Featured Joint Publications

Atmospheric Physics, Informatics & Bioinformatics

Automatic shape detection of ice crystals. L. Grulich, R. Weigel, A. Hildebrandt, M. Wand, and P. Spichtinger. Journal of Computational Science, 54, 101429 (2021).
doi: 10.1016/j.jocs.2021.101429

Clouds have a crucial impact on the energy budget of the Earth Atmosphere system; they can cool the system by partly scattering solar radiation or warm the system by absorption and re-emission of thermal emission of the Earth. For clouds consisting exclusively of ice particles, both effects are of the same order of magnitude but have different signs. Thus, for these clouds details of the ice crystals, as e.g. the shape, become more important for the estimation of the climate effect of ice clouds. Ice crystals can be measured with optical array probes (OAPs) onboard research aircraft, producing a huge amount of 2D shadow images. In this study we present two new computational analysis methods,  combined in an hybrid approach, for the automatic shape detection of ice crystals. The first method computes the principal components of a cloud particle and uses them to determine an ellipse, which can then be used to filter for spherical particles. The second method uses convolutional neural networks for the classification of columns and rosettes, respectively. The evaluation of the methods in a field test leads to precision better than 81% with values up to 98%, demonstrating that the new methods are suitable for providing profound shape classifications of cloud particle images obtained by OAP measurements.

Bioinformatics & Physics

Low complexity induces structure in protein regions predicted as intrinsically disordered. Gonçalves-Kulik, M., P. Mier, K. Kastano, J. Cortés, P. Bernadó, F. Schmid and M.A. Andrade-Navarro. Biomolecules. 12, 1098 (2022).
PMID=36008992; doi:10.3390/biom12081098

Intrinsically disordered regions in proteins (IDRs) are predicted as disordered but they often contain low complexity regions (LCRs), which we hypothesized could provide them with structural propensity. To test this, we obtained the experimental structures of human proteins (or homologs) and observed that LCRs (formed by one or two amino acids, polyX and polyXY, respectively) within IDRs are more often represented in structures than other parts of IDRs. Glu and Gly were the amino acids more often observed in these LCRs and polyEK was found to induce alpha helical conformation, with coils as the most frequent structure, although beta-strands were also observed.

 

Reliability of genomic variants across different next-generation sequencing platforms and bioinformatic processing pipelines. S, Weißbach, S. Sys, C. Hewel1, H. Todorov, S. Schweiger, J. Winter, M. Pfenninger, A. Torkamani, D. Evans, J. Burger, K. Everschor-Sitte, H. May-Simera, and S. Gerber. BMC Genomics 22, 62 (2021).
https://doi.org/10.1186/s12864-020-07362-8

Next Generation Sequencing (NGS) is the fundament of various studies, providing insights into questions from biology and medicine. Nevertheless, integrating data from different experimental backgrounds can introduce strong biases. In order to methodically investigate the magnitude of systematic errors in single nucleotide variant calls, we performed a cross-sectional observational study on a genomic cohort of 99 subjects each sequenced via (i) Illumina HiSeq X, (ii) Illumina HiSeq, and (iii) Complete Genomics and processed with the respective bioinformatic pipeline. We also repeated variant calling for the Illumina cohorts with GATK, which allowed us to investigate the effect of the bioinformatics analysis strategy separately from the sequencing platform's impact. We provide empirical evidence of systematic heterogeneity in variant calls between alternative experimental and data analysis setups. Furthermore, our results demonstrate the benefit of reprocessing genomic data with harmonized pipelines when integrating data from different studies.

Geological Sciences & Mathematics

Investigating the effects of intersection flow localization in equivalent-continuum-based upscaling of flow in discrete fracture networks. M.O. Kottwitz, A.A. Popov, S. Abe, B.J.P. Kaus. Solid Earth 12, 2235–2254 (2021).
https://doi.org/10.5194/se-12-2235-2021

Predicting effective permeabilities of fractured rock masses is a crucial component of reservoir modeling. Its often realized with the discrete fracture network (DFN) method, whereby single-phase incompressible fluid flow is modeled in discrete representations of individual fractures in a network. Depending on the overall number of fractures, this can result in high computational costs. Equivalent con- tinuum models (ECMs) provide an alternative approach by subdividing the fracture network into a grid of continuous medium cells, over which hydraulic properties are averaged for fluid flow simulations. While continuum methods have the advantage of lower computational costs and the possi- bility of including matrix properties, choosing the right cell size to discretize the fracture network into an ECM is cru- cial to provide accurate flow results and conserve anisotropic flow properties. Whereas several techniques exist to map a fracture network onto a grid of continuum cells, the com- plexity related to flow in fracture intersections is often ig- nored. Here, numerical simulations of Stokes flow in simple fracture intersections are utilized to analyze their effect on permeability. It is demonstrated that intersection lineaments oriented parallel to the principal direction of flow increase permeability in a process we term intersection flow localiza- tion (IFL). We propose a new method to generate ECMs that includes this effect with a directional pipe flow parameteriza- tion: the fracture-and-pipe model. Our approach is compared against an ECM method that does not take IFL into account by performing ECM-based upscaling with a massively paral- lelized Darcy flow solver capable of representing permeability anisotropy for individual grid cells. While IFL results in an increase in permeability at the local scale of the ECM cell (fracture scale), its effects on network-scale flow are minor. We investigated the effects of IFL for test cases with orthog- onal fracture formations for various scales, fracture lengths, hydraulic apertures, and fracture densities. Only for global fracture porosities above 30 % does IFL start to increase the systems permeability. For lower fracture densities, the effects of IFL are smeared out in the upscaling process. However, we noticed a strong dependency of ECM-based upscaling on its grid resolution. Resolution tests suggests that, as long as the cell size is smaller than the minimal fracture length and larger than the maximal hydraulic aperture of the considered fracture network, the resulting effective permeabilities and anisotropies are resolution-independent. Within that range, ECMs are applicable to upscale flow in fracture networks.

---
Statistical and deterministic inverse methods in the geosciences: introduction, review, and application to the nonlinear diffusion equation. G.S. Reuber. GEM International Journal on Geomathematics 12:19 (2021).
https://doi.org/10.1007/s13137-021-00186-y

Using one and two dimensional steady and advective nonlinear diffusion equation as examples, both, statistical and deterministic inversion approaches are explored. Inversions for material parameters, including but not limited to pure geometrical reconstructions or pointwise parameter updates, are presented. Following a gentle theoretical introduction to the inverse problems multiple algorithms, ranging from basic sampling to Hamiltonian Monte Carlo Markov Chains, are explored in terms of their implementation, are of application and performance. Additionally, a discussion of two sensitivity analysis methods using derivative information to gain deeper insights into the dynamics of the forwards problem is given.

---
Inferring rheology and geometry of subsurface structures by adjoint-based inversion of principal stress directions. G.S. Reuber, L. Holbach, A.A. Popov, M. Hanke, and B.J.P. Kaus. Geophys. J. International 223.2 (2020): 851-861.
https://doi.org/10.1093/gji/ggaa344

Imaging subsurface structures, such as salt domes, magma reservoirs or subducting plates, is a major challenge in geophysics. Seismic imaging methods are, so far, the most precise methods to open a window into the Earth. However, the methods may not yield the exact depth or size of the imaged feature and may become distorted by phenomena such as seismic anisotropy, fluid flow, or compositional variations. A useful complementary method is therefore to simulate the mechanical behaviour of rocks on large timescales, and compare model predictions with observations. Recent studies have used the (non-linear) Stokes equations and geometries from seismic studies in combination with an adjoint-based approach to invert for rheological parameters that are consistent with surface observations such as GPS velocities. Nevertheless, it would be useful to use other surface observations, such as principal stress directions, as constraints as well. Here, we derive the adjoint formulation for the case that principal stress directions are used as observables with respect to rheological parameters. Both an algebraic and a discretized derivation of the adjoint equations are described. This thus enables the usage of two data fields - surface velocities and stress directions - as a misfit for the inversion. We test the performance of the inversion for principal stress directions on simplified 3-D test cases.
Finally, we demonstrate how the adjoint approach can be used to compute 3-D geodynamic sensitivity kernels, which highlight the areas in the model domain that have the largest impact on the misfit value of a particular point. This provides a simple, yet powerful, way to visualize which parts of the model domain are of key importance if changing rheological constants.