Natural products (NPs) have long been used for their pharmacological properties, to cure a range of different ailments. From the ancient and traditional uses of natural extracts (NEs) to the first isolation of a NP over 200 years ago, NPs have consistently delivered new therapeutic solutions. They have had a paradigm-changing effect on the fight against infectious diseases, especially after the first isolation of penicillin in 1940. In that context, tuberculosis has been a particular burden for centuries, one which has been greatly reduced thanks to antibiotics of natural origin. Over time though, bacteria have developed resistance to antibiotics, which prompted a need for new therapeutic solutions such as anti-infective agents that target malignant bacteria more selectively. This thesis focuses on isolating anti-infective agents through a collaboration with the Soldati lab (University of Geneva) that has developed an in cellulo growth assay in which Mycobacterium marinum is infecting the amoeba Dictyostelium discoideum. The source of chemical diversity used was a collection of 1600 NEs obtained through a collaboration with the company Pierre Fabre^®. One of the main objectives of the thesis was to transform the content of this NEs collection into a virtual chemical library that could be explored. In-silico tools have been used to achieve this goal, alongside mass spectrometry based high-resolution metabolite profiling methods together with molecular networks. HPLC bioactivity profiling and advanced chromatographic techniques were also used at various scales, with the end goal always being the isolation of bioactive NPs. All data generated was also systematically organized in a knowledge graph framework in partnership with the Swiss institute of bioinformatics (SIB). This innovative way of connecting heterogeneous data in a semantic-web format (resource description framework, RDF) was used as a proof-of-concept to help NE prioritization. SPARQL queries were developed to search across diverse RDF datasources gathered in the KG to make rational decisions in the isolation of bioactive NPs. Complementary biological activity assessments were carried out on a case-by-case basis, in order to deepen the understanding of the biological processes involved in the observed anti-infective effects. These tests included bacterial growth assays on other types of bacteria (Pseudomonas aeruginosa and Staphylococcus aureus) as well as pro-metabolic assays (lipid droplet behavior, propidium iodide staining and autophagy) that were carried out. In addition to bio-guided isolations, chemically informed extract selections were used to direct the isolation of anti-infective agents. A total of 115 NPs with various degrees of anti-infective activities have been isolated from extracts of 21 different plant parts during the course of the thesis. Synthetic and hemi-synthetic derivatization were also elaborated to improve the activity patterns based on promising scaffolds. Complete structural elucidation was carried out using NMR, HRMS/MS and circular dichroism to confidently establish the chemical structures of NPs up to the assignment of their absolute stereochemistry. This set of NPs contains a number of different chemical scaffolds highlighted as potential anti-infectives that are likely to have a range of different mechanisms of action.