Toward Bridging the Simulated-to-Real Gap: Benchmarking Super-Resolution on Real Data.

Capturing ground truth data to benchmark super-resolution (SR) is challenging. Therefore, current quantitative studies are mainly evaluated on simulated data artificially sampled from ground truth images. We argue that such evaluations overestimate the actual performance of SR methods compared to their behavior on real images. Toward bridging this simulated-to-real gap, we introduce the Super-Resolution Erlangen (SupER) database, the first comprehensive laboratory SR database of all-real acquisitions with pixel-wise ground truth. It consists of more than 80k images of 14 scenes combining different facets: CMOS sensor noise, real sampling at four resolution levels, nine scene motion types, two photometric conditions, and lossy video coding at five levels. As such, the database exceeds existing benchmarks by an order of magnitude in quality and quantity. This paper also benchmarks 19 popular single-image and multi-frame algorithms on our data. The benchmark comprises a quantitative study by exploiting ground truth data and qualitative evaluations in a large-scale observer study. We also rigorously investigate agreements between both evaluations from a statistical perspective. One interesting result is that top-performing methods on simulated data may be surpassed by others on real data. Our insights can spur further algorithm development, and the publicy available dataset can foster future evaluations.

Characterization of Azotobacter vinelandii nifZ deletion strains. Indication of stepwise MoFe protein assembly.

The nifZ gene product (NifZ) of Azotobacter vinelandii has been implicated in MoFe protein maturation. However, its exact function in this process remains largely unknown. Here, we report a detailed biochemical/biophysical characterization of His-tagged MoFe proteins purified from A. vinelandii nifZ and nifZ/nifB deletion strains DJ1182 and YM6A (Delta nifZ and Delta nifZ Delta nifB MoFe proteins, respectively). Our data from EPR, metal, activity, and stability analyses indicate that one alpha beta subunit pair of the Delta nifZ MoFe protein contains a P cluster ([8Fe-7S]) and an iron-molybdenum cofactor (FeMoco) ([Mo-7Fe-9S-X-homocitrate]), whereas the other contains a presumed P cluster precursor, possibly comprising a pair of [4Fe-4S]-like clusters, and a vacant FeMoco site. Likewise, the Delta nifZ Delta nifB MoFe protein has the same composition as the Delta nifZ MoFe protein except for the absence of FeMoco, an effect caused by the deletion of the nifB gene. These results suggest that the MoFe protein is likely assembled stepwise, i.e. one alpha beta subunit pair of the tetrameric MoFe protein is assembled prior to the other, and that NifZ might act as a chaperone in the assembly of the second alpha beta subunit pair by facilitating a conformational rearrangement that is required for the formation of the P cluster through the condensation of two [4Fe-4S]-like clusters. The possibility of NifZ exercising its effect through the Fe protein was ruled out because the Fe proteins from nifZ and nifZ/nifB deletion strains are not defective in their normal functions. However, the detailed mechanism of how NifZ carries out its exact function in MoFe protein maturation awaits further investigation.

Comparison of iron-molybdenum cofactor-deficient nitrogenase MoFe proteins by X-ray absorption spectroscopy: implications for P-cluster biosynthesis.

Nitrogenase, the enzyme system responsible for biological nitrogen fixation, is believed to utilize two unique metalloclusters in catalysis. There is considerable interest in understanding how these metalloclusters are assembled in vivo. It has been presumed that immature iron-molybdenum cofactor-deficient nitrogenase MoFe proteins contain the P-cluster, although no biosynthetic pathway for the assembly of this complex cluster has been identified as yet. Through the comparison by iron K-edge x-ray absorption edge and extended fine structure analyses of cofactor-deficient MoFe proteins resulting from nifH and nifB deletion strains of Azotobacter vinelandii, a novel [Fe-S] cluster is identified in the DeltanifH MoFe protein. The iron-iron scattering displayed by the DeltanifH MoFe protein is more similar to that of a standard [Fe(4)S(4)]-containing protein than that of the DeltanifB MoFe protein, which is shown to contain a "normal" P-cluster. The iron-sulfur scattering of the DeltanifH MoFe protein, however, indicates differences in its cluster from an [Fe(4)S(4)](Cys)(4) site that may be consistent with the presence of either oxygenic or nitrogenic ligation. Based on these results, models for the [Fe-S] center in the DeltanifH MoFe protein are constructed, the most likely of which consist of two separate [Fe(4)S(4)] sites, each with some non-cysteinyl coordination. This type of model suggests that the P-cluster is formed by the condensation of two [Fe(4)S(4)] fragments, possibly concomitant with Fe protein (NifH)-induced conformational change.