I am a researcher in psychology and I am trying to come up with a framework/platform in which psychological phenomena can be studied from a dynamical systems / computational perspective in collaboration with many labs. To be more concrete, the question is how to organize knowledge about psychological variables and how they are coupled by functional relationships on different time-scales that may be a function of additional variables.

My hope was to learn from disciplines that already work with a dynamical systems/computational framework for a long time. The first interesting analogy I came across were metabolic networks. I read this review about their reconstruction and while many things are different, the task is very similar to ours in that a given network of entities and functional relationships is continuously refined by the scientific community.

I saw that there are many data bases with genes, pathways, annotated graphs etc. but my main question is:

How is the scientific process of refining such a metabolic network organized? Are there platforms for this? Is there anything in place to quantify the uncertainty associated with parts of the network or to resolve disagreements? Are there any standards for implementations? Or is each group simply creating their own version based on previous literature and then publishing their new network in a paper?

I realize that this is a rather broad question and I apologize if I posted this in the wrong stackexchange forum. But I didn't know where else it would fit. I'd appreciate a lot if someone could give me some basic insights about how this field works or if someone has some pointers to relevant reading material.

  • $\begingroup$ In your comment to the answer from @Nicolai you refer to " a (formalized) dynamical system with many variables and many more parameters which render it impossible for a single research lab to map everything out alone". I don't think this is so. It does not strike me as dynamic and a single person usual writes the program that a single lab uses. One generally works with a network of pathways that is well established and stable, takes a bacterium, sequences it, runs the program to see which enzymes hit the network, and whether they join up to give particular pathways. Game over. $\endgroup$ – David Sep 8 '18 at 21:48

A bit of a preface:

A lot of this large scale collaboration (which I would maybe rather describe as the work of a whole [sub]field over a given time frame) often ends up in some sort of database. Really, there A LOT of those in modern biology - and often with minor but sometimes important differences. One of those would be the level of curation: are results gathered automatically from other sources, or always manually checked before being added? Another thing about databases is that (with certain exceptions of course) they often are only relevant for specific sub field, in some cases this can mean that only one lab maintains a website with a databases that is used by a few others, but that's it.
What I'm trying to get to is that there isn't really a 'one size fits all' solution and that especially the size of your field will impact which kind of solutions are both feasible and actually useful.

Now to answer your actually state question (as far as I can):

Most known biochemical pathways (including the metabolic ones) are described in the KEGG database. Initially this project was started by a single lab (and is still run by that group as far as I can tell), but of course it contains a lot of published information from other sources (I couldn't find information on how they update the networks in the database, you may have to read the papers for this). The database can also be accessed remotely, so it allows people to use the same network for a given reaction, and by incorporating sequence data it also has information on which organisms even contain the given enzymes in the network.
One thing the database does not contain is information about dynamics in the network. In this case I think most groups will generate and publish their own models (e.g. based on ODE's).

  • $\begingroup$ I think the poster is talking about computationally determining which metabolic pathways are present in an uncharacterized organism from the genes detected, using the full set of possible pathways in KEGG or the like as a reference. So the info about the KEGG database isn't particularly relevant to the fitting process. I imagine @jmb should try SE Bioinformatics where there may be someone who knows the approaches used in the programmatic fitting.. $\endgroup$ – David Sep 7 '18 at 14:44
  • $\begingroup$ @ Nicolai thanks for the answer and @David thanks for checking in. I was less interested in the specifics of the field but rather in how efficient collaboration works in fields in which the phenomena of interest is modeled as a (formalized) dynamical system with many variables and many more parameters which render it impossible for a single research lab to map everything out alone. $\endgroup$ – jmb Sep 7 '18 at 21:07
  • $\begingroup$ @jmb — Sorry I misunderstood. I thing you have it wrong, but I'll comment on your question. $\endgroup$ – David Sep 8 '18 at 21:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.