the mathematics of complex systems with his students. Here's a place to start reading[1] but a caveat: it would take a lot of patience and link-following to grok this research from scratch.
[1] https://golem.ph.utexas.edu/category/2015/04/categories_in_c...
Much appreciated.
My suspicion is Machine Learning (IMHO)
You can feed a lot of data then try to derive (maybe human readable) results
The problem as always with machine learning methods is figuring out how to construct your machine correctly and what data to feed it. Mostly this is about good data collection, not about what machine you use. I am somewhat agnostic on whether you train using a linear model, SVM, neural network, etc. - all of these are likely to be analytically useful, provided I can give them an interesting and well-contsructed input set. But if the data I'm feeding into such a system is too noisy, doesn't sample deeply enough to catalog the extent of variation, or we don't have good descriptors for correctly labeling the dataset, it's useless to try to train a machine, we'll just have GIGO.