Dynamic change of pH in acidogenic fermentation of cheese whey towards polyhydroxyalkanoates production: Impact on performance and microbial population.

Polyhydroxyalkanoates (PHA) are a sustainable alternative to conventional plastics that can be obtained from industrial wastes/by-products using mixed microbial cultures (MMC). MMC PHA production is commonly carried out in a 3-stage process of acidogenesis, PHA culture selection and accumulation. This research focused on the possibility of tailoring PHA by controlling the acidogenic reactor operating conditions, namely pH, using cheese whey as model feedstock. The objective was to investigate the impact that dynamically varying the acidogenic pH, when targeting different PHA monomer profiles, had on the performance and microbial community profile of the anaerobic reactor. To accomplish this, an anaerobic reactor was continuously operated under dynamic pH changes, ranging from pH 4 to 7, turning to pH 6 after each change of pH. At pH 6, lactate and acetate were the dominant products (41-48% gCOD basis and 22-44% gCOD basis, respectively). At low pH, lactate production was higher while at high pH acetate production was favoured. Despite the dynamic change of pH, the fermentation product composition at pH 6 was always similar, showing the resilience of the process, i.e. when the same pH value was imposed, the culture produced the same metabolic products independently of the history of changes occurring in the system. The different fermentation product fractions led to PHAs of different compositions. The microbial community, analysed by high throughput sequencing of bacterial 16S rRNA gene fragments, was dominated by Lactobacillus, but varied markedly when subjected to the highest and lowest pH values of the tested range (4 and 7), with increase in the abundance of Lactococcus and a member of the Candidate Division TM7. Different bacterial profiles obtained at pH 6 during this dynamic operation were able to produce a consistent profile of fermentation products (and consequently a constant PHA composition), demonstrating the community's functional redundancy.

Dissecting geographic variation in population synchrony using the common vole in central Europe as a test bed.

Spatial synchrony of population fluctuations is ubiquitous in nature. Theoretical models suggest that correlated environmental stochasticity, dispersal, and trophic interactions are important promoters of synchrony in nature to leave characteristic signatures of distance-dependent decays in synchrony. Recent refinements of this theory have clarified how distance-decay curves may steepen if local dynamics are governed by different density-dependent feedbacks and how synchrony should vary regionally if the importance and correlation of environmental stochasticity is location-specific. We analysed spatiotemporal data for the common vole, Microtus arvalis from 49 districts in the Czech Republic to examine the pattern of population synchrony between 2000 and 2014. By extending the nonparametric covariation function, we develop a quantitative method that allows a dissection of the effects of distance and additional variables such as altitude on synchrony. To examine the pattern of local synchrony, we apply the noncentered local-indicators of spatial association (ncLISA) which highlights areas with different degrees of synchrony than expected by the region-wide average. Additionally, in order to understand the obtained pattern of local spatial correlations, we have regressed LISA results against the proportion of forest in each district. The common vole abundances fluctuated strongly and exhibited synchronous dynamics with the typical tendency for a decline of synchrony with increasing distance but, not with altitude. The correlation between the neighbor districts decreases as the proportion of forest increases. Forested areas are suboptimum habitats and are strongly avoided by common voles. The investigation of spatiotemporal dynamics in animal populations is a key issue in ecology. Although the majority of studies are focused on testing hypotheses about which mechanisms are involved in shaping this dynamics it is crucial to understand the sources of variation involved in order to understand the underlying processes.

Response of a three-stage process for PHA production by mixed microbial cultures to feedstock shift: impact on polymer composition.

Polyhydroxyalkanoates (PHA) can be produced by mixed microbial cultures (MMC) using a three-stage process. An attractive feature of MMC for PHA production is the ability to use waste/surplus feedstocks. In this study, the effect of a feedstock shift, mimicking a seasonal feedstock scenario and/or as a strategy for controlling polymer composition, on a MMC PHA production process was assessed using cheese whey (CW) and sugar cane molasses (SCM) as model feedstocks. The acidogenic stage responded immediately to the feedstock shift by changing the fermented products profile, with acetate and butyrate being the main acids produced from CW, while for SCM propionate and valerate were the dominant products. The fermentation process was then quite stable during long term operation. The PHA culture selection stage also responded quickly to the fermented feestocks shift, generating a polymer whose composition was linearly dependent on the concentration of HV and HB precursors produced in the acidogenic stage. The selected culture reached a maximum PHA content of 56% and 65% with fermented SCM and CW, respectively. Mixing fermented CW and SCM, in equal volume proportions, demonstrated the possibility of using different fermented feedstocks for tailoring polymer composition.

FMRI 3D registration based on Fourier space subsets using neural networks.

In this work, we present a neural network (NN) based method designed for 3D rigid-body registration of FMRI time series, which relies on a limited number of Fourier coefficients of the images to be aligned. These coefficients, which are comprised in a small cubic neighborhood located at the first octant of a 3D Fourier space (including the DC component), are then fed into six NN during the learning stage. Each NN yields the estimates of a registration parameter. The proposed method was assessed for 3D rigid-body transformations, using DC neighborhoods of different sizes. The mean absolute registration errors are of approximately 0.030 mm in translations and 0.030 deg in rotations, for the typical motion amplitudes encountered in FMRI studies. The construction of the training set and the learning stage are fast requiring, respectively, 90 s and 1 to 12 s, depending on the number of input and hidden units of the NN. We believe that NN-based approaches to the problem of FMRI registration can be of great interest in the future. For instance, NN relying on limited K-space data (possibly in navigation echoes) can be a valid solution to the problem of prospective (in frame) FMRI registration.