FLASHIda: Intelligent data acquisition for top-down proteomics that doubles proteoform level identification count

Top-down proteomics (TDP) has gained a lot of interest in biomedical application for detailed analysis and structural characterization of proteoforms. Data-dependent acquisition (DDA) of intact proteins is non-trivial due to the diversity and complex signal of proteoforms. Dedicated acquisition methods thus have the potential to greatly improve TDP. We present FLASHIda, an intelligent online data acquisition algorithm for TDP that ensures the real-time selection of high-quality precursors of diverse proteoforms. FLASHIda combines fast charge deconvolution algorithms and machine learning-based quality assessment for optimal precursor selection. In analysis in E. coli lysates, FLASHIda increased the number of unique proteoform level identifications from 800 to 1,500, or generated a near-identical number of identifications in ⅓ of instrument time when compared to standard DDA mode. Furthermore, FLASHIda enabled sensitive mapping of post translational modifications and detection of chemical adducts. As an extension module to the instrument, FLASHIda can be readily adopted for TDP studies of complex samples to enhance proteoform identification rates.

Multi-Agent Deep Reinforcement Learning for Online 3D Human Poses Estimation

Most multi-view based human pose estimation techniques assume the cameras are fixed. While in dynamic scenes, the cameras should be able to move and seek the best views to avoid occlusions and extract 3D information of the target collaboratively. In this paper, we address the problem of online view selection for a fixed number of cameras to estimate multi-person 3D poses actively. The proposed method exploits a distributed multi-agent based deep reinforcement learning framework, where each camera is modeled as an agent, to optimize the action of all the cameras. An inter-agent communication protocol was developed to transfer the cameras’ relative positions between agents for better collaboration. Experiments on the Panoptic dataset show that our method outperforms other view selection methods by a large margin given an identical number of cameras. To the best of our knowledge, our method is the first to address online active multi-view 3D pose estimation with multi-agent reinforcement learning.

Estimation of the prevalence of respiratory diseases in pigs in north-eastern Poland: Survey of pulmonary lesions in pigs at a slaughterhouse

A total of 29 520 animals, from 164 batches of pigs belonging to an identical number of herds, were involved in the study. The considered population of pigs were limited to the region of north-eastern Poland involving six voivodeships. From each herd, samples of blood were collected to evaluate the antibody titres to Mycoplasma hyopneumoniae, porcine reproductive and respiratory syndrome virus, Aujeszky’s disease virus and swine influenza virus. At an abattoir, the lung lesions of each batch were scored and the enzootic pneumonia-like lesion average value was calculated. Lesions, indicative of enzootic pneumonia, were found in 57.8% of the lungs. For all lungs, the enzootic pneumonia-like lesion average value was 1.74, ranging from 0.42 to 3.56 among the 164 batches. In the examined pig population, 57.8% were considered suffering from swine respiratory disease, the majority of the affected pigs came from the Podlaskie (21.7%) and Greater Poland (17.25%) voivodeships. In the most affected voivodeships, 88.37% and 85.16% of the farms were considered as disease-susceptible for Greater Poland and Podlaskie, respectively. The findings indicate that, in pigs in north-eastern Poland, the major pathogens causing pneumonia-like lesions are Mycoplasma hyopneumoniae (68.9%) and porcine reproductive and respiratory syndrome virus (44%).

White space interpretation in Hermann Rorschach’s inkblot test: An analysis of two male examinees’ responses

Introduction: There are numerous points of controversy over the psychological interpretation of the so-called S responses – namely, the interpretation of white space in the Rorschach test. The aim of my paper was to verify the position held by Charles P. Fonda, who claims that the meaning of S responses depends on the results of the test as a whole. Materials and methods: I have presented the responses of two young men and coded them according to the Comprehensive System developed by John E. Exner. I chose this way of processing data as the most precise, relatively well formalized, and detailed one. The two examinees gave an almost identical number of responses in the whole test and a similar number of S responses, similarly distributed across the cards of the test. Results: I assessed and compared the results of the two examinees in terms of the effectiveness of their cognitive processes, self-control ability, mental resources, social adjustment, and self-esteem. Finally, I looked closely at the contents of a few responses, as in my opinion they symbolically expressed the examinees’ basic problems. Discussion: I tried to avoid describing the examinees’ psyche in nosological terms and to focus on presenting the way in which they experienced the world, as well as on how this experience affected their behavior. Conclusions: I found that in Examinee 1 criticism towards other people may stem from an excessively idealistic attitude to the world and from the ensuing disappointments. Good cognitive functioning, resistance to stress, positive self-esteem, and the socialized emotional sphere make this man’s S responses a sign of creative engagement in the problems encountered rather than a sign of maladjustment. In the case of Examinee 2, S responses can be understood as defiance and a generalized attitude of negativism, which play the role of defense against the excessively complicated, not fully comprehensible, and inhospitable world; these responses may, in their turn, contribute to the intensification of problems and to an increase in maladjustment.

Machine learning in spectral domain

Deep neural networks are usually trained in the space of the nodes, by adjusting the weights of existing links via suitable optimization protocols. We here propose a radically new approach which anchors the learning process to reciprocal space. Specifically, the training acts on the spectral domain and seeks to modify the eigenvalues and eigenvectors of transfer operators in direct space. The proposed method is ductile and can be tailored to return either linear or non-linear classifiers. Adjusting the eigenvalues, when freezing the eigenvectors entries, yields performances that are superior to those attained with standard methods restricted to operate with an identical number of free parameters. To recover a feed-forward architecture in direct space, we have postulated a nested indentation of the eigenvectors. Different non-orthogonal basis could be employed to export the spectral learning to other frameworks, as e.g. reservoir computing.

Comparison of the Risk for Peripheral Vertigo between Physicians and the General Population

Background: Because of the limited information available regarding peripheral vertigo (PV) in physicians, we conducted this study to clarify this issue. Methods: We used Taiwan National Health Insurance Research Database to identify 26,309 physicians and an identical number of general population matched by age and sex. All the participants who had PV before 2007 and residents were excluded. By tracing their medical histories between 2007 and 2013, comparisons of PV risk between physicians and general population and among physicians were performed. Results: Physicians had a significantly lower PV risk than the general population (adjusted odds ratio [AOR]: 0.811; 95% confidence interval [CI]: 0.662–0.994). In comparison among physicians, otolaryngologists had a significantly higher PV risk than other specialties. Physicians who were older or served in local hospitals or clinics had a significantly higher PV risk than physicians in medical centers. Conclusion: Physicians had a significantly lower PV risk than the general population. Better medical knowledge in physicians than in the general population may explain the findings; however, further studies are warranted for elucidating the detailed mechanisms.

Resolving the Equal Number Density Puzzle: Molecular Picture from Simulations of LiCl(aq) and NaCl(aq)

The change in number densities of aqueous solutions of alkali chlorides should be qualitatively predictable. Typically, as cations get bigger the number density of the solution decreases. However, aqueous solutions of lithium and sodium chloride exhibit at ambient conditions practically identical number densities at equal molalities despite different ionic sizes. Here, we provide an atomistic interpretation of this experimentally observed anomalous behavior using molecular dynamics simulations. The obtained results show that rigidity of the Li⁺ first and second solvation shells and associated compromised hydrogen bonding result in practically equal average water densities in the local hydration regions for Li+ and Na+ despite different sizes of the cations. In addition, in more distant regions from the cations, the water densities of these two solutions also coincide. These findings thus provide an atomistic interpretation for matching number densities of LiCl and NaCl solutions. In contrast, the number density differences between NaCl and KCl solutions, as well as between LiCl and KCl solutions behave in a regular fashion with lower number densities of solutions observed for larger cations.

Resolving the Equal Number Density Puzzle: Molecular Picture from Simulations of LiCl(aq) and NaCl(aq)

The change in number densities of aqueous solutions of alkali chlorides should be qualitatively predictable. Typically, as cations get bigger the number density of the solution decreases. However, aqueous solutions of lithium and sodium chloride exhibit at ambient conditions practically identical number densities at equal molalities despite different ionic sizes. Here, we provide an atomistic interpretation of this experimentally observed anomalous behavior using molecular dynamics simulations. The obtained results show that rigidity of the Li⁺ first and second solvation shells and associated compromised hydrogen bonding result in practically equal average water densities in the local hydration regions for Li+ and Na+ despite different sizes of the cations. In addition, in more distant regions from the cations, the water densities of these two solutions also coincide. These findings thus provide an atomistic interpretation for matching number densities of LiCl and NaCl solutions. In contrast, the number density differences between NaCl and KCl solutions, as well as between LiCl and KCl solutions behave in a regular fashion with lower number densities of solutions observed for larger cations.

Comparison of 68Ga-PSMA-11 PET/CT with 11C-acetate PET/CT in re-staging of prostate cancer relapse

Positron emission tomography (PET) imaging is used to localize recurrent disease in prostate cancer (PCa). The tracer 68Ga-PSMA-11 visualizes lesions overexpressing prostate-specific membrane antigen (PSMA), while 11C-acetate visualizes lesions with increased anabolic metabolism. The aim of this study was to compare the performance of PSMA-PET and acetate-PET in re-staging patients with biochemical relapse. Thirty PCa patients with prostate-specific antigen (PSA) relapse after primary curative therapy were prospectively evaluated. PET/CT examinations using 11C-acetate and 68Ga-PSMA-11 were performed. Identified lesions were categorized according to anatomical location and PET measurements were correlated with PSA at time of scan. Tumour lesions showed higher semi-quantitative uptake values on PSMA-PET than acetate-PET. PSMA-PET identified more lesions in 11 patients, fewer lesions in eight patients, and identical number of lesions in 11 patients. This study indicates better diagnostic performance of PSMA-PET, particularly in detecting lymph node (81% vs 60%, p = 0.02) and bone metastasis (95% vs 61%, p = 0.0001) compared to acetate-PET. However, 38% of PSMA-expressing metastases appear to be metabolically inactive and 15% of metabolically active metastases lack PSMA expression. Addition of PET with a metabolic tracer, such as 11C-acetate, might be beneficial before making treatment decisions.

A device for exploring the full angular excitation space - Can more angular projections improve determination of a molecules 3D-orientation in the presence of noise?

In the past, different methods have been presented to determine the 3D orientation of single molecules in a microscopic set-up by excitation polarization modulation. Using linearly polarized illumination from different directions and thereby measuring different 2D projections enables reconstructing the full 3D orientation. Theoretically, two projections suffice for a full 3D orientation determination if the intensities are properly calibrated. If they are not, a third projection will enable unambiguous orientation measurements. The question arises if three projections already contain the maximum information on the 3D orientation when also considering the limited number of available photons and shot noise in an experiment, or if detecting more projections or even continuously changing the projection direction during a measurement provides more information with an identical number of available photons. To answer this principle question, we constructed a simple device allowing for exploring any projection direction available with a particular microscope objective and tested several different excitation modulation schemes using simulated as well as experimental single molecule data. We found that three different projections in fact already do provide the maximum information also for noisy data. Our results do not indicate a significant improvement in angular precision in comparison to three projections, both when increasing the number of projections and when modulating the projection direction and polarization simultaneously during the measurement.In fluorescence microscopy polarized illumination from different directions enables the determination of the 3D orientation of single molecules by combining the 2D information of different projection directions. Ambiguities that emerge when using only two projections can be eliminated using a third projection. In a systematic study we show that – also considering the limited number of available photons and shot noise in an experiment – three projection directions already contain the maximum information on the 3D orientation. Our results do not indicate a significant improvement in angular precision in comparison to three projections, both when increasing the number of projections and when modulating the projection direction and polarization simultaneously during the measurement.