Similarities

Being able to identify molecules that are similar to the one we study can allow to infer some of its properties. There are several ways of measuring the similarity between molecules, by their structure, effects, pharmacological interactions etc. In the following, you can find similar molecules according to various criterions and tools that we developed. To understand the limitations of these comparisons, it is crucial to always refer to the methodology that was used to measure those similarities. Please note: This information is provided solely for informational purposes and should not be interpreted as medical advice.

To measure structural similarity, we use the Mol2vec method, which is a neural network that processes molecules and transform them into points in spaces, such that molecules with chemically related substructures are transformed into points that are close in space.

Molecule properties

Using the KGPT Deep Learning model, we predict several property of the molecule. Predictions are grouped by the dataset that was used to get those prediction. Along with each prediction, we provide a plot that shows the distribution of predicted values on the train/test/val dataset. This gives an estimate of the reliability of the model.

Description: A dataset focused on predicting the inhibitory effects of molecules on the enzyme beta-secretase 1 (BACE1). BACE1 inhibition is a potential target for Alzheimer's disease treatment.
Class
TrainValTest
2.88-3.70-1.36
Description: A dataset providing insights on the ability of molecules to penetrate the blood-brain barrier. Crucial for understanding the potential of molecules as central nervous system drugs.
p_np
TrainValTest
-2.24-3.290.40
Description: This dataset deals with the FDA approval status and clinical trial toxicity of molecules. Important for understanding the safety and regulatory status of compounds.
CT_TOX
TrainValTest
-0.49-0.78-0.91
FDA_APPROVED
TrainValTest
0.53-1.27-0.50
Description: A dataset that predicts the solubility of molecules in water. Solubility is an essential property influencing bioavailability and the potential formulation of a drug.
logSolubilitylog(mol/L)
1.11
TrainValTest
Description: This dataset is centered on predicting the free energy when a molecule is dissolved in water. The energy changes can affect molecular interactions in biological systems.
freesolvkcal/mol
-0.74
TrainValTest
Description: A dataset predicting the lipophilicity of molecules. Lipophilicity is a crucial factor affecting the distribution, metabolism, and excretion of drugs in the body.
lipoAlogP
-0.63
TrainValTest
Description: This dataset gives insights into the metabolic stability of molecules. High metabolic stability often results in a longer half-life, influencing drug dosage and frequency.
low
TrainValTest
-0.17-3.50-2.88
high
TrainValTest
0.72-3.55-0.49
27 entries
12 entries
617 entries

Advanced insights

In the following we provide more advanced analysis about the interactions of a molecule with the human metabolism, docking sites etc.

Interaction of this molecule with metabolism

We use a Deep Learning model to predict the interactions of this molecule with metabolism. Refer to the source to understand the methodology.

Reactions that metabolism produce from this molecule
Aliphatic hydroxylation of methyl carbon adjacent to aromatic ring BTMR1058
Products