Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

< Back to Article

Fig 1.

Flowchart with the different steps implicated in the study.

More »

Fig 1 Expand

Fig 2.

Flowchart including the steps implicated in the calculation of different similarity measures.

Drugs were represented as fingerprints, i.e. bit vector codifying the presence or absence (1, 0) of structural keys, adverse effects, targets, drug-drug interactions or ATC codes. The Tanimoto coefficient (Tc) between all the fingerprint pairs is calculated and placed in a drug-drug similarity matrix (M2). Different M2 matrices are calculated weighted with the different similarity measures.

More »

Fig 2 Expand

Fig 3.

ROC results using different methods to rank the 386 TWOSIDES candidates: PRR (Proportional Reporting Ratio), p-values, 2D structural similarity (MACCS), 3D structural similarity, ADEPF (Adverse Drug Effect Profile Fingerprint), TPF (Target Profile Fingerprint), DDIPF (Drug-Drug Interaction Profile Fingerprint), ATC-code fingerprint, PCA (Principal Component Analysis) and LDA (Linear Discriminant Analysis).

Panel (a) shows pAUROCs with 95% confidence intervals. Panels (b) and (c) show the ROC curves for the different methods.

More »

Fig 3 Expand

Table 1.

pAUROCs using different methods to rank the DDIs.

Drugdex (sets 1–3) and Drugs.com (sets 4–5) were used as reference standards.

More »

Table 1 Expand

Fig 4.

Precision of the different methods in test set 1 with all the interactions described in the reference standard Drugdex (interactions well_established+probable+theoretical).

More »

Fig 4 Expand

Fig 5.

Precision of the different methods in test set 4 with all the interactions described in the reference standard Drugs.com (high and moderate clinically significant interactions).

More »

Fig 5 Expand

Table 2.

Example of some arrhythmia DDIs described in Drugdex and detected by the different similarity-based models (2D MACCS, 3D similarity, ADEPF, TPF, DDIPF and ATC)

.

More »

Table 2 Expand

Fig 6.

Examples of different pairs of similar drugs with different pharmacological profile detected by our models.

Panel (a): methadone is similar to amitriptyline and predicted to interact with fluconazole (reference standard amitriptyline-fluconazole). Panel (b): amitriptyline is similar to disopyramide and predicted by the 3D model to interact with gatifloxacin (reference standard disopyramide—gatifloxacin). Panel (c): citalopram, was found to be similar to disopyramide and hence, to interact with ranolazine (reference standard disopyramide-ranolazine). Panel (d): diltiazem was found to be similar to fluconazole and predicted to interact with imipramine (reference standard fluconazole-imipramine).

More »

Fig 6 Expand

Table 3.

Correlation coefficients (r) between the six drug similarity measures.

More »

Table 3 Expand