Facilitating clinically relevant skin tumor diagnostics with spectroscopy-driven machine learning.

In the dawning era of artificial intelligence (AI), health care stands to undergo a significant transformation with the increasing digitalization of patient data. Digital imaging, in particular, will serve as an important platform for AI to aid decision making and diagnostics. A growing number of studies demonstrate the potential of automatic pre-surgical skin tumor delineation, which could have tremendous impact on clinical practice. However, current methods rely on having ground truth images in which tumor borders are already identified, which is not clinically possible. We report a novel approach where hyperspectral images provide spectra from small regions representing healthy tissue and tumor, which are used to generate prediction maps using artificial neural networks (ANNs), after which a segmentation algorithm automatically identifies the tumor borders. This circumvents the need for ground truth images, since an ANN model is trained with data from each individual patient, representing a more clinically relevant approach.

Elucidating fungal decomposition of organic matter at sub-micrometer spatial scales using optical photothermal infrared (O-PTIR) microspectroscopy.

In microbiological studies, a common goal is to link environmental factors to microbial activities. Both environmental factors and microbial activities are typically derived from bulk samples. It is becoming increasingly clear that such bulk environmental parameters poorly represent the microscale environments microorganisms experience. Using infrared (IR) microspectroscopy, the spatial distribution of chemical compound classes can be visualized, making it a useful tool for studying the interactions between microbial cells and their microenvironments. The spatial resolution of conventional IR microspectroscopy has been limited by the diffraction limit of IR light. The recent development of optical photothermal infrared (O-PTIR) microspectroscopy has pushed the spatial resolution of IR microspectroscopy beyond this diffraction limit, allowing the distribution of chemical compound classes to be visualized at sub-micrometer spatial scales. To examine the potential and limitations of O-PTIR microspectroscopy to probe the interactions between fungal cells and their immediate environments, we imaged the decomposition of cellulose films by cells of the ectomycorrhizal fungus Paxillus involutus and compared O-PTIR results using conventional IR microspectroscopy. Whereas the data collected with conventional IR microspectroscopy indicated that P. involutus has only a very limited ability to decompose cellulose films, O-PTIR data suggested that the ability of P. involutus to decompose cellulose was substantial. Moreover, the O-PTIR method enabled the identification of a zone located outside the fungal hyphae where the cellulose was decomposed by oxidation. We conclude that O-PTIR can provide valuable new insights into the abilities and mechanisms by which microorganisms interact with their surrounding environments.IMPORTANCEInfrared (IR) microspectroscopy allows the spatial distribution of chemical compound classes to be visualized. The use of conventional IR microspectroscopy in microbiological studies has been restricted by limited spatial resolution. Recent developments in laser technology have enabled a new class of IR microspectroscopy instruments to be developed, pushing the spatial resolution beyond the diffraction limit of IR light to approximately 500 nm. This improved spatial resolution now allows microscopic observations of changes in chemical compounds to be made, making IR microspectroscopy a useful tool to investigate microscale changes in chemistry that are caused by microbial activity. We show these new possibilities using optical photothermal infrared microspectroscopy to visualize the changes in cellulose substrates caused by oxidation by the ectomycorrhizal fungus Paxillus involutus at the interface between individual fungal hyphae and cellulose substrates.

OCTAVVS: A Graphical Toolbox for High-Throughput Preprocessing and Analysis of Vibrational Spectroscopy Imaging Data.

Modern vibrational spectroscopy techniques enable the rapid collection of thousands of spectra in a single hyperspectral image, allowing researchers to study spatially heterogeneous samples at micrometer resolution. A number of algorithms have been developed to correct for effects such as atmospheric absorption, light scattering by cellular structures and varying baseline levels. After preprocessing, spectra are commonly decomposed and clustered to reveal informative patterns and subtle spectral changes. Several of these steps are slow, labor-intensive and require programming skills to make use of published algorithms and code. We here present a free and platform-independent graphical toolbox that allows rapid preprocessing of large sets of spectroscopic images, including atmospheric correction and a new algorithm for resonant Mie scattering with improved speed. The software also includes modules for decomposition into constituent spectra using the popular Multivariate Curve Resolution-Alternating Least Squares (MCR-ALS) algorithm, augmented by region-of-interest selection, as well as clustering and cluster annotation.

Regulation of fungal decomposition at single-cell level.

Filamentous fungi play a key role as decomposers in Earth's nutrient cycles. In soils, substrates are heterogeneously distributed in microenvironments. Hence, individual hyphae of a mycelium may experience very different environmental conditions simultaneously. In the current work, we investigated how fungi cope with local environmental variations at single-cell level. We developed a method based on infrared spectroscopy that allows the direct, in-situ chemical imaging of the decomposition activity of individual hyphal tips. Colonies of the ectomycorrhizal Basidiomycete Paxillus involutus were grown on liquid media, while parts of colonies were allowed to colonize lignin patches. Oxidative decomposition of lignin by individual hyphae growing under different conditions was followed for a period of seven days. We identified two sub-populations of hyphal tips: one with low decomposition activity and one with much higher activity. Active cells secreted more extracellular polymeric substances and oxidized lignin more strongly. The ratio of active to inactive hyphae strongly depended on the environmental conditions in lignin patches, but was further mediated by the decomposition activity of entire mycelia. Phenotypic heterogeneity occurring between genetically identical hyphal tips may be an important strategy for filamentous fungi to cope with heterogeneous and constantly changing soil environments.

Homology and linkage in crossover for linear genomes of variable length.

The use of variable-length genomes in evolutionary computation has applications in optimisation when the size of the search space is unknown, and provides a unique environment to study the evolutionary dynamics of genome structure. Here, we revisit crossover for linear genomes of variable length, identifying two crucial attributes of successful recombination algorithms: the ability to retain homologous structure, and to reshuffle variant information. We introduce direct measures of these properties-homology score and linkage score-and use them to review existing crossover algorithms, as well as two novel ones. In addition, we measure the performance of these crossover methods on three different benchmark problems, and find that variable-length genomes out-perform fixed-length variants in all three cases. Our homology and linkage scores successfully explain the difference in performance between different crossover methods, providing a simple and insightful framework for crossover in a variable-length setting.

The soil organic matter decomposition mechanisms in ectomycorrhizal fungi are tuned for liberating soil organic nitrogen.

Many trees form ectomycorrhizal symbiosis with fungi. During symbiosis, the tree roots supply sugar to the fungi in exchange for nitrogen, and this process is critical for the nitrogen and carbon cycles in forest ecosystems. However, the extents to which ectomycorrhizal fungi can liberate nitrogen and modify the soil organic matter and the mechanisms by which they do so remain unclear since they have lost many enzymes for litter decomposition that were present in their free-living, saprotrophic ancestors. Using time-series spectroscopy and transcriptomics, we examined the ability of two ectomycorrhizal fungi from two independently evolved ectomycorrhizal lineages to mobilize soil organic nitrogen. Both species oxidized the organic matter and accessed the organic nitrogen. The expression of those events was controlled by the availability of glucose and inorganic nitrogen. Despite those similarities, the decomposition mechanisms, including the type of genes involved as well as the patterns of their expression, differed markedly between the two species. Our results suggest that in agreement with their diverse evolutionary origins, ectomycorrhizal fungi use different decomposition mechanisms to access organic nitrogen entrapped in soil organic matter. The timing and magnitude of the expression of the decomposition activity can be controlled by the below-ground nitrogen quality and the above-ground carbon supply.

Fenton reaction facilitates organic nitrogen acquisition by an ectomycorrhizal fungus.

Boreal trees rely on their ectomycorrhizal fungal symbionts to acquire growth-limiting nutrients, such as nitrogen (N), which mainly occurs as proteins complexed in soil organic matter (SOM). The mechanisms for liberating this N are unclear as ectomycorrhizal fungi have lost many genes encoding lignocellulose-degrading enzymes present in their saprotrophic ancestors. We hypothesized that hydroxyl radicals (˙ OH), produced by the ectomycorrhizal fungus Paxillus involutus during growth on SOM, are involved in liberating organic N. Paxillus involutus was grown for 7 d on N-containing or N-free substrates that represent major organic compounds of SOM. ˙ OH production, ammonium assimilation, and proteolytic activity were measured daily. ˙ OH production was strongly induced when P. involutus switched from ammonium to protein as the main N source. Extracellular proteolytic activity was initiated shortly after the oxidation. Oxidized protein substrates induced higher proteolytic activity than unmodified proteins. Dynamic modeling predicted that ˙ OH production occurs in a burst, regulated mainly by ammonium and ferric iron concentrations. We propose that the production of ˙ OH and extracellular proteolytic enzymes are regulated by similar nutritional signals. Oxidation works in concert with proteolysis, improving N liberation from proteins in SOM. Organic N mining by ectomycorrhizal fungi has, until now, only been attributed to proteolysis.

Selection Shapes Transcriptional Logic and Regulatory Specialization in Genetic Networks.

BACKGROUND: Living organisms need to regulate their gene expression in response to environmental signals and internal cues. This is a computational task where genes act as logic gates that connect to form transcriptional networks, which are shaped at all scales by evolution. Large-scale mutations such as gene duplications and deletions add and remove network components, whereas smaller mutations alter the connections between them. Selection determines what mutations are accepted, but its importance for shaping the resulting networks has been debated. METHODOLOGY: To investigate the effects of selection in the shaping of transcriptional networks, we derive transcriptional logic from a combinatorially powerful yet tractable model of the binding between DNA and transcription factors. By evolving the resulting networks based on their ability to function as either a simple decision system or a circadian clock, we obtain information on the regulation and logic rules encoded in functional transcriptional networks. Comparisons are made between networks evolved for different functions, as well as with structurally equivalent but non-functional (neutrally evolved) networks, and predictions are validated against the transcriptional network of E. coli. PRINCIPAL FINDINGS: We find that the logic rules governing gene expression depend on the function performed by the network. Unlike the decision systems, the circadian clocks show strong cooperative binding and negative regulation, which achieves tight temporal control of gene expression. Furthermore, we find that transcription factors act preferentially as either activators or repressors, both when binding multiple sites for a single target gene and globally in the transcriptional networks. This separation into positive and negative regulators requires gene duplications, which highlights the interplay between mutation and selection in shaping the transcriptional networks.

Rethinking transcriptional activation in the Arabidopsis circadian clock.

Circadian clocks are biological timekeepers that allow living cells to time their activity in anticipation of predictable daily changes in light and other environmental factors. The complexity of the circadian clock in higher plants makes it difficult to understand the role of individual genes or molecular interactions, and mathematical modelling has been useful in guiding clock research in model organisms such as Arabidopsis thaliana. We present a model of the circadian clock in Arabidopsis, based on a large corpus of published time course data. It appears from experimental evidence in the literature that most interactions in the clock are repressive. Hence, we remove all transcriptional activation found in previous models of this system, and instead extend the system by including two new components, the morning-expressed activator RVE8 and the nightly repressor/activator NOX. Our modelling results demonstrate that the clock does not need a large number of activators in order to reproduce the observed gene expression patterns. For example, the sequential expression of the PRR genes does not require the genes to be connected as a series of activators. In the presented model, transcriptional activation is exclusively the task of RVE8. Predictions of how strongly RVE8 affects its targets are found to agree with earlier interpretations of the experimental data, but generally we find that the many negative feedbacks in the system should discourage intuitive interpretations of mutant phenotypes. The dynamics of the clock are difficult to predict without mathematical modelling, and the clock is better viewed as a tangled web than as a series of loops.

Light and circadian regulation of clock components aids flexible responses to environmental signals.

The circadian clock measures time across a 24 h period, increasing fitness by phasing biological processes to the most appropriate time of day. The interlocking feedback loop mechanism of the clock is conserved across species; however, the number of loops varies. Mathematical and computational analyses have suggested that loop complexity affects the overall flexibility of the oscillator, including its responses to entrainment signals. We used a discriminating experimental assay, at the transition between different photoperiods, in order to test this proposal in a minimal circadian network (in Ostreococcus tauri) and a more complex network (in Arabidopsis thaliana). Transcriptional and translational reporters in O. tauri primarily tracked dawn or dusk, whereas in A. thaliana, a wider range of responses were observed, consistent with its more flexible clock. Model analysis supported the requirement for this diversity of responses among the components of the more complex network. However, these and earlier data showed that the O. tauri network retains surprising flexibility, despite its simple circuit. We found that models constructed from experimental data can show flexibility either from multiple loops and/or from multiple light inputs. Our results suggest that O. tauri has adopted the latter strategy, possibly as a consequence of genomic reduction.

SBSI: an extensible distributed software infrastructure for parameter estimation in systems biology.

SUMMARY: Complex computational experiments in Systems Biology, such as fitting model parameters to experimental data, can be challenging to perform. Not only do they frequently require a high level of computational power, but the software needed to run the experiment needs to be usable by scientists with varying levels of computational expertise, and modellers need to be able to obtain up-to-date experimental data resources easily. We have developed a software suite, the Systems Biology Software Infrastructure (SBSI), to facilitate the parameter-fitting process. SBSI is a modular software suite composed of three major components: SBSINumerics, a high-performance library containing parallelized algorithms for performing parameter fitting; SBSIDispatcher, a middleware application to track experiments and submit jobs to back-end servers; and SBSIVisual, an extensible client application used to configure optimization experiments and view results. Furthermore, we have created a plugin infrastructure to enable project-specific modules to be easily installed. Plugin developers can take advantage of the existing user-interface and application framework to customize SBSI for their own uses, facilitated by SBSI's use of standard data formats. AVAILABILITY AND IMPLEMENTATION: All SBSI binaries and source-code are freely available from http://sourceforge.net/projects/sbsi under an Apache 2 open-source license. The server-side SBSINumerics runs on any Unix-based operating system; both SBSIVisual and SBSIDispatcher are written in Java and are platform independent, allowing use on Windows, Linux and Mac OS X. The SBSI project website at http://www.sbsi.ed.ac.uk provides documentation and tutorials.

On evaluating models in Computational Morphodynamics.

Recent advances in experimental plant biology have led to an increased potential to investigate plant development at a systems level. The emerging research field of Computational Morphodynamics has the aim to lead this development by combining dynamic spatial experimental data with computational models of molecular networks, growth, and mechanics in a multicellular context. The increased number of published models may lead to a diversification of our understanding of the systems, and methods for evaluating, comparing, and sharing models are main challenges for the future. We will discuss this problem using ideas originating from physics and use recent computational models of plant development as examples.

Mutation-induced fold switching among lattice proteins.

Recent experiments uncovered a mutational pathway between two proteins, along which a single mutation causes a switch in fold. Searching for such paths between real proteins remains, despite this achievement, a true challenge. Here, we analyze fold switching in the minimalistic hydrophobic/polar model on a square lattice. For this analysis, we generate a comprehensive sequence-structure database for chains of length ≤ 30, which exceeds previous work by five units. Single-mutation-induced fold switching turns out to be quite common in the model. The switches define a fold network, whose topology is roughly similar to what one would expect for a set of randomly connected nodes. In the combinatorially challenging search for fold switches between two proteins, a tempting strategy is to only consider paths containing the minimum number of mutations. Such a restricted search fails to correctly identify 40% of the single-mutation-linked fold pairs that we observe. The thermodynamic stability is correlated with mutational stability and is, on average, markedly reduced at the observed fold switches.

Proteasome function is required for biological timing throughout the twenty-four hour cycle.

Circadian clocks were, until recently, seen as a consequence of rhythmic transcription of clock components, directed by transcriptional/translational feedback loops (TTFLs). Oscillations of protein modification were then discovered in cyanobacteria. Canonical posttranslational signaling processes have known importance for clocks across taxa. More recently, evidence from the unicellular eukaryote Ostreococcus tauri revealed a transcription-independent, rhythmic protein modification shared in anucleate human cells. In this study, the Ostreococcus system reveals a central role for targeted protein degradation in the mechanism of circadian timing. The Ostreococcus clockwork contains a TTFL involving the morning-expressed CCA1 and evening-expressed TOC1 proteins. Cellular CCA1 and TOC1 protein content and degradation rates are analyzed qualitatively and quantitatively using luciferase reporter fusion proteins. CCA1 protein degradation rates, measured in high time resolution, feature a sharp clock-regulated peak under constant conditions. TOC1 degradation peaks in response to darkness. Targeted protein degradation, unlike transcription and translation, is shown to be essential to sustain TTFL rhythmicity throughout the circadian cycle. Although proteasomal degradation is not necessary for sustained posttranslational oscillations in transcriptionally inactive cells, TTFL and posttranslational oscillators are normally coupled, and proteasome function is crucial to sustain both.

Circadian rhythms persist without transcription in a eukaryote.

Circadian rhythms are ubiquitous in eukaryotes, and coordinate numerous aspects of behaviour, physiology and metabolism, from sleep/wake cycles in mammals to growth and photosynthesis in plants. This daily timekeeping is thought to be driven by transcriptional-translational feedback loops, whereby rhythmic expression of 'clock' gene products regulates the expression of associated genes in approximately 24-hour cycles. The specific transcriptional components differ between phylogenetic kingdoms. The unicellular pico-eukaryotic alga Ostreococcus tauri possesses a naturally minimized clock, which includes many features that are shared with plants, such as a central negative feedback loop that involves the morning-expressed CCA1 and evening-expressed TOC1 genes. Given that recent observations in animals and plants have revealed prominent post-translational contributions to timekeeping, a reappraisal of the transcriptional contribution to oscillator function is overdue. Here we show that non-transcriptional mechanisms are sufficient to sustain circadian timekeeping in the eukaryotic lineage, although they normally function in conjunction with transcriptional components. We identify oxidation of peroxiredoxin proteins as a transcription-independent rhythmic biomarker, which is also rhythmic in mammals. Moreover we show that pharmacological modulators of the mammalian clock mechanism have the same effects on rhythms in Ostreococcus. Post-translational mechanisms, and at least one rhythmic marker, seem to be better conserved than transcriptional clock regulators. It is plausible that the oldest oscillator components are non-transcriptional in nature, as in cyanobacteria, and are conserved across kingdoms.

Multiple light inputs to a simple clock circuit allow complex biological rhythms.

Circadian clocks are biological timekeepers that allow living cells to time their activity in anticipation of predictable environmental changes. Detailed understanding of the circadian network of higher plants, such as Arabidopsis thaliana, is hampered by the high number of partially redundant genes. However, the picoeukaryotic alga Ostreococcus tauri, which was recently shown to possess a small number of non-redundant clock genes, presents an attractive alternative target for detailed modelling of circadian clocks in the green lineage. Based on extensive time-series data from in vivo reporter gene assays, we developed a model of the Ostreococcus clock as a feedback loop between the genes TOC1 and CCA1. The model reproduces the dynamics of the transcriptional and translational reporters over a range of photoperiods. Surprisingly, the model is also able to predict the transient behaviour of the clock when the light conditions are altered. Despite the apparent simplicity of the clock circuit, it displays considerable complexity in its response to changing light conditions. Systematic screening of the effects of altered day length revealed a complex relationship between phase and photoperiod, which is also captured by the model. The complex light response is shown to stem from circadian gating of light-dependent mechanisms. This study provides insights into the contributions of light inputs to the Ostreococcus clock. The model suggests that a high number of light-dependent reactions are important for flexible timing in a circadian clock with only one feedback loop.

Weather and seasons together demand complex biological clocks.

The 24-hour rhythms of the circadian clock [1] allow an organism to anticipate daily environmental cycles, giving it a competitive advantage [2, 3]. Although clock components show little protein sequence homology across phyla, multiple feedback loops and light inputs are universal features of clock networks [4, 5]. Why have circadian systems evolved such a complex structure? All biological clocks entrain a set of regulatory genes to the environmental cycle, in order to correctly time the expression of many downstream processes. Thus the question becomes: What aspects of the environment, and of the desired downstream regulation, are demanding the observed complexity? To answer this, we have evolved gene regulatory networks in silico, selecting for networks that correctly predict particular phases of the day under light/dark cycles. Gradually increasing the realism of the environmental cycles, we have tested the networks for the minimal characteristics of clocks observed in nature: oscillation under constant conditions, entrainment to light signals, and the presence of multiple feedback loops and light inputs. Realistic circadian gene networks are found to require a nontrivial combination of conditions, with seasonal differences in photoperiod as a necessary but not sufficient component.

Is transcriptional regulation of metabolic pathways an optimal strategy for fitness?

BACKGROUND: Transcriptional regulation of the genes in metabolic pathways is a highly successful strategy, which is virtually universal in microorganisms. The lac operon of E. coli is but one example of how enzyme and transporter production can be made conditional on the presence of a nutrient to catabolize. METHODOLOGY: With a minimalist model of metabolism, cell growth and transcriptional regulation in a microorganism, we explore how the interaction between environmental conditions and gene regulation set the growth rate of cells in the phase of exponential growth. This in silico model, which is based on biochemical rate equations, does not describe a specific organism, but the magnitudes of its parameters are chosen to match realistic values. Optimizing the parameters of the regulatory system allows us to quantify the fitness benefit of regulation. When a second nutrient and its metabolic pathway are introduced, the system must further decide whether and how to activate both pathways. CONCLUSIONS: Even the crudest transcriptional network is shown to substantially increase the fitness of the organism, and this effect persists even when the range of nutrient levels is kept very narrow. We show that maximal growth is achieved when pathway activation is a more or less steeply graded function of the nutrient concentration. Furthermore, we predict that bistability of the system is a rare phenomenon in this context, but outline a situation where it may be selected for.

Transcriptional dynamics of the embryonic stem cell switch.

Recent ChIP experiments of human and mouse embryonic stem cells have elucidated the architecture of the transcriptional regulatory circuitry responsible for cell determination, which involves the transcription factors OCT4, SOX2, and NANOG. In addition to regulating each other through feedback loops, these genes also regulate downstream target genes involved in the maintenance and differentiation of embryonic stem cells. A search for the OCT4-SOX2-NANOG network motif in other species reveals that it is unique to mammals. With a kinetic modeling approach, we ascribe function to the observed OCT4-SOX2-NANOG network by making plausible assumptions about the interactions between the transcription factors at the gene promoter binding sites and RNA polymerase (RNAP), at each of the three genes as well as at the target genes. We identify a bistable switch in the network, which arises due to several positive feedback loops, and is switched on/off by input environmental signals. The switch stabilizes the expression levels of the three genes, and through their regulatory roles on the downstream target genes, leads to a binary decision: when OCT4, SOX2, and NANOG are expressed and the switch is on, the self-renewal genes are on and the differentiation genes are off. The opposite holds when the switch is off. The model is extremely robust to parameter changes. In addition to providing a self-consistent picture of the transcriptional circuit, the model generates several predictions. Increasing the binding strength of NANOG to OCT4 and SOX2, or increasing its basal transcriptional rate, leads to an irreversible bistable switch: the switch remains on even when the activating signal is removed. Hence, the stem cell can be manipulated to be self-renewing without the requirement of input signals. We also suggest tests that could discriminate between a variety of feedforward regulation architectures of the target genes by OCT4, SOX2, and NANOG.

An introduction to BioArray Software Environment.

BioArray Software Environment (BASE) is a web-based software package for storing, searching, and analyzing locally generated microarray data and information surrounding microarray production. The workflow begins in sample management and, optionally, microtiter plate tracking and ends in visualization and analysis of entire experiments. The relative ease with which new analysis plug-ins can be added has given rise to a plethora of third-party tools, and the licensing terms (GNU GPL) encourage local modifications of the software. This introduction to BASE describes the basics of working with the software, both in general and in more detail for the various parts. It also provides some hints about more advanced usage and a section on what is needed to set up your own BASE server. The information is current as of BASE version 1.2.17b, which was released on November 6, 2005.