Deep-Manager: a versatile tool for optimal feature selection in live-cell imaging analysis.

One of the major problems in bioimaging, often highly underestimated, is whether features extracted for a discrimination or regression task will remain valid for a broader set of similar experiments or in the presence of unpredictable perturbations during the image acquisition process. Such an issue is even more important when it is addressed in the context of deep learning features due to the lack of a priori known relationship between the black-box descriptors (deep features) and the phenotypic properties of the biological entities under study. In this regard, the widespread use of descriptors, such as those coming from pre-trained Convolutional Neural Networks (CNNs), is hindered by the fact that they are devoid of apparent physical meaning and strongly subjected to unspecific biases, i.e., features that do not depend on the cell phenotypes, but rather on acquisition artifacts, such as brightness or texture changes, focus shifts, autofluorescence or photobleaching. The proposed Deep-Manager software platform offers the possibility to efficiently select those features having lower sensitivity to unspecific disturbances and, at the same time, a high discriminating power. Deep-Manager can be used in the context of both handcrafted and deep features. The unprecedented performances of the method are proven using five different case studies, ranging from selecting handcrafted green fluorescence protein intensity features in chemotherapy-related breast cancer cell death investigation to addressing problems related to the context of Deep Transfer Learning. Deep-Manager, freely available at https://github.com/BEEuniroma2/Deep-Manager , is suitable for use in many fields of bioimaging and is conceived to be constantly upgraded with novel image acquisition perturbations and modalities.

Machine learning phenomics (MLP) combining deep learning with time-lapse-microscopy for monitoring colorectal adenocarcinoma cells gene expression and drug-response.

High-throughput phenotyping is becoming increasingly available thanks to analytical and bioinformatics approaches that enable the use of very high-dimensional data and to the availability of dynamic models that link phenomena across levels: from genes to cells, from cells to organs, and through the whole organism. The combination of phenomics, deep learning, and machine learning represents a strong potential for the phenotypical investigation, leading the way to a more embracing approach, called machine learning phenomics (MLP). In particular, in this work we present a novel MLP platform for phenomics investigation of cancer-cells response to therapy, exploiting and combining the potential of time-lapse microscopy for cell behavior data acquisition and robust deep learning software architectures for the latent phenotypes extraction. A two-step proof of concepts is designed. First, we demonstrate a strict correlation among gene expression and cell phenotype with the aim to identify new biomarkers and targets for tailored therapy in human colorectal cancer onset and progression. Experiments were conducted on human colorectal adenocarcinoma cells (DLD-1) and their profile was compared with an isogenic line in which the expression of LOX-1 transcript was knocked down. In addition, we also evaluate the phenotypic impact of the administration of different doses of an antineoplastic drug over DLD-1 cells. Under the omics paradigm, proteomics results are used to confirm the findings of the experiments.

Accelerating the experimental responses on cell behaviors: a long-term prediction of cell trajectories using Social Generative Adversarial Network.

The incremented uptake provided by time-lapse microscopy in Organ-on-a-Chip (OoC) devices allowed increased attention to the dynamics of the co-cultured systems. However, the amount of information stored in long-time experiments may constitute a serious bottleneck of the experimental pipeline. Forward long-term prediction of cell trajectories may reduce the spatial-temporal burden of video sequences storage. Cell trajectory prediction becomes crucial especially to increase the trustworthiness in software tools designed to conduct a massive analysis of cell behavior under chemical stimuli. To address this task, we transpose here the exploitation of the presence of "social forces" from the human to the cellular level for motion prediction at microscale by adapting the potential of Social Generative Adversarial Network predictors to cell motility. To demonstrate the effectiveness of the approach, we consider here two case studies: one related to PC-3 prostate cancer cells cultured in 2D Petri dishes under control and treated conditions and one related to an OoC experiment of tumor-immune interaction in fibrosarcoma cells. The goodness of the proposed strategy has been verified by successfully comparing the distributions of common descriptors (kinematic descriptors and mean interaction time for the two scenarios respectively) from the trajectories obtained by video analysis and the predicted counterparts.

Machine Learning (ML) based-method applied in recurrent pregnancy loss (RPL) patients diagnostic work-up: a potential innovation in common clinical practice.

RPL is a very debated condition, in which many issues concerning definition, etiological factors to investigate or therapies to apply are still controversial. ML could help clinicians to reach an objectiveness in RPL classification and access to care. Our aim was to stratify RPL patients in different risk classes by applying an ML algorithm, through a diagnostic work-up to validate it for the appropriate prognosis and potential therapeutic approach. 734 patients were enrolled and divided into 4 risk classes, according to the numbers of miscarriages. ML method, called Support Vector Machine (SVM), was used to analyze data. Using the whole set of 43 features and the set of the most informative 18 features we obtained comparable results: respectively 81.86 ± 0.35% and 81.71 ± 0.37% Unbalanced Accuracy. Applying the same method, introducing the only features recommended by ESHRE, a correct classification was obtained only in 58.52 ± 0.58%. ML approach could provide a Support Decision System tool to stratify RPL patients and address them objectively to the proper clinical management.

Sociability and gazing toward humans in dogs and wolves: Simple behaviors with broad implications.

Sociability, defined as the tendency to approach and interact with unfamiliar people, has been found to modulate some communicative responses in domestic dogs, including gaze behavior toward the human face. The objective of this study was to compare sociability and gaze behavior in pet domestic dogs and in human-socialized captive wolves in order to identify the relative influence of domestication and learning in the development of the dog-human bond. In Experiment 1, we assessed the approach behavior and social tendencies of dogs and wolves to a familiar and an unfamiliar person. In Experiment 2, we compared the animal's duration of gaze toward a person's face in the presence of food, which the animals could see but not access. Dogs showed higher levels of interspecific sociability than wolves in all conditions, including those where attention was unavailable. In addition, dogs gazed longer at the person's face than wolves in the presence of out-of-reach food. The potential contributions of domestication, associative learning, and experiences during ontogeny to prosocial behavior toward humans are discussed.

[Delirium and psychotropic substance abuse]. / Il delirio e l'abuso di sostanze.

Delirium represents a common symptom in psychiatric diseases such as psychotic and affective disorders, organic illnesses and psychotropic substance abuse. The literature shows a high risk of developing psychosis in psychotropic substance abusers and a higher proneness to psychotropic substance abuse in people suffering from psychosis. The aim of this review is to discuss the relationship between substance abuse (cocaine, amphetamine, alcohol, cannabinoids, opioids, etc.) and the development of delirium, the way each molecule influences the pathogenesis of delirium, the neurochemical basis of delirium induced by psychotropic substances and the potential endophenotypes involved in a biologically plausible mechanism of mental diseases' pathogenesis.

[The psychopathology of delirium. Evolution of the research from Jaspers to Kapur]. / Psicopatologia del delirio. Evoluzione storica della ricerca da Jaspers a Kapur.

Delirium represents a symptom of a thought disorder in which a belief is strongly held in spite of invalidating evidence. Many varieties of delirium have been described and this symptom is common in psychotic disorders (schizophrenia, brief psychotic disorder, chronic delusional disorders, etc.), affective disorders (major depressive disorder, psychotic depression, melancholic depression, bipolar disorder, etc.), organic illnesses and psychotropic substance abuse. Delirium has been deeply studied by psychopathology. Our work offers an overview of the most relevant psychopathological descriptions of delirium in the last century, from Jaspers to Kapur.

[Teaching how to ask for help: social sensitiveness and early intervention in psychosis]. / Educare alla richiesta di aiuto: sensibilizzazione sociale e intervento precoce negli esordi psicotici.

Psychosis is the most traumatic mental illness influencing both sufferers and their families' quality of life, because of the symptoms and the social stigma. A delay in the recognition of fi rst episode psychosis is unfortunately common being often associated with social functioning decline and worse prognosis. Early detection and intervention could potentially alter the course of this serious illness. We reviewed evidence from the literature on the most recent examples of early intervention in psychosis and on its effectiveness in delaying transition to psychosis, reducing the duration of untreated psychosis, the admission rates, the suicide rates, the treatment costs, as well as preventing relapses and improving the short and long-term symptomatic and functional outcome. The morbidity and mortality associated with schizophrenia spectrum may be improved by a multidisciplinary approach, involving the School, the Primary care and the Information in order to detect as soon as possible the prodromal feelings associated with early psycho-sis. General practitioners are indeed ideally placed to identify mental and emotional changes in the emerging psychotic illnesses and could bridge the gap between specialist mental health services, patients and their families. These are the reasons why mental health should be demanding commitment for both psychiatrists and primary care.

The dimeric assembly of Photobacterium leiognathi and Salmonella typhimurium SodC1 Cu,Zn superoxide dismutases is affected differently by active site demetallation and pH: an analytical ultracentrifuge study.

To establish whether the species-specific variations at the subunit interface of bacterial Cu,Zn superoxide dismutases affect dimer assembly, the association state of the Photobacterium leiognathi (PlSOD) and Salmonella typhimurium (StSOD) enzymes, which differ in 11 out of 19 interface residues, was investigated by analytical ultracentrifugation. The same linkage pattern correlates quaternary assembly, active site metallation, and pH in the two enzymes albeit with quantitative differences. Both holo-enzymes are stable dimers at pH 6.8 and 8.0, although their shape is altered at alkaline pH. In contrast, dimer stability is affected differently by metal removal. Thus, apo-StSOD is a stable dimer at pH 6.8 whereas apo-PlSOD is in reversible monomer-dimer equilibrium. In both apoproteins a pH increase to 8.0 favors monomerization. These effects prove the existence of long-range communication between the active site and the subunit interface and provide a structural explanation for the known functional differences between the two enzymes.

Lipid modification of the Cu,Zn superoxide dismutase from Mycobacterium tuberculosis.

The leader sequence of Mycobacterium tuberculosis Cu,Zn superoxide dismutase (Cu,ZnSOD) contains a prokaryotic membrane lipoprotein attachment site. In the present study, we have found that the protein, which exhibits detectable SOD activity, is lipid-modified and associated with the bacterial membrane when expressed either in M. tuberculosis or in Escherichia coli. These results provide the first demonstration of lipid modification of a Cu,ZnSOD. An analysis of the sodC genes present in available databases indicates that the same signal for lipid modification is also present in the sodC gene products from other mycobacteria and Gram-positive bacteria and, uniquely, in two distinct sodC gene products from the Gram-negative bacterium Salmonella typhimurium. Evidence is also provided for an up-regulation of M. tuberculosis sodC in response to phagocytosis by human macrophages, suggesting that Cu,ZnSOD is involved in the mechanisms that facilitate mycobacterial intracellular growth.

Single mutations at the subunit interface modulate copper reactivity in Photobacterium leiognathi Cu,Zn superoxide dismutase.

The functional properties and X-ray structures of five mutant forms of Photobacterium leiognathi Cu,Zn superoxide dismutase carrying single mutations at residues located at the dimer association interface have been investigated. When compared to the wild-type enzyme, the three-dimensional structures of the mutants show structural perturbations limited to the proximity of the mutation sites and substantial identity of active site geometry. Nonetheless, the catalytic rates of all mutants, measured at neutral pH and low ionic strength by pulse radiolysis, are higher than that of the wild-type protein. Such enzymatic activity increase is paralleled by enhanced active site accessibility to external chelating agents, which, in the mutated enzyme, remove more readily the active site copper ion. It is concluded that mutations at the prokaryotic Cu,Zn superoxide dismutase subunit interface can transduce dynamical perturbation to the active site region, promoting substrate active site accessibility. Such long-range intramolecular communication effects have not been extensively described before within the Cu,Zn superoxide dismutase homology family.

Single mutation induces a metal-dependent subunit association in dimeric Cu,Zn superoxide dismutase.

Tryptophan 83, a residue strongly involved in the intersubunit interaction of the Cu,Zn superoxide dismutases from Photobacterium leiognathi, has been selectively mutated to phenylalanine or tyrosine. The recombinant mutant enzymes expressed in Escherichia coli were purified in two well distinct and stable forms, one dimeric and fully active and the other monomeric and devoid of metals. In agreement, in vitro experiments indicate that the removal and addition of zinc in the mutant enzymes induces monomerization and dimerization, respectively, while does not perturb the dimeric association of the native protein. This is the first unambiguous experimental proof of a direct communication between the intersubunit interface and the metal active site.

Evidence that a guanine nucleotide-binding protein linked to a muscarinic receptor inhibits directly phospholipase C.

The mechanism of phospholipase C regulation by inhibitory receptors was analyzed both in intact and in permeabilized rat thyroid cells (FRTL5). In this system, the muscarinic agonist carbachol inhibited phospholipase C, as indicated by the decrease in the basal levels of inositol 1,4,5-trisphosphate as well as by the reduced adrenergic stimulation of phosphoinositol accumulation, which was paralleled by a fall in the cytosolic Ca2+ levels. This inhibition involved an M2 muscarinic receptor because it was abolished by atropine but not by the M1 antagonist pirenzepine. Cells pretreated with pertussis toxin were not responsive to carbachol, indicating the involvement of a guanine nucleotide-binding protein in this inhibitory process. This possibility was further evaluated in permeabilized cells, where the carbachol inhibition was shown to be completely dependent on GTP. Known second messengers were not involved in this inhibitory process since Ca2+, cAMP, and activators of protein kinases were not able to mimic or prevent the carbachol effect either in intact or in permeabilized FRTL5 cells. In this system, the phospholipases C and A2 are coupled to two classes of muscarinic receptors that display a different sensitivity to pertussis toxin. The carbachol inhibitory effect occurred under conditions that prevented activation of phospholipase A2, excluding a role of the arachidonic acid metabolism in this process. Taken together these data provide the strongest support to date that an inhibitory guanine nucleotide-binding protein sensitive to pertussis toxin can directly mediate receptor-induced inhibition of phospholipase C.

Effect of exogenous hydrogen peroxide on human erythrocytes.

Circulating erythrocytes are drastically susceptible to peroxidative reactions. To examine the extent of the damage induced by exogenous H2O2 we limited the catalase activity in order to study the extent of lysis, the lipid peroxidation and namely the behaviour of membrane micro-viscosity. Our data showed that the erythrocytes can efficiently scavenge exogenous H2O2 without significant damage of the cells and/or their membranes. These findings could confirm the important role of the erythrocytes as extracellular-antioxidant defense.