search

Open-Systems-Pharmacology / ICRP-Extended

On going work regarding ICRP Population improvement
0 stars 0 forks source link

Update volume with new data points #1

Open msevestre opened 6 years ago

msevestre commented 6 years ago

@JanSchlender @Erik-Sjogren I think we can now start to work on this feature. I have updated a copy of the PKSimDB as well as instruction on how to update the database use by PK-Sim with this one. The PKSimDB.mdb file contains at the moment an exact copy of the ICRP Database called ICRP Extended. This pop should be changed in order to get the new behavior

Erik-Sjogren commented 6 years ago

Great! One thing before we start is to decide to what level we would like to do this update, i.e. what information to include. For instance, before I learnt about the PK-Sim pop-generation-algorithm I made a function to include the "mass of body fat"-development profile to accommodate the possibility to distribute BWT between fat and muscles. There are obviously additional tissue-mass-development profiles to consider if wanted. I also wonder if this should be considered as a stand-alone project or if we must keep an eye on the default populations to keep some consistency within PK-Sim?

msevestre commented 6 years ago

I also wonder if this should be considered as a stand-alone project or if we must keep an eye on the default populations to keep some consistency within PK-Sim?

Right now there are 2 populations in the PKSimDB defined in this repo:

Once we are done with the implementation of the ICRP Extended population and we feel that this is ready to go into PK-Sim, my suggestion would be to keep 2 different populations. That way skeptical can always used the previous version (or simply to reproduce previous results). New projects can be done using either. As long as we document clearly what the difference is, this should be great.

So my 2 cents: default population will remain unchanged and we should provided a new population in PK-Sim when the implementation is finished

what information to include

I had started a while ago a wiki describing some internal algorithms. This has not been published yet but will be soon hopefully. (This is a private repo of mine but you should have access normally)

In particular check Create Individual and Create Population.

Fat weight is being used as a compensation parameter to achieve the desire output weight. We should not change this strategy at the moment.

In my eyes, but I'd like some input from @JanSchlender and @Yuri05 here, there are only a hand full of parameters to update with new data points between 15 and 30 years of age.

Also we have some discrete values for age 15 and 30 for the following parameters (i.e. a value based on age but no variability). We may have to add some new data points here as well (I have no idea if this data actually come from ICRP....anyone?). Alternatively we are happy with the linear interpolation between age 15 and 30. Not my expertise here :)