Chemical objects consideration

[oolonek]

The chemical constitution of biological organisms could be considered in multiple ways.
For now we have mostly thought of :

• 1. individual chemical structures (described by their InChiKey for example)
• 2. chemical classes were 1. would fit in a given chemical classification framework (e.g. NPClassifier, yaccl see ..)

As discussed in #7 (comment) and before, consideration of 2. should be the first way to go because 1. is thought to be at a to "precise" level (maybe it's wrong ...)

But other ways to consider the chemical constitution could be explored:

• 3. Fingerprints of 1. Which would allow us to consider the structural relationships among individual structure in a chemical space and present an alternative to 2. for individual compound organisation

• 4. Vectors of :

  • 4.a. aggregated 1.
  • 4.b. aggregated 2.
  • 4.c. aggregated 3.

Any of 4. offers the advantage to have a single vector describing the chemical content of biological organism.

[Adafede]

We have the TMAP of all structures in LOTUS with related MAP4 fingerprints in case...

[oolonek]

However now that I think about it a problem with 4. is that if it captures and groups more chemical information it is also harder (not possible ?) to come back to the individual 1. or 2. constituting anticipated 4.'s !