Diamond

There are two customizable steps in the OrthoFinder method: (1) the sequence search method and (2) the orthogroup tree inference method. The default option for step 1 is DIAMOND [5 (link)]. The default option for step 2 is DendroBLAST [24 (link)]. The default options are recommended by the authors as they are fast and achieve high accuracy on the Quest for Orthologs benchmarks [1 (link)] (Fig. 4a–d). However, the user is free to substitute any alternative methods for these steps. Currently, supported methods for step 1 include BLAST [4 (link)] and MMseqs2 [6 (link)]. Similarly, any combination of multiple sequence alignment and tree inference method can be substituted in for step 2. For illustrative purposes, the default multiple sequence alignment method is MAFFT [35 (link)] and the default tree inference method is FastTree [25 (link)]; this combination is benchmarked above. It is impossible for the authors to test all possible combinations of multiple sequence alignment and tree inference methods, and the selected methods were chosen because of their speed and scalability characteristics [25 (link), 35 (link)]. OrthoFinder provides flexibility for the user to select their preferred method. More accurate multiple sequence alignment and tree inference methods should give more accurate ortholog inference, and many studies exist comparing the accuracy and runtime characteristics of the available methods [36 (link), 37 (link)]. A user-editable configuration file is provided in JSON format that allows new sequence search, multiple sequence alignment, and tree inference methods to be added to OrthoFinder. To facilitate the trialing of alternative multiple sequence alignment and tree inference methods, OrthoFinder provides the option to restart an existing analysis after the orthogroup inference stage. This skips the requirement to compute the all-vs-all sequence search and orthogroup inference and thus accelerates testing of different internal steps.

On the basis of human gut (stool) metagenomics data from the MetaHIT project, the ten highest mean abundance species (TaxIDs 435591, 515620, 470145, 457412, 657321, 445970, 657317, 585543, 537011, and 43559) with a sequenced genome in the set of representative species defined by Mende et al. (2013) (link) were used for simulation using a previously published metagenomics simulation tool (Mende et al. 2012 (link)). On the basis of these simulated metagenomes, assembly and gene calling was performed using MOCAT2 (Kultima et al. 2016 (link)). The predicted genes were assigned GO terms using alternatively eggNOG-mapper or InterProScan as described below. The underlying gold standard was defined by counting, for each GO, the number of reads which were simulated from a gene annotated with that GO. This vector was compared, using Spearman’s rank correlation, to the predicted functional abundance profile of each sample. Predicted functional profiles were obtained by converting predicted gene abundances in the simulated metagenome (estimated with MOCAT2) to GO term abundances. Only GO terms appearing in either the gold set or the prediction were taken into consideration for the correlation analysis. eggNOG-mapper was executed in DIAMOND mode with the following parameters: “-m diamond-–tax_scope auto-–target_orthologs all –-go_evidence non-electronic –-cpu 20”. InterProScan v5.19-58.0 was configured to used 20 cpu workers and called with parameters –-goterms –-iprlookup -pa.

We constructed sets of fungal proteomes of increasing size for performance testing. Ensembl Genomes was interrogated on 6 November 2017 using its REST API [44 (link)] to identify all available fungal genomes. To achieve an even sampling of species, we selected 1 species per genera and excluded genomes from candidate phyla or phyla with fewer than 3 sequenced genomes. This gave a set of 272 species which were downloaded from the Ensembl FTP site [45 (link)]. We created datasets of increasing size by randomly selecting 4, 8, 16, 32, 64, 128, and 256 species such that the last common ancestor was the same for each dataset. Each dataset was analyzed using a single Intel E5-2640v3 Haswell node (16 cores) on the Oxford University ARCUS-B server using 16 parallel threads for OrthoFinder with DIAMOND (arguments: “-S diamond -t 16 -a 16”). The complete datasets for all analyzed species subsets are available for download from Zenodo at 10.5281/zenodo.1481147. All methods submitted to Quest for Orthologs that provided a user-runnable implementation of the method were tested on the same fungi datasets and the same ARCUS-B server nodes and run in parallel using 16 threads (when supported by the method).

Single crystal sized (0.35, 0.25, 0.20) mm³ was carefully selected to perform its structural analysis by X-ray diffraction. The crystallographic data were collected on a Bruker AXS CCD diffractometer at room temperature using graphite-monochromated Mo Kα radiation (λ = 0.71073 Å). All intensities were corrected for Lorentz, polarization and absorption effects.^{15 (link)} The structural determination procedure was carried out using SHELXS97 program.¹⁶ The structure was solved by direct method and refined with full-matrix least squares methods based on F² using SHELXL97.¹⁷ The space group was determined to be P2₁/n. A total of 54568 reflections were collected in the θ range 2.2–27.5°. In this structure, all non-hydrogen atoms were refined with anisotropic displacement parameters. H-atoms were set in calculated positions and treated as riding on their parent atom with constrained thermal parameters. The final discrepancy factors R₁ and wR₂ are 0.053 and 0.134, respectively. Crystal data of (C₁₂H₁₇N₂)₂ZnBr₄ are given in Table 1. Molecular plots were made with ORTEP^{18 (link)} and Diamond.¹⁹ Atomic coordinates anisotropic displacement parameters, tables for all bond distances, and angles have been deposited at the Cambridge Crystallographic Data Centre (deposition number: CCDC 2090035).

All surface resistance measurements were performed with a 2400 Sourcemeter (Co. Keithley Instruments) with a 4-wire sense mode configuration in combination with a custom made 4-point measuring probe. The cylindrical probe tips were arranged in line with a diameter of 0.8 mm and a distance between the probes of 2.1 mm. A constant current was applied through the outer probe tips to the sample. The voltage drop from the outer to the inner probe tips was used for resistance calculation based on a method described by Schroder et al.⁵⁶ Each measurement was performed with a constant current in the range of 1 μA to 1 mA at 2.1 V for 200 s of measurement time. The average value of 6 measurements at different locations on the sample surface is reported for each of the samples.
Attenuated total reflection Fourier transformed infrared spectroscopy (ATR-FTIR) was recorded with a 1.9 cm⁻¹ spectral resolution on a 670 FT-IR spectrometer (Co. Varian Inc. (now: Agilent Technologies)). The assignment of measured vibrations were supported by DFT-based calculations on a D3(BJ)-BP86-def2-SVP level of theory.^57–61 Deviations to the measured spectra were described by Benavides-Garcia and Monroe.⁶² All calculations were carried out using the ORCA computational chemistry program.^63,64 Raman spectra were recorded on an inVia confocal (Co. Renishaw) with an excitation wavelength of 532 nm, 3 times for each specimen with 20 s exposure time. Hardness testing was performed using a Fischerscope H100C XYp Nanoindenter (Co. Helmut Fischer GmbH) using a Vickers diamond indenter. After contact with the surface, the indenter was approached into specimens at a constant rate of 300.00 mN/60 s until 150 mN of force was reached and withdrawn from the surface at the same rate as loading. At least 12 indentations were performed for each specimen and the average value was reported. Surface roughness measurements and optical imaging were performed using a VK-9700 Color 3D-Laser scanning microscope (Co. Keyence Corporation). For each sample, at least five randomly selected areas of the surface were measured and the surface roughness R_a and surface depth R_z were determined. High-resolution images of the composite material were taken using a scanning electron microscope (SEM, S-3400N, Co. Hitatchi Science Systems, Ltd) and spectral maps for sulfur and phosphorus were prepared using energy dispersive X-ray spectroscopy (EDX). The samples were fractured after storage in liquid nitrogen for at least 3 h and the exposed surface was coated with a thin platinum layer using a high vacuum platinum sputter at low voltage (brittle fractures). High-resolution transmission electron microscopy (FEI, Talos 120C, Co. Thermo Fisher Scientifics) images were taken of selected polymer compounds. Therefore, very thin lamellae were sectioned with a diatome diamond knife (Cryo-Mikrotomy, Co. Reichert-Jung Ultracut E and RMC CR-X Cryoattachment) at a temperature of −120 °C. The freshly microtomed sample surfaces were subsequently measured by AFM (MultiMode 8, Co. Brucker). The ultrathin sections (about 60 nm) were collected and used for TEM measurements. By evaluating the distribution of the added liquid and solid lubricants in the bulk material, the tribological mechanisms leading to self-lubrication will be analyzed.