Aquaculture is currently the fastest growing animal production sector. Because the aquaculture sector is growing at rapid rates, certain materials for feed, specifically marine protein sources, are becoming increasingly expensive and unsustainable. To counteract the reliance on fishmeal (FM) and other marine protein sources in the industry plant protein (PP), specifically soybean meal (SBM), has been investigated to replace FM as a protein source. Unfortunately, SBM when given in high quantities (greater than 30%) has been shown to negatively affect fish performance including retarded growth, intestinal inflammation, reduction of spawn quality, as well as dysbiosis in the gut microbiome, most likely due to presence of antinutritional factors such as saponins and tannins in SBM. The goal of this thesis was to investigate the effect of nutritional programming (NP) with SBM-based diet on gut microbiota in broodstock and progeny fish. Three feedings trials were conducted to test the efficacy of 3 approaches towards improving the use of PP in fish.The first trial (Chapter 2), tested the effect of NP on larval zebrafish (Danio rerio). NP is the theory of introducing an early nutritional stimulus to an animal that will ‘program’ the animal to better utilize the stimuli later in its adult life. The zebrafish were programmed in their larval stages, and the trial lasted for 65 days. There was a significant effect on growth performance for the programmed group (NP-PP) in terms of weight gains, as the NP-PP group grew better compared to the non-programmed group (NP-FM) and negative control (-control). There was no significant effect on the gut microbiome in terms of alpha or beta diversity, however, there were significant changes in the relative abundance (RA) of the gut microbiome throughout time in the NP-PP and the NP-FM groups. The findings of the study support that early NP of zebrafish with SBM improves growth performance on PP diet, but the gut microbiome does not seem to be a mechanism for NP.The second feeding trial (Chapter 3) focused on NP induced in the zebrafish broodstock with dietary SBM. For two weeks, the broodstock fish were fed with either a SBM diet or a FM diet so that gametogenesis occurred with either a FM or PP diet. This phase was called the ‘broodstock programming’ stage. The broodstock were then spawned, and the larval fish were separated into four different treatments: 1) SBM broodstock progeny, fed SBM for the entire trial (PPBS-PP) 2) SBM broodstock progeny fed FM the entire trial (PPBS-FM), 3) FM broodstock progeny fed FM the entire trial (+ control, FMBS-FM), and 4) FM broodstock progeny fed SBM the entire trial (- control, FMBS-PP). The PPBS-PP group achieved similar weight gains compared to all other treatments in terms of grams, but was numerically greater than the FMBS-PP treatment. There were no differences detected in gut microbiome alpha or beta diversity in any of the groups, however, there was significant change observed of certain bacterial phyla between the ‘programmed broodstock’, larval fish, and fish at the end of the trial, 48 days post hatch. Overall, this trial suggests that parental programming does not improve PP utilization in the progeny of zebrafish. It also appears that the gut microbiome is not a mechanism of parental programming. The third feeding trial (Chapter 4), was conducted on largemouth bass (Micropterus salmoides). This chapter had a similar experimental design as the first trial (Chapter 2), and larval largemouth bass were programmed with dietary SBM. This trial had an additional group added to it, which included a dietary saponin-programmed group. The study found that the NP with SBM diet or dietary saponin did not improve PP utilization and growth performance of largemouth bass in its pre-adult age. The study also found that the NP with SBM diet or dietary saponin did not have any effect on the largemouth bass gut microbiome, and there does not seem to be any gut microbiome modification associated with the NP in this fish species. Overall, NP can be used to improve dietary PP utilization but optimal timing and PP delivery method must be well assessed to ensure successful PP exposure and adaptation in different species. Nevertheless, the gut microbiome does not seem to be affected by NP and therefore is not considered the mechanism behind NP. Finally, studies on both zebrafish and largemouth bass presented major shifts in the gut microbiome as the fish aged. In addition, the core microbiomes of both species appeared to become more pronounced as the fish become adults. There seem to be an evolutionary tie between host and its gut microbiome. More studies, however, should further investigate this and the genetic effects on gut microbiota development and its heritability.

No description available.

We extended the parallel skeleton programming framework SkePU 2 with a new back-end utilizing StarPU, a task programming framework for hybrid and distributed architectures. The aim was to allow SkePU to run on distributed clusters, using MPI through StarPU. The implemented back-end distributes data and work across participating ranks. While we did not implement the full SkePU API, the Map and Reduce1D skeletons were successfully implemented. During the implementation, we discovered some differences in API design between SkePU and StarPU. We combine the type-safe templates used in the SkePU API with the C-style void*-heavy API of StarPU. This requires the implementation to use more complex templates than normally desired. While we could preserve most of the SkePU 2 API when moving to a distributed memory situation, some parts had to change. In particular, we needed to change the semantics of SkePU 2 containers with regards to iterators and random access. We benchmarked the performance of the implemented back-end against an MPI+OpenMP reference implementation on two problems, n-body and a simple reduction. While the n-body problem demonstrates promising scaling properties, reductions do not scale well to larger number of ranks. A performance comparison against the MPI+OpenMP reference implementation reveals that, aside from the higher communication overhead, there may also be some overhead in the work performed between communications, potentially performing at below 60-70% of the reference. In most cases, the new back-end to SkePU exhibits significantly lower performance than the reference. Extending the implemented solution to cover the full API and improving performance could provide a high level interface to distributed programming for application programmers. Indeed, subsequent developments of SkePU 3 extend and improve our StarPU back-end.