The *object-oriented annotation database* concept about compounddb HOT 6 CLOSED

Sounds very reasonable to me. so XCompDb Could basically be any info you wanted to add?

from compounddb.

That's the plan. Having a base CompDb class and then e.g. a RtimeCompDb class that provides in addition retention times for some of the compounds available in CompDb. In such a RtimeCompDb users could put their own data and perform an annotation based on m/z and rtime. The benefit I think is that we don't have to think of all possible scenarios for the class/database layout. The base layout should simply reflect what is publicly available, and we build on that and extend it (without altering the base layout). Thus it's somehow like object oriented programming.

from compounddb.

But RtimeCompDb implies a specific object for RT right? What if you wanted to add random annotation? From whatever prediction tool for example?

from compounddb.

The rtime was just an example - you could have additional objects/classes, I would not want to limit to e.g. the RtimeCompDb as only class. The idea is to have several classes which could be sort of plugged-in to the annotation process.

from compounddb.

Got it. So content specific class (not an attach whatever data.frame class).
Yes this would be very nice.

Perhaps there could be a formal way to link feature tables entries and CompDb entries?

from compounddb.

This is now implemented with the CompDb and the IonDb.

from compounddb.