Les personnes qui vivent avec le VIH (PVVIH) doivent prendre un traitement d’antirétroviraux combinés (ART) pour contrôler la réplication virale et empêcher le développement d’une immunodéficience dont l’issue est fatale. Les ART protègent les cellules saines de l’infection et permettent ainsi de rendre la charge virale plasmatique indétectable. Cependant, l’arrêt des ART entraine presque inévitablement un rebond de la charge virale, puisque le virus n’est jamais complètement éliminé par le système immunitaire. En effet, de multiples cellules infectées, où le virus s’est intégré et demeure dans un état de latence, restent présentes tout au long de la vie des PVVIH. Une partie de ces cellules infectées forme le réservoir compétent pour la réplication. Les provirus responsables du rebond de la charge virale possèdent trois caractéristiques : ils gardent la capacité d’être réactivés (sortie de latence), ils sont génétiquement intacts, et ils peuvent produire de nouvelles particules virales infectieuses. Afin de guérir les PVVIH de l’infection, il faut donc cibler les quelques rares cellules portant un provirus intact et inductible. Pour ce faire, il est impératif de comprendre comment ces cellules sont maintenues pendant les années de ART, de les localiser dans tout l’organisme, et d’identifier ce qui peut les distinguer des autres. Ce sont ces trois aspects que nous avons abordés au cours des travaux de recherche présentés dans cette thèse, autant à l’échelle de la cellule unique que de l’organisme entier. Nos résultats montrent que les provirus compétents pour la réplication persistent dans des lymphocytes T CD4+ mémoires exprimant l’intégrine VLA-4 en grande quantité, que les provirus intacts peuvent subsister au sein de différents compartiments anatomiques, que les provirus inductibles et compétents pour la traduction de la protéine virale p24 sont majoritairement défectifs, et que l’expansion clonale est un mécanisme important qui favorise le maintien du réservoir viral dans le sang et dans les tissus tout en favorisant la diversité phénotypique de ces cellules. / People with HIV (PWH) must take combinational antiretroviral therapy (ART) to control viral replication and avoid developing fatal immunodeficiency. ART allows achieving undetectable plasma viral load, and thus protects uninfected cells from HIV. However, ART interruption will almost inevitably result in a viral rebound since HIV is never completely cleared by the immune system. Indeed, a group of infected cells, where the virus has integrated and remains in a latent state, persists throughout the life course of PWH, and some of these cells form the replicationcompetent reservoir. Proviruses responsible for viral rebound have three characteristics: they can be induced to exit their latent state, they are genetically intact, and they are able to produce new infectious viral particles. Therefore, in order to cure PWH, it is essential to target the few cells with intact and inducible provirus, and to be able to do so, it is imperative to understand how these cells are maintained during years of ART, to localize them throughout the body, and to identify what distinguishes them from other cells. These three aspects are the focus of the work presented in this thesis, whether at the single-cell level or looking through the whole body. Our results show that replication-competent proviruses persist in memory CD4+ T cells expressing high levels of the integrin VLA-4, that intact proviruses can persist among various anatomical compartments, that inducible and translation-competent proviruses are predominantly defective, and that clonal expansion is an important mechanism that favors the maintenance of reservoir cells both in the blood and in deep tissues in addition to diversify phenotypically those cells.

In forensic autopsies, accurately estimating the postmortem interval (PMI) is crucial. Traditional methods, relying on physical parameters and police data, often lack precision, particularly after approximately two days have passed since the person's death. New methods are increasingly focusing on analyzing postmortem metabolomics in biological systems, acting as a 'fingerprint' of ongoing processes influenced by internal and external molecules. By carefully analyzing these metabolomic profiles, which span a diverse range of information from events preceding death to postmortem changes, there is potential to provide more accurate estimates of the PMI. The limitation of available real human data has hindered comprehensive investigation until recently. Large-scale metabolomic data collected by the National Board of Forensic Medicine (RMV, Rättsmedicinalverket) presents a unique opportunity for predictive analysis in forensic science, enabling innovative approaches for improving  PMI estimation. However, the metabolomic data appears to be large, complex, and potentially nonlinear, making it difficult to interpret. This underscores the importance of effectively employing machine learning algorithms to manage metabolomic data for the purpose of PMI predictions, the primary focus of this project.  In this study, a dataset consisting of 4,866 human samples and 2,304 metabolites from the RMV was utilized to train a model capable of predicting the PMI. Random Forest (RF) and Artificial Neural Network (ANN) models were then employed for PMI prediction. Furthermore, feature selection and incorporating sex and age into the model were explored to improve the neural network's performance.  This master's thesis shows that ANN consistently outperforms RF in PMI estimation, achieving an R2 of 0.68 and an MAE of 1.51 days compared to RF's R2 of 0.43 and MAE of 2.0 days across the entire PMI-interval. Additionally, feature selection indicates that only 35% of total metabolites are necessary for comparable results with maintained predictive accuracy. Furthermore, Principal Component Analysis (PCA) reveals that these informative metabolites are primarily located within a specific cluster on the first and second principal components (PC), suggesting a need for further research into the biological context of these metabolites.  In conclusion, the dataset has proven valuable for predicting PMI. This indicates significant potential for employing machine learning models in PMI estimation, thereby assisting forensic pathologists in determining the time of death. Notably, the model shows promise in surpassing current methods and filling crucial gaps in the field, representing an important step towards achieving accurate PMI estimations in forensic practice. This project suggests that machine learning will play a central role in assisting with determining time since death in the future.