A principal component is a new feature formed by combining original features. PCA orders these features by the amount of data variance they explain, allowing a high-dimensional data set to be represented with fewer dimensions. section 1.2.3 section 6.2.3
PCA creates orthogonal axes in a new coordinate system. Retaining the first k components preserves the largest available variation while reducing the computational and visualization burden. section 6.2.3
A practical molecular workflow is:
- Represent molecules with high-dimensional fingerprints, such as 1024-bit ECFP6 features. section 1.2.3
- Fit PCA and transform the feature matrix into a smaller representation, such as four principal components. section 1.2.3 section 6.2.3
- Inspect component cross-sections to explore chemical space or provide inputs for a classifier. section 1.2.3
FDA-approved drug chemical space
A data set of FDA-approved molecules is converted from SMILES into RDKit molecules, encoded as 1024-dimensional Morgan fingerprints, and reduced to four components with sklearn PCA. Cross-sections reveal distinct regions for several USAN stem classes, including “-caine,” “-cillin,” “-pine,” and “-sulfa”; the resulting components can also support logistic-regression decision boundaries. section 1.2.3
Listing 1.2 Example Applications with USAN Stems
from rdkit.Chem.rdFingerprintGenerator import GetMorganGenerator
from sklearn.decomposition import PCA
from sklearn.linear_model import LogisticRegression
fda_approved_df = pd.read_csv(
'data/fda_approved_drugs/fda_approved_drugs.csv', sep=','
)
PandasTools.AddMoleculeColumnToFrame(
fda_approved_df, smilesCol='Smiles', molCol='ROMol'
)
fda_approved_df = fda_approved_df[~fda_approved_df['ROMol'].isnull()]
morgan_gen = GetMorganGenerator(radius=3, fpSize=1024)
fda_approved_df['ECFP6'] = fda_approved_df['ROMol'].apply(
lambda mol: morgan_gen.GetFingerprint(mol)
)
X = np.array([x for x in fda_approved_df['ECFP6']])
pca = PCA(n_components=4, random_state=42)
pca_drugs = pca.fit_transform(X)
fda_approved_df['PC1'] = pca_drugs.T[0]
fda_approved_df['PC2'] = pca_drugs.T[1]
fda_approved_df['PC3'] = pca_drugs.T[2]
fda_approved_df['PC4'] = pca_drugs.T[3]
sns.pairplot(
fda_approved_df,
hue='USAN Stem',
vars=['PC1', 'PC2', 'PC3', 'PC4'],
palette='viridis'
)
fda_approved_df['cillin'] = (
fda_approved_df['USAN Stem'] == '-cillin'
).astype(int)
fda_approved_df['olol'] = (
fda_approved_df['USAN Stem'] == '-olol'
).astype(int)
lr_model_cillin = LogisticRegression()
lr_model_cillin.fit(fda_approved_df[['PC1', 'PC3']], fda_approved_df.cillin)
xx = np.mgrid[fda_approved_df.PC1.min():fda_approved_df.PC1.max():.01
yy = fda_approved_df.PC3.min():fda_approved_df.PC3.max():.01]
grid = np.c_[xx.ravel(), yy.ravel()]
probs = lr_model_cillin.predict_proba(grid)[:, 1].reshape(xx.shape)
Runnable code shows RDKit molecular conversion, Morgan fingerprint generation, sklearn PCA fitting, and assignment of PC1–PC4 coordinates. section 1.2.3





