Polyhydroxyalkanotes (PHAs) are naturally occurring polymers synthesized by a wide range of microorganisms. Their physiological role is to act as carbon and energy reserves, and their mechanical and physical properties are similar to those of petrochemical plastics. PHAs can be synthesized from renewable materials and they are biodegradable. Considering these properties, PHAs are commonly known as bioplastics. Commercial processes for the production of PHA are based on the use of pure cultures of microorganisms and pure substrates, leading to a very high price of PHA compared to conventional petroleum-based plastics. During the last years, extensive research has been conducted with the aim of developing a low-cost PHA production process. To lower the costs associated to the production of PHA the use of bacterial enrichment cultures combined with the use of waste as substrate has been proposed. The conceptual idea of this approach is to naturally select for microorganisms with the ability to produce PHA, thus preventing the need of sterile equipment. The use of waste as substrate will considerably diminish, and even eliminate, costs associated to the raw materials. In addition, cheaper equipment will be needed. So far, the production of PHA by bacterial enrichment cultures has only been studied using fermentative products, mainly volatile fatty acids (VFAs), as substrate. The production strategy is based on a two-step process consisting on (1) the selection of a bacterial enrichment culture with the ability to produce PHA, and (2) the maximization of the PHA content of the enrichment selected in the first step. The selection step is based on the supply of a selective pressure in the form of alternating periods of presence and absence of the carbon source under fully aerobic conditions (feast-famine regime). During the second step, the presence of an essential nutrient is limited in order to prevent growth and maximize biopolymer content by forcing the conversion of the substrate into PHA. The first aim of the research described in this thesis was to test the feasibility of using a non-fermented substrate, glycerol, for the production of PHA by bacterial enrichment cultures. Glycerol was chosen because of its large availability as waste derived from the biofuel industry. Biodiesel production generates large amounts of a side stream that contains up to 70 wt % of glycerol. The suitability of the selective pressure applied when VFAs were used as substrate was tested with glycerol. Since the main goal of this research was to understand the mechanisms related to the production of PHA using non-fermented substrates, a synthetic medium was used rather than industrial waste. As a proof of concept, it was demonstrated that glycerol can be used as substrate for PHA production applying the same strategy as with the extensively studied VFAs (Chapter 2). Up to 80 wt % of PHA after the accumulation step was attained, which is comparable with the highest accumulation values reported in the literature when acids were used as substrates. PHA was found in the form of polyhydroxybutyrate, PHB. An outstanding fact was that, besides PHB, a polymer of glucose, referred to as polyglucose (PG), was simultaneously produced. The results obtained in Chapter 2 settled the next step in the research: to study the regulation between PHB and PG production. The metabolic pathways involved in glycerol transformation into PG and PHB were examined. This analysis showed that the PG production pathway requires less oxygen respiration than PHB production. The hypothesis that the dissolved oxygen (DO) concentration can have a major impact on the type of storage polymer produced was studied (Chapter 3). The effect of the DO concentration was studied on the maximal PHB storage capacity of the community previously established under a feast-famine regime with no oxygen limiting conditions, and during the enrichment step in a feast-famine regime with low oxygen supply. It was found that decreasing DO concentration during the accumulation step of the enrichment selected in excess of oxygen did not significantly affect glycerol partitioning among PHB and PG, although conversion rates were lower. The limitation of oxygen during the enrichment step favoured PG production over PHB. Worse overall accumulation performances were found when DO was limited during the selection step. When glycerol was used as substrate for the enrichment of a biopolymer-producing community in a feast-famine system, the limitation of the oxygen supply lead to lower accumulation capacities compared to biomass enriched in excess of oxygen. The bacterial composition of the two enrichments was examined and compared, and different microorganisms were dominant in each system. The production of PHB and PG in activated sludge systems had been studied using mixtures of substrates. It had been reported that the kinetics associated to the production of these two polymers, from different substrates, were markedly different. This phenomena was also observed in the enrichments studied in the present work when these polymers were synthesized from a single substrate, i.e. glycerol. Based on the existing differences in the polymer production and degradation kinetics, the effect of the cycle length on the selection and performance of the bacterial enrichment culture, when the hydraulic (HRT) and solid retention times (SRT) remained unaltered, was studied. The selective pressure, based on the feast-famine regime, was changed in two different enrichments: one operated with cycles of 24 h length, and a second reactor, operated with cycles of 6 h length (Chapter 4). The modification of the cycle length while maintaining the HRT and SRT had a direct impact on the food to microorganism (F/M) ratio in the reactors. It was found that the operation at shorter cycles favoured PG production over PHB: the observed yield of PHB during the feast phase decreased from 0.34 g PHB/g glycerol in the 24 h cycle to 0.20 g PHB/glycerol in the 6h cycle, whereas an increase from 0.17 to 0.55 g PG/g glycerol was registered from the 24 h cycle to the 6h cycle. In addition, the microbial community structure present in both enrichments was compared to check if the enrichment strategy influenced the dominant microorganism that was selected or if different metabolic responses were developed by the same population. A species hybridising with Defluviicoccus vanus FISH probe was found to be present, and dominant, in both systems. Despite the fact that the dominant microorganism did not change, the number of cycles/SRT, and thus the F/M ratio, exerted a direct impact on the type of polymer that was produced. Chapter 5 examines the effect of significantly changing the composition or the nature of the substrate between the enrichment and the accumulation step. Different substrates were supplied during the accumulation to a glycerol-grown bacterial enrichment culture. The biopolymer production performance was tested using glycerol, glucose, acetate, xylitol and lactate as substrate. Apart from xylitol, that did not show good results, the accumulation capacity with the rest of the tested substrates was high, between 40-70 wt % PHB. When glucose or glycerol was used, apart from PHB, also PG was produced. The use of glycerol as substrate allowed for the selection of a bacterial community with the ability to cope with different substrates. Considering the intrinsic changing nature of wastewaters, to know the effect of changes in substrate composition is very interesting for the establishment of a waste-based biopolymer production process. In conclusion, the feasibility of producing PHA by bacterial enrichment cultures using non-fermented substrates has been proven in this thesis. The scarce knowledge about the use of non-fermented substrates has been considerably improved. Herewith, the development of sustainable biopolymer production processes will enhance the shift towards a bio-based economy.