Testing free energy of hydration features via Jazzy for solubility prediction.

 This is a first of the “lets-test-this” posts: Google Colab notebook is here.

TLDR: Jazzy features does not have positive impact on solubility prediction.

Intro: hydrogen bonds drive all kind of interactions, and no doubt are important for drug design process. A nice package from AstraZeneca was published recently that predict of hydrogen-bond strengths and free energies of hydration of small molecules. It looks like that the main application of Jazzy is for local settings – protein-ligand strength estimation, SAR studies, but my interest was to test it on compound solubility prediction.

Goal: testing if free energy of hydration (without free parameters) from Jazzy has an impact on prediction of compound solubility.

Setup:

  • Data was taken from SolCuration, that's a derivate from Meng et al paper (AqSol) + additional curated datasets. In case of duplicated measurements, a median was taken and compounds with SD > 0.5 were removed (~1500 drop).
  • Three types of descriptors were used: Morgan fingerprints, 2D descriptors from rdkit, Jazzy features
  • 5-fold random cross-validation splits (no repeats)

Results:

Despite authors showed some positive correlation between compound activities and Jazzy features but for “global” solubility prediction (logS) results are not very promising, and Morgan alone or combination of Morgan + 2D descriptors showing better performance in a default  catboost regression setup (no HP search).


Descriptors

R2

Mean absolute percentage error

RMSE

Jazzy

0.483166

0.697701

1.209996

Morgan

0.647810

0.636827

1.005246

combo_Morgan_Jazzy

0.709274

0.522402

0.912780

combo_2D_Jazzy

0.755690

0.422069

0.832237

2D

0.756753

0.416461

0.830284

combo_2D_Jazzy_Morgan

0.770472

0.431028

0.806384

combo_Morgan_2D

0.772232

0.412204

0.803286

It might be way more impactful if Jazzy features are used for GNN-like architecture, but for tabular data representation 2D/Morgan features are providing better results. 

Links:

 

 

Comments

Post a Comment

Popular posts from this blog

Paper comment: "MELLODDY: cross pharma federated learning at unprecedented scale unlocks benefits in QSAR without compromising proprietary information"

Image
 Intro I finally got my hand on carefully reading papers by the MELLODDY consortium led by my ex-supervisor (and fantastic visionary) - Hugo Ceulemans. I participated in some very-very early discussions and saw the painful process of the project's initiation. What Hugo did is colossal, and he helped to build a PRECEDENT! Honestly, the amount of PR the project got is too little! (Compared to PR of some other ML-for-chemistry companies). The project united ten big pharma companies to build a federated multi-task model for activity assays (activity on proteins, PK/PD, toxicity):  40000+ assays on 21+ million compounds with 2.6+ billion end-points . MELLODDY used federated learning to improve individual models of the consortium partners by confidential sharing of the "model"  without sharing private data itself. An exciting approach that has a lot of tiny, small, and big challenges related to fairness, data security, MLOps, ML engineering, ML algorithms, and legal questions. ...
Read more

How many actual bits in unfolded Morgan fingerprint?

Image
The easiest  possible descriptors used in ML for chemistry are fragments, and they are usually used in a folded form. The default folded size is 1024, with some recommendations to use higher dimensionality (>=4096) for computational chemogenomics tasks.  Folding of descriptors is done for several reasons: to save space, get a uniform length, with a price of  bit collision , and sometimes insensitivity to small changes.  Let's check how many actual bits are in drugs as an example of FDA-approved chemical space vs random selection of 300K compounds from Enamine REAL virtual space of 3B compounds commonly used in virtual screening. Number of unique bits So, as we can see, on average, for radius 3, there are 16 bits per compound on average for drug space that's going down 4 and 2 bits for increased chemical space of 100K and 300K diverse compounds from the Enamine REAL library.  Enamine virtual space is, of course, repetitive, but it still indicates that with a l...
Read more

ChatGPT: Practical AI revolution is here... (1)

The text below is a private opinion not linked in any form to my current employee.   So, all hands on deck… the AI revolution everyone was talking about for a long time arrived. I played a bit with ChatGPT for some of the general R&D and chemoinformatics tasks… and saying that impressed is nothing, I am shocked by the quality of the generated text and code, and the future possibilities of those products. Some people claim that ChatGPT will not replace search engines. With a bit of time, it will likely be a merge between personal assistance and worldwide web search engines. ChatGPT and its successors will revolutionize how software is built, increasing productivity while reducing costs. That, in turn, gives programmers more flexibility and time for more complex tasks. Furthermore, these tools can help bridge gaps between different programming languages, making it easier for coders to operate in heterogeneous environments. ChatGPT-like models will overtake enterprise text mining ...
Read more