A dataset for illuminant- and device- invariant colour barcode decoding with cameras.

Barcodes are visual representations of data widely used in commerce and administration to compactly codify information about objects, services, and people. Specifically, a barcode is an image composed of parallel lines, with different widths, spacing and sizes. Generally, the lines are dark (usually black) on a bright background (usually white) or vice-versa. Thanks to this representation, barcodes can be detected and decoded in a way robust to changes of light and noise. However, using barcodes with several colours for the lines is quite intriguing because it enables boosting the barcode's data capacity. Colour barcodes still pose a challenge today, even though numerous studies on the topic were conducted between 1990 and 2022. The main issue that needs to be solved is the creation of an optical technology able to decode colour sequences regardless of the ambient light, the acquisition and printing or visualisation device, and the physical support on which the barcode is printed or displayed. To the best of our knowledge, the studies currently available in literature do not provide the experimental data on which they are based, nor are there online databases that can be used for further studies or for training data analysis procedures based on artificial intelligence techniques. To fill this gap and push further research in this technology, we built COCO-10, a public dataset of colour barcode images, that would like to become a testbench for the development and testing of colour barcode decoding algorithms, taking into account the colour variability due to the light, to the printer and camera gamuts and to the quality of the paper on which the barcode is printed. COCO-10 contains 5400 images of 150 colour barcodes, each of one printed on two white papers with different density and printers and acquired under six illuminations by three smartphones' cameras. For each colour barcode image, a mask identifying the region occupied by the barcode is released too. The 150 colour barcodes have been generated by colouring the lines of black & white barcodes with colours randomly selected from a palette of ten colours including both warm and colour hues. The name COCO-10 just refers to the fact that the dataset contains COlor BarCOdes with 10 possible colours for each line. We also provide a set of 300 images created as follows. The 150 COCO-10 colour barcodes were synthetically superimposed on 150 cluttered backgrounds, resulting in 150 images. The first 75 (group 1) were printed on thick paper, the others (group 2) on plain paper. Each group was further subdivided into subsets of 25 images, resulting in 3 subgroups, each of which was captured by 2 smartphones' cameras under one of the 6 illuminants mentioned above. We also provide masks for these images. These images would like to be a benchmark for testing the accuracy of barcode decoding algorithms, bearing in mind that the performance of these algorithms may be influenced by the accuracy of the previous detection of the barcodes themselves in the background. The total number of images in COCO-10 is 11700, including the 300 synthetic images of the colour barcodes displayed on white and cluttered background, the 5700 real-world images of the colour barcodes printed on white papers and with cluttered backgrounds and their corresponding 5700 masks. We finally highlight that COCO-10 can be also used for developing and testing algorithms for gamut and tone mapping, machine colour constancy, and colour correction.

How the latent geometry of a biological network provides information on its dynamics: the case of the gene network of chronic myeloid leukaemia.

Background: The concept of the latent geometry of a network that can be represented as a graph has emerged from the classrooms of mathematicians and theoretical physicists to become an indispensable tool for determining the structural and dynamic properties of the network in many application areas, including contact networks, social networks, and especially biological networks. It is precisely latent geometry that we discuss in this article to show how the geometry of the metric space of the graph representing the network can influence its dynamics. Methods: We considered the transcriptome network of the Chronic Myeloid Laeukemia K562 cells. We modelled the gene network as a system of springs using a generalization of the Hooke's law to n-dimension (n ≥ 1). We embedded the network, described by the matrix of spring's stiffnesses, in Euclidean, hyperbolic, and spherical metric spaces to determine which one of these metric spaces best approximates the network's latent geometry. We found that the gene network has hyperbolic latent geometry, and, based on this result, we proceeded to cluster the nodes according to their radial coordinate, that in this geometry represents the node popularity. Results: Clustering according to radial coordinate in a hyperbolic metric space when the input to network embedding procedure is the matrix of the stiffnesses of the spring representing the edges, allowed to identify the most popular genes that are also centres of effective spreading and passage of information through the entire network and can therefore be considered the drivers of its dynamics. Conclusion: The correct identification of the latent geometry of the network leads to experimentally confirmed clusters of genes drivers of the dynamics, and, because of this, it is a trustable mean to unveil important information on the dynamics of the network. Not considering the latent metric space of the network, or the assumption of a Euclidean space when this metric structure is not proven to be relevant to the network, especially for complex networks with hierarchical or modularised structure can lead to unreliable network analysis results.

Graph embedding and geometric deep learning relevance to network biology and structural chemistry.

Graphs are used as a model of complex relationships among data in biological science since the advent of systems biology in the early 2000. In particular, graph data analysis and graph data mining play an important role in biology interaction networks, where recent techniques of artificial intelligence, usually employed in other type of networks (e.g., social, citations, and trademark networks) aim to implement various data mining tasks including classification, clustering, recommendation, anomaly detection, and link prediction. The commitment and efforts of artificial intelligence research in network biology are motivated by the fact that machine learning techniques are often prohibitively computational demanding, low parallelizable, and ultimately inapplicable, since biological network of realistic size is a large system, which is characterised by a high density of interactions and often with a non-linear dynamics and a non-Euclidean latent geometry. Currently, graph embedding emerges as the new learning paradigm that shifts the tasks of building complex models for classification, clustering, and link prediction to learning an informative representation of the graph data in a vector space so that many graph mining and learning tasks can be more easily performed by employing efficient non-iterative traditional models (e.g., a linear support vector machine for the classification task). The great potential of graph embedding is the main reason of the flourishing of studies in this area and, in particular, the artificial intelligence learning techniques. In this mini review, we give a comprehensive summary of the main graph embedding algorithms in light of the recent burgeoning interest in geometric deep learning.

Gene Expression Landscape of Chronic Myeloid Leukemia K562 Cells Overexpressing the Tumor Suppressor Gene PTPRG.

This study concerns the analysis of the modulation of Chronic Myeloid Leukemia (CML) cell model K562 transcriptome following transfection with the tumor suppressor gene encoding for Protein Tyrosine Phosphatase Receptor Type G (PTPRG) and treatment with the tyrosine kinase inhibitor (TKI) Imatinib. Specifically, we aimed at identifying genes whose level of expression is altered by PTPRG modulation and Imatinib concentration. Statistical tests as differential expression analysis (DEA) supported by gene set enrichment analysis (GSEA) and modern methods of ontological term analysis are presented along with some results of current interest for forthcoming experimental research in the field of the transcriptomic landscape of CML. In particular, we present two methods that differ in the order of the analysis steps. After a gene selection based on fold-change value thresholding, we applied statistical tests to select differentially expressed genes. Therefore, we applied two different methods on the set of differentially expressed genes. With the first method (Method 1), we implemented GSEA, followed by the identification of transcription factors. With the second method (Method 2), we first selected the transcription factors from the set of differentially expressed genes and implemented GSEA on this set. Method 1 is a standard method commonly used in this type of analysis, while Method 2 is unconventional and is motivated by the intention to identify transcription factors more specifically involved in biological processes relevant to the CML condition. Both methods have been equipped in ontological knowledge mining and word cloud analysis, as elements of novelty in our analytical procedure. Data analysis identified RARG and CD36 as a potential PTPRG up-regulated genes, suggesting a possible induction of cell differentiation toward an erithromyeloid phenotype. The prediction was confirmed at the mRNA and protein level, further validating the approach and identifying a new molecular mechanism of tumor suppression governed by PTPRG in a CML context.

Dynamic Modelling of DNA Repair Pathway at the Molecular Level: A New Perspective.

DNA is the genetic repository for all living organisms, and it is subject to constant changes caused by chemical and physical factors. Any change, if not repaired, erodes the genetic information and causes mutations and diseases. To ensure overall survival, robust DNA repair mechanisms and damage-bypass mechanisms have evolved to ensure that the DNA is constantly protected against potentially deleterious damage while maintaining its integrity. Not surprisingly, defects in DNA repair genes affect metabolic processes, and this can be seen in some types of cancer, where DNA repair pathways are disrupted and deregulated, resulting in genome instability. Mathematically modelling the complex network of genes and processes that make up the DNA repair network will not only provide insight into how cells recognise and react to mutations, but it may also reveal whether or not genes involved in the repair process can be controlled. Due to the complexity of this network and the need for a mathematical model and software platform to simulate different investigation scenarios, there must be an automatic way to convert this network into a mathematical model. In this paper, we present a topological analysis of one of the networks in DNA repair, specifically homologous recombination repair (HR). We propose a method for the automatic construction of a system of rate equations to describe network dynamics and present results of a numerical simulation of the model and model sensitivity analysis to the parameters. In the past, dynamic modelling and sensitivity analysis have been used to study the evolution of tumours in response to drugs in cancer medicine. However, automatic generation of a mathematical model and the study of its sensitivity to parameter have not been applied to research on the DNA repair network so far. Therefore, we present this application as an approach for medical research against cancer, since it could give insight into a possible approach with which central nodes of the networks and repair genes could be identified and controlled with the ultimate goal of aiding cancer therapy to fight the onset of cancer and its progression.

Identification of Potential Leukocyte Biomarkers Related to Drug Recovery of CFTR: Clinical Applications in Cystic Fibrosis.

The aim of this study was the identification of specific proteomic profiles, related to a restored cystic fibrosis transmembrane conductance regulator (CFTR) activity in cystic fibrosis (CF) leukocytes before and after ex vivo treatment with the potentiator VX770. We used leukocytes, isolated from CF patients carrying residual function mutations and eligible for Ivacaftor therapy, and performed CFTR activity together with proteomic analyses through micro-LC-MS. Bioinformatic analyses of the results obtained revealed the downregulation of proteins belonging to the leukocyte transendothelial migration and regulation of actin cytoskeleton pathways when CFTR activity was rescued by VX770 treatment. In particular, we focused our attention on matrix metalloproteinase 9 (MMP9), because the high expression of this protease potentially contributes to parenchyma lung destruction and dysfunction in CF. Thus, the downregulation of MMP9 could represent one of the possible positive effects of VX770 in decreasing the disease progression, and a potential biomarker for the prediction of the efficacy of therapies targeting the defect of Cl- transport in CF.

Machine Learning for Causal Inference in Biological Networks: Perspectives of This Challenge.

Most machine learning-based methods predict outcomes rather than understanding causality. Machine learning methods have been proved to be efficient in finding correlations in data, but unskilful to determine causation. This issue severely limits the applicability of machine learning methods to infer the causal relationships between the entities of a biological network, and more in general of any dynamical system, such as medical intervention strategies and clinical outcomes system, that is representable as a network. From the perspective of those who want to use the results of network inference not only to understand the mechanisms underlying the dynamics, but also to understand how the network reacts to external stimuli (e. g. environmental factors, therapeutic treatments), tools that can understand the causal relationships between data are highly demanded. Given the increasing popularity of machine learning techniques in computational biology and the recent literature proposing the use of machine learning techniques for the inference of biological networks, we would like to present the challenges that mathematics and computer science research faces in generalising machine learning to an approach capable of understanding causal relationships, and the prospects that achieving this will open up for the medical application domains of systems biology, the main paradigm of which is precisely network biology at any physical scale.

Composite nanocellulose-based hydrogels with spatially oriented degradation and retarded release of macromolecules.

The oral delivery of macromolecular therapeutics to the intestinal tract requires novel, robust, and controlled formulations. Here, we report on fabrication by molding of composite hydrogel cylinders made of cellulose nanocrystals (CNCs) and chitosan (Cht) and their performance as delivery vehicles. CNCs provide excellent mechanical and chemical stress resistance, whereas Cht allows scaffold degradation by enzyme digestion. The release of a representative medium size protein (bovine serum albumin) dispersed in the hydrogel is slow and shows a sigmoidal profile; meanwhile, the hydrogel scaffold degrades according to a preferred route, that is the cylinder is eroded along the vertical axis. The cup-like, scarcely interconnected porous network, with a gradient of hardness along the cylinder axis, and the compact skin-like layer covering the lateral wall which stayed in contact with the mold during gelification, explain the preferred erosion direction and the long-term protein release. The possible effect of the molding process on hydrogel structure suggests that molding could be a simple and cheap way to favor surface compaction and directional scaffold degradation.

Control Theory and Cancer Chemotherapy: How They Interact.

Control theory arises in most modern real-life applications, not least in biological and medical applications. In particular, in biological and medical contexts, the role of control theory began to take shape in the early 1980s when the first works appeared on the application of control theory in models of pharmacokinetics and pharmacodynamics for antitumor therapies. Forty years after those first works, the theory of control continues to be considered a mathematical analysis tool of extreme importance and usefulness, but the challenges it must overcome in order to manage the complexity of biological processes are in fact not yet overcome. In this article, we introduce the reader to the basic ideas of control theory, its aims and its mathematical formalization, and we review its use in cell phase-specific models for cancer chemotherapy. We discuss strengths and limitations of the control theory approach to the analysis pharmacokinetics and pharmacodynamics models, and we will see that most of them are strongly related to data availability and mathematical form of the model. We propose some future research directions that could prove useful in overcoming the these limitations and we indicate the crucial steps preliminary to a useful and informative application of control theory to cancer chemotherapy modeling.

On TD-WGcluster: Theoretical Foundations and Guidelines for the User.

We review the TD-WGcluster (time delayed weighted edge clustering) software integrating static interaction networks with time series data in order to detect modules of nodes between which the information flows at similar time delays and intensities. The software has represented an advancement of the state of the art in the software for the identification of connected components due to its peculiarity of dealing with direct and weighted graphs, where the attributes of the physical entities represented by nodes vary over time. This chapter aims to deepen those theoretical aspects of the clustering model implemented by TD-WGcluster that may be of greater interest to the user. We show the instructions necessary to run the software through some exploratory cases and comment on the results obtained.

A Reaction-Based Model of the State Space of Chemical Reaction Systems Enables Efficient Simulations.

The choice of the state space representation of a system can turn into a prominent advantage or burden in any endeavour to mathematically model dynamical systems since it entails the analytical tractability of the related modelling formalism and the efficiency of the numerical computation. The Reaction-Based Model (RBM) of the state space, which is presented in this article, is a novel formalization of the kinetics of a system of interacting molecules. According to our representation, the state Sµ of a system of M reactions and N molecular species, is identified with the occurrence of the reaction Rµ ( µ = 1, ..., M). The transition between any two states Sµ and Sν is modelled as a first-order reaction Sµ → Sν and described by mass action-like equation for the partial time derivative of the variables P(Sµ, t) and P(Sν, t), which denote the probabilities that the system lies in the two states, respectively. The rate equations for the state probabilities are coupled with those for the abundance of molecular species. Altogether, the rate equations along with the specification of the initial conditions define the Cauchy problem whose solution describes the time-evolution of the system. The RBM has been applied to a typical severely stiff biological case study. The numerical solutions of the system's dynamics turned out to be computationally more efficient and in agreement with the results of the stochastic and hybrid stochastic/deterministic simulation algorithms.

Identifying necessary and sufficient conditions for the observability of models of biochemical processes.

The notions of observability and controllability of non-linear systems are a cornerstone of mathematical control theory and cover a wide scope of applications including process design, characterization, monitoring and control. Synthetic biology - which aims to (re)-program living functionalities - and bio-based process engineering - which aims to develop biotechnological manufacturing processes based on industrial and natural living agents - remarkably benefit of methodological improvements inspired to control theory for countless reasons including the huge variety of control mechanisms in living organisms, experimental limitations in terms of measurement feasibility, design of controllers - at single cell or population level - of synthetic production processes and process optimization purposes. Many fundamental problems of control theory such as stabilisability of unstable systems and optimal control may be solved under the assumption that the system is observable/controllable. Observability and controllability are mathematical duals, that means that the observability property can be determined analysing the controllability of the dual system and vice versa. Given this duality, we focus on observability. In this work, we revisit a generalization of the Fujisawa and Kuh theorem as a tool to explore the possibility that a system is observable. We show that the theorem, when applicable, is a sufficient but not necessary condition for observability. We revisit the theorem to propose a necessary and sufficient condition for observability for non-linear systems. Finally, we show how it is possible to identify regions of the phase space of the model in which the model is observable.

How Monte Carlo heuristics aid to identify the physical processes of drug release kinetics.

We implement a Monte Carlo heuristic algorithm to model drug release from a solid dosage form. We show that with Monte Carlo simulations it is possible to identify and explain the causes of the unsatisfactory predictive power of current drug release models. It is well known that the power-law, the exponential models, as well as those derived from or inspired by them accurately reproduce only the first 60% of the release curve of a drug from a dosage form. In this study, by using Monte Carlo simulation approaches, we show that these models fit quite accurately almost the entire release profile when the release kinetics is not governed by the coexistence of different physico-chemical mechanisms. We show that the accuracy of the traditional models are comparable with those of Monte Carlo heuristics when these heuristics approximate and oversimply the phenomenology of drug release. This observation suggests to develop and use novel Monte Carlo simulation heuristics able to describe the complexity of the release kinetics, and consequently to generate data more similar to those observed in real experiments. Implementing Monte Carlo simulation heuristics of the drug release phenomenology may be much straightforward and efficient than hypothesizing and implementing from scratch complex mathematical models of the physical processes involved in drug release. Identifying and understanding through simulation heuristics what processes of this phenomenology reproduce the observed data and then formalize them in mathematics may allow avoiding time-consuming, trial-error based regression procedures. Three bullet points, highlighting the customization of the procedure. •An efficient heuristics based on Monte Carlo methods for simulating drug release from solid dosage form encodes is presented. It specifies the model of the physical process in a simple but accurate way in the formula of the Monte Carlo Micro Step (MCS) time interval.•Given the experimentally observed curve of drug release, we point out how Monte Carlo heuristics can be integrated in an evolutionary algorithmic approach to infer the mode of MCS best fitting the observed data, and thus the observed release kinetics.•The software implementing the method is written in R language, the free most used language in the bioinformaticians community.

Hybrid deterministic/stochastic simulation of complex biochemical systems.

In a biological cell, cellular functions and the genetic regulatory apparatus are implemented and controlled by complex networks of chemical reactions involving genes, proteins, and enzymes. Accurate computational models are indispensable means for understanding the mechanisms behind the evolution of a complex system, not always explored with wet lab experiments. To serve their purpose, computational models, however, should be able to describe and simulate the complexity of a biological system in many of its aspects. Moreover, it should be implemented by efficient algorithms requiring the shortest possible execution time, to avoid enlarging excessively the time elapsing between data analysis and any subsequent experiment. Besides the features of their topological structure, the complexity of biological networks also refers to their dynamics, that is often non-linear and stiff. The stiffness is due to the presence of molecular species whose abundance fluctuates by many orders of magnitude. A fully stochastic simulation of a stiff system is computationally time-expensive. On the other hand, continuous models are less costly, but they fail to capture the stochastic behaviour of small populations of molecular species. We introduce a new efficient hybrid stochastic-deterministic computational model and the software tool MoBioS (MOlecular Biology Simulator) implementing it. The mathematical model of MoBioS uses continuous differential equations to describe the deterministic reactions and a Gillespie-like algorithm to describe the stochastic ones. Unlike the majority of current hybrid methods, the MoBioS algorithm divides the reactions' set into fast reactions, moderate reactions, and slow reactions and implements a hysteresis switching between the stochastic model and the deterministic model. Fast reactions are approximated as continuous-deterministic processes and modelled by deterministic rate equations. Moderate reactions are those whose reaction waiting time is greater than the fast reaction waiting time but smaller than the slow reaction waiting time. A moderate reaction is approximated as a stochastic (deterministic) process if it was classified as a stochastic (deterministic) process at the time at which it crosses the threshold of low (high) waiting time. A Gillespie First Reaction Method is implemented to select and execute the slow reactions. The performances of MoBios were tested on a typical example of hybrid dynamics: that is the DNA transcription regulation. The simulated dynamic profile of the reagents' abundance and the estimate of the error introduced by the fully deterministic approach were used to evaluate the consistency of the computational model and that of the software tool.

Network-Oriented Approaches to Anticancer Drug Response.

A long-standing paradigm in drug discovery has been the concept of designing maximally selective drugs to act on individual targets considered to underlie a disease of interest. Nonetheless, although some drugs have proven to be successful, many more potential drugs identified by the "one gene, one drug, one disease" approach have been found to be less effective than expected or to cause notable side effects. Advances in systems biology and high-throughput in-depth genomic profiling technologies along with an analysis of the successful and failed drugs uncovered that the prominent factor to determine drug sensitivity is the intrinsic robustness of the response of biological systems in the face of perturbations. The complexity of the molecular and cellular bases of systems responses to drug interventions has fostered an increased interest in systems-oriented approaches to drug discovery. Consonant with this knowledge of the multifactorial mechanistic basis of drug sensitivity and resistance is the application of network-based approaches for the identification of molecular (multi-)feature signatures associated with desired (multi-)drug phenotypic profiles. This chapter illustrates the principal network analysis and inference techniques which have found application in systems-oriented drug design and considers their benefits and drawbacks in relation to the nature of the data produced by network pharmacology.

Accurate prediction of the age incidence of chronic myeloid leukemia with an improved two-mutation mathematical model.

Chronic myeloid leukemia (CML) is a malignant clonal disorder whose hallmark is a reciprocal translocation between chromosomes 9 and 22 occurring in 95% of affected patients. This translocation causes the expression of a deregulated BCR/ABL fusion oncogene and, interestingly, is detectable in healthy individuals. Based on this information we assumed that the sole presence of the BCR/ABL transcript represents a necessary but not sufficient event for the clonal expansion of CML precursors. We developed a mathematical model introducing a probability that any normal cell undergoes a first aberration, and a probability that a cell that experienced a first mutation undergoes a second mutation as well. Two variants are proposed and analyzed: in the first the probability of the first mutation is supposed to be age independent and in the second, it depends on the hemopoietic cell turnover and mass. The model parameters have been estimated by regression from the observed CML incidence curves. Our models offer a significantly improved version of existing one-step and two-steps models, as they integrate key clinical and biological data reported in the literature and accurately fit the observed incidence. Our models also estimate the increased radiation-associated mutation rate at a younger age in atomic bomb survivors. Although this work focuses on CML, the modelling approach can be applied to all types of leukemia and lymphoma.

Time Series Analysis of the Bacillus subtilis Sporulation Network Reveals Low Dimensional Chaotic Dynamics.

Chaotic behavior refers to a behavior which, albeit irregular, is generated by an underlying deterministic process. Therefore, a chaotic behavior is potentially controllable. This possibility becomes practically amenable especially when chaos is shown to be low-dimensional, i.e., to be attributable to a small fraction of the total systems components. In this case, indeed, including the major drivers of chaos in a system into the modeling approach allows us to improve predictability of the systems dynamics. Here, we analyzed the numerical simulations of an accurate ordinary differential equation model of the gene network regulating sporulation initiation in Bacillus subtilis to explore whether the non-linearity underlying time series data is due to low-dimensional chaos. Low-dimensional chaos is expectedly common in systems with few degrees of freedom, but rare in systems with many degrees of freedom such as the B. subtilis sporulation network. The estimation of a number of indices, which reflect the chaotic nature of a system, indicates that the dynamics of this network is affected by deterministic chaos. The neat separation between the indices obtained from the time series simulated from the model and those obtained from time series generated by Gaussian white and colored noise confirmed that the B. subtilis sporulation network dynamics is affected by low dimensional chaos rather than by noise. Furthermore, our analysis identifies the principal driver of the networks chaotic dynamics to be sporulation initiation phosphotransferase B (Spo0B). We then analyzed the parameters and the phase space of the system to characterize the instability points of the network dynamics, and, in turn, to identify the ranges of values of Spo0B and of the other drivers of the chaotic dynamics, for which the whole system is highly sensitive to minimal perturbation. In summary, we described an unappreciated source of complexity in the B. subtilis sporulation network by gathering evidence for the chaotic behavior of the system, and by suggesting candidate molecules driving chaos in the system. The results of our chaos analysis can increase our understanding of the intricacies of the regulatory network under analysis, and suggest experimental work to refine our behavior of the mechanisms underlying B. subtilis sporulation initiation control.