Experimental evaluation of similarity indices about the MIA molecular family - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2022

Experimental evaluation of similarity indices about the MIA molecular family

Évaluation expérimentale des indices de similarité concernant la famille moléculaire MIA

Résumé

The purpose of this article is to evaluate the performance of the Morgan/Tanimoto, modified cosine similarity, Spec2Vec and MS2Deep Score similarity indices that aim at measuring structural and spectral similarities between molecules. To illustrate this evaluation, we used a spectral database containing about a hundred of the emblematic monoterpene indole alkaloids (MIADB). Comparing two molecules in order to see how similar they are is an important issue for chemists. Such comparisons are used in order to detect new molecules, classify known molecules with respect to their potential use, etc. Overall, two types of information may be used in order to compute similarities between molecules: information on their atomic structures or information obtained from their analytical read-out such as their tandem mass spectrometyr (MS/MS) and NMR spectra, to name a few. There exist different way to compare such information. In this article we analyse the performance of four similarity indices (generally denoted by $\sigma$): one based on their atomic structure and three based on their MS/MS spectra.
Fichier principal
Vignette du fichier
Skeletons___abstract.pdf (302.59 Ko) Télécharger le fichier
Skeletons___abstract (1).pdf (302.59 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04272008 , version 1 (01-12-2023)

Identifiants

  • HAL Id : hal-04272008 , version 1

Citer

Yassine M Mejri, Mehdi A Beniddir, Olivier Cailloux, Meltem Öztürk. Experimental evaluation of similarity indices about the MIA molecular family. Multiple-Criteria Decision Aid to Preference Learning, Nov 2022, Compiègne, France. ⟨hal-04272008⟩
69 Consultations
22 Téléchargements

Partager

Gmail Mastodon Facebook X LinkedIn More