We developed a computational platform including machine learning and a mechanistic mathematical model to find the optimal protocol for administration of platinum-doublet chemotherapy in a palliative setting. The platform has been applied to advanced metastatic non-small cell lung cancer (NSCLC). The 42 NSCLC patients treated with palliative intent at Maria Sklodowska-Curie National Research Institute of Oncology, Gliwice Branch, were collected from a retrospective cohort of patients diagnosed in 2004-2014. Patients were followed-up, for three years. Clinical data collected include complete information about the clinical course of the patients including treatment schedule, response according to RECIST classification, and survival. The core of the platform is the mathematical model, in the form of a system of ordinary differential equations, describing dynamics of platinum-sensitive and platinum-resistant cancer cells and interactions reflecting competition for space and resources. The model is simulated stochastically by sampling the parameter values from a joint probability distribution function. The machine learning model is applied to calibrate the mathematical model and to fit it to the overall survival curve. The model simulations faithfully reproduce the clinical cohort at three levels long-term response (OS), the initial response (according to RECIST criteria), and the relationship between the number of chemotherapy cycles and time between two consecutive chemotherapy cycles. In addition, we investigated the relationship between initial and long-term response. We showed that those two variables do not correlate which means that we cannot predict patient survival solely based on the initial response. We also tested several chemotherapy schedules to find the best one for patients treated with palliative intent. We found that the optimal treatment schedule depends, among others, on the strength of competition among various subclones in a tumor. The computational platform developed allows optimizing chemotherapy protocols, within admissible limits of toxicity, for palliative treatment of metastatic NSCLC. The simplicity of the method allows its application to chemotherapy optimization in different cancers.

The cell cycle is the fundamental process of cell populations, it is regulated by environmental cues and by intracellular checkpoints. Cell cycle variability in clonal cell population is caused by stochastic processes such as random partitioning of cellular components to progeny cells at division and random interactions among biomolecules in cells. One of the important biological questions is how the dynamics at the cell cycle scale, which is related to family dependencies between the cell and its descendants, affects cell population behavior in the long-run. We address this question using a "mechanistic" model, built based on observations of single cells over several cell generations, and then extrapolated in time. We used cell pedigree observations of NIH 3T3 cells including FUCCI markers, to determine patterns of inheritance of cell-cycle phase durations and single-cell protein dynamics. Based on that information we developed a hybrid mathematical model, involving bifurcating autoregression to describe stochasticity of partitioning and inheritance of cell-cycle-phase times, and an ordinary differential equation system to capture single-cell protein dynamics. Long-term simulations, concordant with in vitro experiments, demonstrated the model reproduced the main features of our data and had homeostatic properties. Moreover, heterogeneity of cell cycle may have important consequences during population development. We discovered an effect similar to genetic drift, amplified by family relationships among cells. In consequence, the progeny of a single cell with a short cell cycle time had a high probability of eventually dominating the population, due to the heritability of cell-cycle phases. Patterns of epigenetic heritability in proliferating cells are important for understanding long-term trends of cell populations which are either required to provide the influx of maturing cells (such as hematopoietic stem cells) or which started proliferating uncontrollably (such as cancer cells).

Cancer development is driven by series of events involving mutations, which may become fixed in a tumor via genetic drift and selection. This process usually includes a limited number of driver (advantageous) mutations and a greater number of passenger (neutral or mildly deleterious) mutations. We focus on a real-world leukemia model evolving on the background of a germline mutation. Severe congenital neutropenia (SCN) evolves to secondary myelodysplastic syndrome (sMDS) and/or secondary acute myeloid leukemia (sAML) in 30-40%. The majority of SCN cases are due to a germline ELANE mutation. Acquired mutations in CSF3R occur in >70% sMDS/sAML associated with SCN. Hypotheses underlying our model are: an ELANE mutation causes SCN; CSF3R mutations occur spontaneously at a low rate; in fetal life, hematopoietic stem and progenitor cells expands quickly, resulting in a high probability of several tens to several hundreds of cells with CSF3R truncation mutations; therapeutic granulocyte colony-stimulating factor (G-CSF) administration early in life exerts a strong selective pressure, providing mutants with a growth advantage. Applying population genetics theory, we propose a novel two-phase model of disease development from SCN to sMDS. In Phase 1, hematopoietic tissues expand and produce tens to hundreds of stem cells with the CSF3R truncation mutation. Phase 2 occurs postnatally through adult stages with bone marrow production of granulocyte precursors and positive selection of mutants due to chronic G-CSF therapy to reverse the severe neutropenia. We predict the existence of the pool of cells with the mutated truncated receptor before G-CSF treatment begins. The model does not require increase in mutation rate under G-CSF treatment and agrees with age distribution of sMDS onset and clinical sequencing data.

BACKGROUND: Because driver mutations provide selective advantage to the mutant clone, they tend to occur at a higher frequency in tumor samples compared to selectively neutral (passenger) mutations. However, mutation frequency alone is insufficient to identify cancer genes because mutability is influenced by many gene characteristics, such as size, nucleotide composition, etc. The goal of this study was to identify gene characteristics associated with the frequency of somatic mutations in the gene in tumor samples. RESULTS: We used data on somatic mutations detected by genome wide screens from the Catalog of Somatic Mutations in Cancer (COSMIC). Gene size, nucleotide composition, expression level of the gene, relative replication time in the cell cycle, level of evolutionary conservation and other gene characteristics (totaling 11) were used as predictors of the number of somatic mutations. We applied stepwise multiple linear regression to predict the number of mutations per gene. Because missense, nonsense, and frameshift mutations are associated with different sets of gene characteristics, they were modeled separately. Gene characteristics explain 88% of the variation in the number of missense, 40% of nonsense, and 23% of frameshift mutations. Comparisons of the observed and expected numbers of mutations identified genes with a higher than expected number of mutations- positive outliers. Many of these are known driver genes. A number of novel candidate driver genes was also identified. CONCLUSIONS: By comparing the observed and predicted number of mutations in a gene, we have identified known cancer-associated genes as well as 111 novel cancer associated genes. We also showed that adding the number of silent mutations per gene reported by genome/exome wide screens across all cancer type (COSMIC data) as a predictor substantially exceeds predicting accuracy of the most popular cancer gene predicting tool - MutsigCV.

Elevated temperature induces the heat shock (HS) response, which modulates cell proliferation, apoptosis, the immune and inflammatory responses. However, specific mechanisms linking the HS response pathways to major cellular signaling systems are not fully understood. Here we used integrated computational and experimental approaches to quantitatively analyze the crosstalk mechanisms between the HS-response and a master regulator of inflammation, cell proliferation, and apoptosis the Nuclear Factor κB (NF-κB) system. We found that populations of human osteosarcoma cells, exposed to a clinically relevant 43°C HS had an attenuated NF-κB p65 response to Tumor Necrosis Factor α (TNFα) treatment. The degree of inhibition of the NF-κB response depended on the HS exposure time. Mathematical modeling of single cells indicated that individual crosstalk mechanisms differentially encode HS-mediated NF-κB responses while being consistent with the observed population-level responses. In particular "all-or-nothing" encoding mechanisms were involved in the HS-dependent regulation of the IKK activity and IκBα phosphorylation, while others involving transport were "analogue". In order to discriminate between these mechanisms, we used live-cell imaging of nuclear translocations of the NF-κB p65 subunit. The single cell responses exhibited "all-or-nothing" encoding. While most cells did not respond to TNFα stimulation after a 60 min HS, 27% showed responses similar to those not receiving HS. We further demonstrated experimentally and theoretically that the predicted inhibition of IKK activity was consistent with the observed HS-dependent depletion of the IKKα and IKKβ subunits in whole cell lysates. However, a combination of "all-or-nothing" crosstalk mechanisms was required to completely recapitulate the single cell data. We postulate therefore that the heterogeneity of the single cell responses might be explained by the cell-intrinsic variability of HS-modulated IKK signaling. In summary, we show that high temperature modulates NF-κB responses in single cells in a complex and unintuitive manner, which needs to be considered in hyperthermia-based treatment strategies.

BACKGROUND: Detection of tandem duplication within coding exons, referred to as internal tandem duplication (ITD), remains challenging due to inefficiencies in alignment of ITD-containing reads to the reference genome. There is a critical need to develop efficient methods to recover these important mutational events. RESULTS: In this paper we introduce ITD Assembler, a novel approach that rapidly evaluates all unmapped and partially mapped reads from whole exome NGS data using a De Bruijn graphs approach to select reads that harbor cycles of appropriate length, followed by assembly using overlap-layout-consensus. We tested ITD Assembler on The Cancer Genome Atlas AML dataset as a truth set. ITD Assembler identified the highest percentage of reported FLT3-ITDs when compared to other ITD detection algorithms, and discovered additional ITDs in FLT3, KIT, CEBPA, WT1 and other genes. Evidence of polymorphic ITDs in 54 genes were also found. Novel ITDs were validated by analyzing the corresponding RNA sequencing data. CONCLUSIONS: ITD Assembler is a very sensitive tool which can detect partial, large and complex tandem duplications. This study highlights the need to more effectively look for ITD's in other cancers and Mendelian diseases.

Recently, Haseltine et al. published a mathematical model of dynamics of viral infection that has consisted of reaction-diffusion type differential equations for wild-type and infected cells, virions and interferon. The model serves as a mathematical

MOTIVATION: Interferon-beta induced JAK-STAT signaling pathways contribute to mucosal immune recognition and an anti-viral state. Though the main molecular mechanisms constituting these pathways are known, neither the detailed structure of the regulatory network, nor its dynamics has yet been investigated. The objective of this work is to build a mathematical model for the pathway that would serve two purposes: (1) to reproduce experimental results in simulation of both early and late response to Interferon-beta stimulation and (2) to explain experimental phenomena generating new hypotheses about regulatory mechanisms that cannot yet be tested experimentally. RESULTS: Experimentally determined time dependent changes in the major components of this pathway were used to build a mathematical model describing pathway dynamics in the form of ordinary differential equations. The experimental results suggested existence of unknown negative control mechanisms that were tested numerically using the model. Together, experimental and numerical data show that the epithelial JAK-STAT pathway might be subjected to previously unknown dynamic negative control mechanisms: (1) activation of dormant phosphatases and (2) inhibition of nuclear import of IRF1.

BACKGROUND: Structural genomics projects such as the Protein Structure Initiative (PSI) yield many new structures, but often these have no known molecular functions. One approach to recover this information is to use 3D templates - structure-function motifs that consist of a few functionally critical amino acids and may suggest functional similarity when geometrically matched to other structures. Since experimentally determined functional sites are not common enough to define 3D templates on a large scale, this work tests a computational strategy to select relevant residues for 3D templates. RESULTS: Based on evolutionary information and heuristics, an Evolutionary Trace Annotation (ETA) pipeline built templates for 98 enzymes, half taken from the PSI, and sought matches in a non-redundant structure database. On average each template matched 2.7 distinct proteins, of which 2.0 share the first three Enzyme Commission digits as the template's enzyme of origin. In many cases (61%) a single most likely function could be predicted as the annotation with the most matches, and in these cases such a plurality vote identified the correct function with 87% accuracy. ETA was also found to be complementary to sequence homology-based annotations. When matches are required to both geometrically match the 3D template and to be sequence homologs found by BLAST or PSI-BLAST, the annotation accuracy is greater than either method alone, especially in the region of lower sequence identity where homology-based annotations are least reliable. CONCLUSION: These data suggest that knowledge of evolutionarily important residues improves functional annotation among distant enzyme homologs. Since, unlike other 3D template approaches, the ETA method bypasses the need for experimental knowledge of the catalytic mechanism, it should prove a useful, large scale, and general adjunct to combine with other methods to decipher protein function in the structural proteome.

BACKGROUND: The NF-kappaB regulatory network controls innate immune response by transducing variety of pathogen-derived and cytokine stimuli into well defined single-cell gene regulatory events. RESULTS: We analyze the network by means of the model combining a deterministic description for molecular species with large cellular concentrations with two classes of stochastic switches: cell-surface receptor activation by TNFalpha ligand, and IkappaBalpha and A20 genes activation by NF-kappaB molecules. Both stochastic switches are associated with amplification pathways capable of translating single molecular events into tens of thousands of synthesized or degraded proteins. Here, we show that at a low TNFalpha dose only a fraction of cells are activated, but in these activated cells the amplification mechanisms assure that the amplitude of NF-kappaB nuclear translocation remains above a threshold. Similarly, the lower nuclear NF-kappaB concentration only reduces the probability of gene activation, but does not reduce gene expression of those responding. CONCLUSION: These two effects provide a particular stochastic robustness in cell regulation, allowing cells to respond differently to the same stimuli, but causing their individual responses to be unequivocal. Both effects are likely to be crucial in the early immune response: Diversity in cell responses causes that the tissue defense is harder to overcome by relatively simple programs coded in viruses and other pathogens. The more focused single-cell responses help cells to choose their individual fates such as apoptosis or proliferation. The model supports the hypothesis that binding of single TNFalpha ligands is sufficient to induce massive NF-kappaB translocation and activation of NF-kappaB dependent genes.

The paper concerns the problem of fitting mathematical models of cell signaling pathways. Such models frequently take the form of sets of nonlinear ordinary differential equations. While the model is continuous in time, the performance index used in the fitting procedure, involves measurements taken at discrete time moments. Adjoint sensitivity analysis is a tool, which can be used for finding the gradient of a performance index in the space of parameters of the model. In the paper a structural formulation of adjoint sensitivity analysis called the Generalized Backpropagation Through Time (GBPTT) is used. The method is especially suited for hybrid, continuous-discrete time systems. As an example we use the mathematical model of the NF-kB regulatory module, which plays a major role in the innate immune response in animals.

SUMMARY: simuPOP is a forward-time population genetics simulation environment. The core of simuPOP is a scripting language (Python) that provides a large number of objects and functions to manipulate populations, and a mechanism to evolve populations forward in time. Using this R/Splus-like environment, users can create, manipulate and evolve populations interactively, or write a script and run it as a batch file. Owing to its flexible and extensible design, simuPOP can simulate large and complex evolutionary processes with ease. At a more user-friendly level, simuPOP provides an increasing number of built-in scripts that perform simulations ranging from implementation of basic population genetics models to generating datasets under complex evolutionary scenarios. AVAILABILITY: simuPOP is freely available at http://simupop.sourceforge.net, distributed under GPL license.