| 1. |
- Bellisario, Alfredo
(författare)
-
Deep learning assisted phase retrieval and computational methods in coherent diffractive imaging
- 2024
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- In recent years, advances in Artificial Intelligence and experimental techniques have revolutionized the field of structural biology. X-ray crystallography and Cryo-EM have provided unprecedented insights into the structures of biomolecules, while the unexpected success of AlphaFold has opened up new avenues of investigation. However, studying the dynamics of proteins at high resolution remains a significant obstacle, especially for fast dynamics. Single-particle imaging (SPI) or Flash X-ray Imaging (FXI) is an emerging technique that may enable the mapping of the conformational landscape of biological molecules at high resolution and fast time scale. This thesis discusses the potential of SPI/FXI, its challenges, recent experimental successes, and the advancements driving its development. In particular, machine learning and neural networks could play a vital role in fostering data analysis and improving SPI/FXI data processing. In Paper I, we discuss the problem of noise and detector masks in collecting FXI data. I simulated a dataset of diffraction patterns and used it to train a Convolutional Neural Network (U-Net) to restore data by denoising and filling in detector masks. As a natural continuation of this work, I trained another machine learning model in Paper II to estimate 2D protein densities from diffraction intensities. In the final chapter, corresponding to Paper III, we discuss another experimental method, time-resolved Small Angle X-ray Scattering (SAXS), and a new algorithm recently developed for SAXS data, the DENsity from Solution Scattering (DENSS) algorithm. I discuss the potential of DENSS in time-resolved SAXS and its application for structural fitting of AsLOV2, a Light-Oxygen-Voltage (LOV) protein domain from Avena sativa.
|
|
| 2. |
- Bellisario, Alfredo, et al.
(författare)
-
Deep learning phase retrieval in X-ray single-particle imaging and object support from autocorrelations
-
Annan publikation (övrigt vetenskapligt/konstnärligt)abstract
- Phase retrieval is an important optimization problem that occurs, for example, in the analysis of coherent diffraction patterns from isolated proteins. All iterative algorithms employed for phase retrieval in this context require some a priori knowledge of the object, usually in the form of a support that describes the extent of the particle. Phase retrieval is a time-consuming task that can often fail, particularly if the support is too loose or of bad quality. In this paper, we present a neural network that can produce low-resolution estimates of the phased object in a fraction of the time that it takes for a full phase retrieval and that can also successfully be used as support for further analysis. Our network is trained on simulated data from biological macromolecules and is thus tailored to the type of data seen in a typical CDI experiment. Other approaches to support finding either require very accurate data without missing regions or require the full phase-retrieval algorithm to be run for a long time. Our network could both speed up off-line analysis, and provide real-time feedback during data collection.
|
|
| 3. |
- Bellisario, Alfredo, et al.
(författare)
-
Noise reduction and mask removal neural network for X-ray single-particle imaging
- 2022
-
Ingår i: Journal of applied crystallography. - : International Union of Crystallography (IUCr). - 0021-8898 .- 1600-5767. ; 55, s. 122-132
-
Tidskriftsartikel (refereegranskat)abstract
- Free-electron lasers could enable X-ray imaging of single biological macro-molecules and the study of protein dynamics, paving the way for a powerful new imaging tool in structural biology, but a low signal-to-noise ratio and missing regions in the detectors, colloquially termed 'masks', affect data collection and hamper real-time evaluation of experimental data. In this article, the challenges posed by noise and masks are tackled by introducing a neural network pipeline that aims to restore diffraction intensities. For training and testing of the model, a data set of diffraction patterns was simulated from 10 900 different proteins with molecular weights within the range of 10-100 kDa and collected at a photon energy of 8 keV. The method is compared with a simple low-pass filtering algorithm based on autocorrelation constraints. The results show an improvement in the mean-squared error of roughly two orders of magnitude in the presence of masks compared with the noisy data. The algorithm was also tested at increasing mask width, leading to the conclusion that demasking can achieve good results when the mask is smaller than half of the central speckle of the pattern. The results highlight the competitiveness of this model for data processing and the feasibility of restoring diffraction intensities from unknown structures in real time using deep learning methods. Finally, an example is shown of this preprocessing making orientation recovery more reliable, especially for data sets containing very few patterns, using the expansion-maximization-compression algorithm.
|
|
| 4. |
- Konold, Patrick E., et al.
(författare)
-
3D-printed sheet jet for stable megahertz liquid sample delivery at X-ray free-electron lasers
- 2023
-
Ingår i: IUCrJ. - : International Union Of Crystallography. - 2052-2525. ; 10, s. 662-670
-
Tidskriftsartikel (refereegranskat)abstract
- X-ray free-electron lasers (XFELs) can probe chemical and biological reactions as they unfold with unprecedented spatial and temporal resolution. A principal challenge in this pursuit involves the delivery of samples to the X-ray interaction point in such a way that produces data of the highest possible quality and with maximal efficiency. This is hampered by intrinsic constraints posed by the light source and operation within a beamline environment. For liquid samples, the solution typically involves some form of high-speed liquid jet, capable of keeping up with the rate of X-ray pulses. However, conventional jets are not ideal because of radiation-induced explosions of the jet, as well as their cylindrical geometry combined with the X-ray pointing instability of many beamlines which causes the interaction volume to differ for every pulse. This complicates data analysis and contributes to measurement errors. An alternative geometry is a liquid sheet jet which, with its constant thickness over large areas, eliminates the problems related to X-ray pointing. Since liquid sheets can be made very thin, the radiation-induced explosion is reduced, boosting their stability. These are especially attractive for experiments which benefit from small interaction volumes such as fluctuation X-ray scattering and several types of spectroscopy. Although their use has increased for soft X-ray applications in recent years, there has not yet been wide-scale adoption at XFELs. Here, gas-accelerated liquid sheet jet sample injection is demonstrated at the European XFEL SPB/SFX nano focus beamline. Its performance relative to a conventional liquid jet is evaluated and superior performance across several key factors has been found. This includes a thickness profile ranging from hundreds of nanometres to 60 nm, a fourfold increase in background stability and favorable radiation-induced explosion dynamics at high repetition rates up to 1.13 MHz. Its minute thickness also suggests that ultrafast single-particle solution scattering is a possibility.
|
|
| 5. |
- Konold, Patrick, et al.
(författare)
-
Microsecond time-resolved X-ray scattering by utilizing MHz repetition rate at second-generation XFELs
-
Annan publikation (övrigt vetenskapligt/konstnärligt)abstract
- Detecting microsecond structural perturbations in biomolecules has wide relevance inbiology, chemistry, and medicine. Here, we show how MHz repetition rates at X-ray freeelectron lasers (XFELs) can be used to produce microsecond time-series of proteinscattering with exceptionally low noise levels of 0.001%. We demonstrate the approach byderiving new mechanistic insight into Jɑ helix unfolding of a Light-Oxygen-Voltage (LOV)photosensory domain. This time-resolved acquisition strategy is easy to implement andwidely applicable for direct observation of structural dynamics of many biochemicalprocesses.
|
|
| 6. |
- Malpetti, D., et al.
(författare)
-
Multipartite entanglement in qudit hypergraph states
- 2022
-
Ingår i: Journal of Physics A. - : Institute of Physics (IOP). - 1751-8113 .- 1751-8121. ; 55:41
-
Tidskriftsartikel (refereegranskat)abstract
- We study entanglement properties of hypergraph states in arbitrary finite dimension. We compute multipartite entanglement of elementary qudit hypergraph states, namely those endowed with a single maximum-cardinality hyperedge. We show that, analogously to the qubit case, also for arbitrary dimension there exists a lower bound for multipartite entanglement of connected qudit hypergraph states; this is given by the multipartite entanglement of an equal-dimension elementary hypergraph state featuring the same number of qudits as the largest-cardinality hyperedge. We highlight interesting differences between prime and non-prime dimension in the entanglement features.
|
|
