La caractérisation à haut débit des interactions protéines-protéines a permis d'établir les premières cartes d'interactions de différents organismes modèles, y compris l'homme. Cependant, la caractérisation structurale des assemblages protéiques reste limitée à un nombre très faible de ces interactions. En mettant en évidence une pression de sélection évolutive spécifique aux interfaces de complexes protéiques, ce travail a permis d'élucider certains mécanismes évolutifs essentiels à l'association entre protéines qui n'avaient pas été décrits jusqu'à présent. Sur cette base, une nouvelle approche bioinformatique, nommée SCOTCH (Surface COmplementarity Trace in Complex History), a été développée pour prédire la structure des assemblages macromoléculaires. Couplée à un programme d'amarrage moléculaire, tel que SCOTCHer, également développé au cours de cette thèse, cette approche a permis de prédire efficacement la structure d'un grand nombre de complexes. Ce travail de thèse s'est également concentré sur l'inhibition des interactions protéiques par des mini-protéines, conçues de façon rationnelle sur la base des structures de complexes. Les résultats obtenus pour deux exemples, celui du complexe Asf1 – Histone H3/H4 et du complexe gp120 – CD4 témoignent du fort potentiel du design rationnel d'interfaces de complexes pour le développement de nouvelles stratégies thérapeutiques.

Background: Anorexia nervosa (AN) is characterized by sizable, widespread gray matter (GM) reductions in the acutely underweight state. However, evidence for persistent alterations after weight-restoration has been surprisingly scarce despite high relapse rates, frequent transitions to other psychiatric disorders, and generally unfavorable outcome. While most studies investigated brain regions separately (univariate analysis), psychiatric disorders can be conceptualized as brain network disorders characterized by multivariate alterations with only subtle local effects. We tested for persistent multivariate structural brain alterations in weight-restored individuals with a history of AN, investigated their putative biological substrate and relation with 1-year treatment outcome. Methods: We trained machine learning models on regional GM measures to classify healthy controls (HC) (N = 289) from individuals at three stages of AN: underweight patients starting intensive treatment (N = 165, used as baseline), patients after partial weight-restoration (N = 115), and former patients after stable and full weight-restoration (N = 89). Alterations after weight-restoration were related to treatment outcome and characterized both anatomically and functionally. Results: Patients could be classified from HC when underweight (ROC-AUC = 0.90) but also after partial weight-restoration (ROC-AUC = 0.64). Alterations after partial weight-restoration were more pronounced in patients with worse outcome and were not detected in long-term weight-recovered individuals, i.e. those with favorable outcome. These alterations were more pronounced in regions with greater functional connectivity, not merely explained by body mass index, and even increases in cortical thickness were observed (insula, lateral orbitofrontal, temporal pole). Conclusions.: Analyzing persistent multivariate brain structural alterations after weight-restoration might help to develop personalized interventions after discharge from inpatient treatment.

The goal of the thesis is to create a predictor of the effect of a mutation on protein solubility given its initial 3D structure. Protein solubility prediction is a bioinformatics problem which is still considered unsolved. Especially a prediction using a 3D structure has not gained much attention yet. A relevant knowledge about proteins, protein solubility and existing predictors is included in the text. The principle of the designed predictor is inspired by the Surface Patches article and therefore it also aims to validate the results achieved by its authors. The designed tool uses changes of positive regions of the electric potential above the protein's surface to make a prediction. The tool has been successfully implemented and series of computationally expensive experiments have been performed. It was shown that the electric potential, hence the predictor itself too, can be successfully used just for a limited set of proteins. On top of that, the method used in the article correlates with a much simpler variable - the protein's net charge.