Approximate parameter inference in systems biology using gradient matching: a comparative evaluation

Background

A challenging problem in current systems biology is that of parameter inference in biological pathways expressed as coupled ordinary differential equations (ODEs). Conventional methods that repeatedly numerically solve the ODEs have large associated computational costs. Aimed at reducing this cost, new concepts using gradient matching have been proposed, which bypass the need for numerical integration. This paper presents a recently established adaptive gradient matching approach, using Gaussian processes (GPs), combined with a parallel tempering scheme, and conducts a comparative evaluation with current state-of-the-art methods used for parameter inference in ODEs. Among these contemporary methods is a technique based on reproducing kernel Hilbert spaces (RKHS). This has previously shown promising results for parameter estimation, but under lax experimental settings. We look at a range of scenarios to test the robustness of this method. We also change the approach of inferring the penalty parameter from AIC to cross validation to improve the stability of the method.

Methods

Methodology for the recently proposed adaptive gradient matching method using GPs, upon which we build our new method, is provided. Details of a competing method using RKHS are also described here.

Results

We conduct a comparative analysis for the methods described in this paper, using two benchmark ODE systems. The analyses are repeated under different experimental settings, to observe the sensitivity of the techniques.

Conclusions

Our study reveals that for known noise variance, our proposed method based on GPs and parallel tempering achieves overall the best performance. When the noise variance is unknown, the RKHS method proves to be more robust.

     

Validated numerical fluid simulation of a thin‐layer cascade photobioreactor in OpenFOAM

Recently, microalgae have been considered as a promising alternative for the production of biofuels from CO₂. For the efficient cultivation of these microalgae, several types of photobioreactors have been designed and Pilot scale photobioreactors have been used to assess the performance of these reactors. Therein the primarily investigated reactor type is the Raceway Pond. However, the less researched Thin‐Layer Cascade Photobioreactor (TLC) shows a high potential for efficient production processes. Unfortunately, for low‐value products like biofuels costs must be kept to a minimum for an economic operation. To facilitate this, 3D Computational Fluid Dynamic simulations can be employed to estimate performance of reactor variants e.g. with respect to power input and mixing. Since up to now little effort has been put into the modelling of TLC reactors, this report aims to present a simulation approach for these reactors types that allows simple adaptation to different geometric or operational boundary conditions. All models have been generated for a two‐phase mixture in OpenFOAM. To demonstrate its applicability, validation measurements with a physical unit have been performed and were compared to the simulation results. With errors in the order of 10 % a successful simulation of the reactor geometry could be proven.     

Same but different–Scale up and numbering up in electrobiotechnology and photobiotechnology

Facing energy problems, there is a strong demand for new technologies dealing with the replacement of fossil fuels. The emerging fields of biotechnology, photobiotechnology and electrobiotechnology, offer solutions for the production of fuels, energy, or chemicals using renewable energy sources (light or electrical current e.g. produced by wind or solar power) or organic (waste) substrates. From an engineering point of view both technologies have analogies and some similar challenges, since both light and electron transfer are primarily surface‐dependent. In contrast to that, bioproduction processes are typically volume dependent. To allow large scale and industrially relevant applications of photobiotechnology and electrobiotechnology, this opinion first gives an overview over the current scales reached in these areas. We then try to point out the challenges and possible methods for the scale up or numbering up of the reactors used. It is shown that the field of photobiotechnology is by now much more advanced than electrobiotechnology and has achieved industrial applications in some cases. We argue that transferring knowledge from photobiotechnology to electrobiotechnology can speed up the development of the emerging field of electrobiotechnology. We believe that a combination of scale up and numbering up, as it has been shown for several photobiotechnological reactors, may well lead to industrially relevant scales in electrobiotechnological processes allowing an industrial application of the technology in near future.     

Bidirectional Propagation of Signals and Nutrients in Fungal Networks via Specialized Hyphae

1. Download high-res image (131KB)
2. Download full-size image

     

A gradient‐forming MipZ protein mediating the control of cell division in the magnetotactic bacterium Magnetospirillum gryphiswaldense

Cell division needs to be tightly regulated and closely coordinated with other cellular processes to ensure the generation of fully viable offspring. Here, we investigate division site placement by the cell division regulator MipZ in the alphaproteobacterium Magnetospirillum gryphiswaldense, a species that forms linear chains of magnetosomes to navigate within the geomagnetic field. We show that M. gryphiswaldense contains two MipZ homologs, termed MipZ1 and MipZ2. MipZ2 localizes to the division site, but its absence does not cause any obvious phenotype. MipZ1, by contrast, forms a dynamic bipolar gradient, and its deletion or overproduction cause cell filamentation, suggesting an important role in cell division. The monomeric form of MipZ1 interacts with the chromosome partitioning protein ParB, whereas its ATP‐dependent dimeric form shows non‐specific DNA‐binding activity. Notably, both the dimeric and, to a lesser extent, the monomeric form inhibit FtsZ polymerization in vitro. MipZ1 thus represents a canonical gradient‐forming MipZ homolog that critically contributes to the spatiotemporal control of FtsZ ring formation. Collectively, our findings add to the view that the regulatory role of MipZ proteins in cell division is conserved among many alphaproteobacteria. However, their number and biochemical properties may have adapted to the specific needs of the host organism.     

Synthesis and antimicrobial activities of N6-hydroxyagelasine analogs and revision of the structure of ageloximes

(+)-N⁶-Hydroxyagelasine D, the enantiomer of the proposed structure of (?)-ageloxime D, as well as N⁶-hydroxyagelasine analogs were synthesized by selective N-7 alkylation of N⁶-[tert-butyl(dimethyl)silyloxy]-9-methyl-9H-purin-6-amine in order to install the terpenoid side chain, followed by fluoride mediated removal of the TBDMS-protecting group. N⁶-Hydroxyagelasine D and the analog carrying a geranylgeranyl side chain displayed profound antimicrobial activities against several pathogenic bacteria and protozoa and inhibited bacterial biofilm formation. However these compounds were also toxic towards mammalian fibroblast cells (MRC-5). The spectral data of N⁶-hydroxyagelasine D did not match those reported for ageloxime D before. Hence, a revised structure of ageloxime D was proposed. Basic hydrolysis of agelasine D gave (+)-N-[4-amino-6-(methylamino)pyrimidin-5-yl]-N-copalylformamide, a compound with spectral data in full agreement with those reported for (?)-ageloxime D.     

Lipidomic Analysis of α-Synuclein Neurotoxicity Identifies Stearoyl CoA Desaturase as a Target for Parkinson Treatment

1. Download high-res image (214KB)
2. Download full-size image