1. Introduction Fermented foods are complex biological ecosystems, harbouring diverse microbial communities which contribute to the quality of these food products. In most industrial processes, these communities are advisedly chosen as starter culture for their adaptation to: (i) the substrates to be fermented (milk, vegetable) (ii) the process (°C, O2, NaCl) and (iii) the desired overall quality of the food product in terms of texture, flavour, nutrition and health benefits. For this purpose, technological properties have been mainly assessed by a time-consuming in vitro screening. We hypothesis that the huge amount of bibliographic and microbial genomic data available in public databases would enable an in silico preselection of the strains of interest before wet laboratory screening experiments. 2. Objectives Our aim was to provide (i) adapted bioinformatic tools to mine the diversity of microbial technological properties and metabolisms (ii) a strategy to preselect strains based on their phenotypes and on their genetic potential (sugar utilization, aroma production). 3. Material and Methods We used a bioanalysis approach to provide insight into the presence of enzyme-encoding genes responsible for targeted technological properties. KEGG, Metacyc, NCBI and CAZY databases were queried by a search on enzyme numbers, names and sequence alignments. We developed Florilege (http://migale.jouy.inra.fr/Florilege/), a database of microbial phenotypes based on text-mining to gather microbial growth conditions. 4. Results Combining both approaches, we set up a bioinformatic preselection of lactic acid bacteria able to degrade sugar from milk, cereal or legume based juice to develop yogurt-like products. Database queries provided a list of microbial species with genes encoding key targeted enzymes. A set of preselected anaerobic, mesophilic and thermophilic strains has been successfully screened on different milk and juice in our wet laboratory providing potential new starter strains. 5. Conclusion The developed tools and the strategy can be usefully applied to other domains like bioremediation and white biotechnology.