Noise propagation and scaling in regulation of gonadotrope biosynthesis

Reproductive physiology depends on the control of biosynthesis in the pituitary gonadotrope by hypothalamic gonadotropin-releasing hormone (GnRH). The responses to GnRH include activation of extracellular signal-regulated kinase (ERK) and induction of Egr1. Using population and single cell signaling assays, we investigated the signal and noise transmission through this signaling and gene circuit, analyzing data obtained from 43,775 individual cells in 40 experiments. After exposure to GnRH, phosphorylated ERK (pERK) is elevated in 50% of the cells at 1.7 (SD = 0.3) min. Studies of the cell-to-cell response showed that for both pERK and for Egr1 protein production the mean response (μ) and standard deviation (σ) within individual cells were linearly related (σ = kμ) and had similar values of k. To understand the basis for the scaling observed for noise propagation through this system, we determined the relationship between pERK and egr1 mRNA levels induced at varying concentration of GnRH. While both pERK and egr1 mRNA show a saturating sigmoidal relationship to the concentration of GnRH exposure, egr1 mRNA is linearly related to the levels of pERK. These results explain the basis for variation in cellular responses in an important mammalian signaling pathway leading to gene induction.     

An improved method for cell-to-cell transmission of infectious prion

Prion diseases are characterized by the accumulation of a pathological form of prion protein (PrP^Sc), which behaves as an infectious agent. Here we developed an in vitro co-culture system to analyze the PrP^Sc transmission from ScN2a cell, which persistently retains PrP^Sc, to naïve N2a cell. In this cell-to-cell transmission system, PrP^Sc transmitted to recipient N2a cell was able to be detected within 5-7 days. Further characterization showed that higher cell density greatly facilitated the transmission of PrP^Sc. This improved in vitro transmission method may become a useful tool for unveiling the molecular mechanism of PrP^Sc transmission.     

Use of AC Impedance Analysis to Study Membrane Changes Related to Acid Secretion in Amphibian Gastric Mucosa

We have applied transepithelial AC impedance techniques to gastric mucosa to reconcile ultrastructural and electrophysiological findings about gastric acid secretion and the mucosal barrier. By fitting impedance data measured at different HCl secretion rates to equivalent circuit models, we extracted capacitances and resistances (as measures of membrane area and ionic conductance, respectively) for the apical and basolateral membranes. The impedance measurements were found to be incompatible with earlier equivalent circuit models that modeled membrane electrical properties as lumped circuits based on one or two cell types. A distributed circuit model was developed that assumed only one dominant electrical pathway (i.e., one cell type), but that incorporated electrical effects arising from long and narrow membrane-lined structures present in the epithelium (e.g., gastric crypts, tubulovesicles, lateral intercellular spaces). This morphologically based model was found to represent the measured data accurately, and to yield values for membrane capacitances consistent with morphometric measurements of membrane areas. The main physiological conclusions from this analysis were as follows: (a) The dominant transepithelial current pathway may reside in the oxyntic cells. (b) The transepithelial conductance increase associated with the onset of acid secretion is entirely due to increased conductance of the apical membrane. This is in turn due entirely to increased area of this membrane, resulting from incorporation of tubulovesicular membrane. (c) When membrane conductances are normalized to actual membrane area by use of membrane capacitances, it turns out that acid secretion is not associated with a change in specific ionic conductance (change in conductance per unit area) at either the apical or basolateral membrane. (d) The puzzlingly low value of transepithelial resistance (≤400 Ω-cm²) arises because there are hundreds or thousands of square centimeters of actual membrane area per square centimeter chamber area. Apical membrane resistance is 25 kΩ-cm² (actual membrane area), implying a tight barrier to back-diffusion of protons.     

The Potential in the Gap between Two Abutting Cardiac Muscle Cells: A Closed Solution

An approximate differential equation is developed describing the potential in the gap (intercalated disc) between two closely abutting, coaxial cylindrical cardiac muscle cells. This permits approximate calculation of the degree of current spread from an active to an inactive cell. The equation has a closed solution in terms of the zero-order Bessel function I₀(x). This result is different from one given by Woodbury and Crill (1961). The source of the original mistake is given and the magnitude of the error estimated. The new solution is compared with the exact, series solution to this problem given by Heppner and Plonsey (1970) in the preceding paper. It is shown analytically that our approximate solution differs negligibly from the series solution for the parameter values chosen. The closed solution not only considerably simplifies calculations but yields additional insights into the nature of the coupling resistances R and r used by Heppner and Plonsey in their detailed analysis of the cell-to-cell transmission process.     

Application of the Hodgkin-Huxley equations to an electric field model for interaction between excitable cells

We previously described a model for the electrical transfer of excitation from one cell to the next which utilized the electric potential generated in the junctional cleft between the cells. Low-resistance connections between the cells were not used in the model, and it was assumed that the junctional membranes were excitable. This model was analyzed for the static case without capacitances and for the dynamic case in which capacitances were part of the circuit elements. For simplicity, the Na⁺ resistance (R_Na), after a threshold potential was exceeded, was allowed to decrease exponentially (to 1% of its initial value) within 0·25–1·0 ms, and possible changes in the K⁺ resistance were ignored. In this paper, we have incorporated the Hodgkin-Huxley equations into the operation of the lumped membrane units for the electrical equivalent circuit of the cell membrane. The parameters varied are the membrane capacitances, resistances, maximum Na⁺ conductance ($\text{g}{}_{\mathrm{<mtext>Na</mtext>}}$), and the radial cleft resistance (R_jc). We demonstrated that our model worked very well, i.e. the successful transfer of action potentials was achieved, with the membrane units following Hodgkin-Huxley dynamics for changes in g_Na and g_K. The calculations indicate that transmission is facilitated when the junctional units have a higher $\text{g}{}_{\mathrm{<mtext>Na</mtext>}}$ and a lower capacitance and when R_jc is elevated. Lowering the resistance of the junctional membrane units several fold, relative to the surface membrane units, also facilitated transmission; however, the absolute resistance of the junctional membrane was still well above the maximum value that would allow sufficient local-circuit current to flow to effect transmission. Thus, the electric field model provides an alternative means of cell-to-cell propagation between myocardial cells which is electrical in nature but does not require the presence of low-resistance connections between cells.     

Relative contribution of free-virus and synaptic transmission to the spread of HIV-1 through target cell populations

Human immunodeficiency virus can spread through target cells by transmission of cell-free virus or directly from cell-to-cell via formation of virological synapses. Although cell-to-cell transmission has been described as much more efficient than cell-free infection, the relative contribution of the two transmission pathways to virus growth during multiple rounds of replication remains poorly defined. Here, we fit a mathematical model to previously published and newly generated in vitro data, and determine that free-virus and synaptic transmission contribute approximately equally to the growth of the virus population.     

Stochastic modeling of the transmission of respiratory syncytial virus (RSV) in the region of Valencia,Spain

In this paper, we study the dynamics of the transmission of respiratory syncytial virus (RSV) in the population using stochastic models. The stochastic models are developed introducing stochastic perturbations on the demographic parameter as well as on the transmission rate of the RSV. Numerical simulations of the deterministic and stochastic models are performed in order to understand the effect of fluctuating birth rate and transmission rate of the RSV on the population dynamics. The numerical solutions of stochastic models are calculated using Euler-Maruyama and Milstein schemes, and confidence intervals for stochastic solutions are given using Monte-Carlo method. Analysis of the numerical results reveals that perturbations on the transmission rate are more decisive in the dynamics of RSV than perturbations on demographic parameters. In addition, the stochastic models show the advantage of reproducing more effectively the noisy RSV hospitalization data. It is concluded that these stochastic models are a viable option to provide a realistic modeling of the RSV dynamics on the population.     

思茅松天然林树冠结构模型

以云南省普洱市思茅区思茅松天然林为研究对象,采用枝解析调查了34株思茅松样木的树冠数据,分析了一级枝枝长、枝径、着枝角度、弦长和树冠半径5个树冠形状变量的变化规律,分别构建其预估模型;分析了树冠结构变化,分别构建了一级枝轮枝高度预估模型、一级枝枝条数量预估模型和一级枝枝条数量累积预估模型,并采用独立样本进行模型统计精度检验。结果表明:8个预估模型的预测效果良好,精度达到91%以上,尤其是一级枝着枝角度模型和一级枝轮枝高度模型预测精度达到97%以上。研究结果合理准确描述思茅松树冠结构的变化,为思茅松天然林的经营管理提供科学依据。     

Estimation of the Size of Genetic Bottlenecks in Cell-to-Cell Movement of Soil-Borne Wheat Mosaic Virus and the Possible Role of the Bottlenecks in Speeding Up Selection of Variations in trans-Acting Genes or Elements

Genetic bottlenecks facilitate the fixation and extinction of variants in populations, and viral populations are no exception to this theory. To examine the existence of genetic bottlenecks in cell-to-cell movement of plant RNA viruses, we prepared constructs for Soil-borne wheat mosaic virus RNA2 vectors carrying two different fluorescent proteins, yellow fluorescent protein (YFP) and cyan fluorescent protein (CFP). Coinoculation of host plant leaves with the two RNA2 vectors and the wild-type RNA1 showed separation of the two vector RNA2s, mostly within seven to nine cell-to-cell movements from individual initially coinfected cells. Our statistical analysis showed that the number of viral RNA genomes establishing infection in adjacent cells after the first cell-to-cell movement from an initially infected cell was 5.97 ± 0.22 on average and 5.02 ± 0.29 after the second cell-to-cell movement. These results indicate that plant RNA viruses may generally face narrow genetic bottlenecks in every cell-to-cell movement. Furthermore, our model suggests that, rather than suffering from fitness losses caused by the bottlenecks, the plant RNA viruses are utilizing the repeated genetic bottlenecks as an essential element of rapid selection of their adaptive variants in trans-acting genes or elements to respond to host shifting and changes in the growth conditions of the hosts.Plant RNA viruses change their genomes so rapidly that variant viruses with altered biological properties are often found after prolonged growth of infected plants or after serial mechanical inoculations (26, 33). Furthermore, inoculation of less-fit artificial mutants produces revertants or pseudo-revertants even after short infection times (12, 14). The rapid evolution of plant RNA viral genomes is achieved not only by high mutation rates due to error-prone replication by the nonproofreading viral RNA-dependent RNA polymerase (19) but also by rapid selection and strong genetic drift. Generally, narrow genetic bottlenecks facilitate the fixation and extinction of variants in populations (15), and viral populations are no exception to this theory.Plant RNA viruses are known to face many narrow genetic bottlenecks during their life cycles (23). The life cycles of most plant RNA viruses are as follows: After replicating in cells, viruses move from cell to cell through plasmodesmata, which connect the cytoplasms of adjacent cells separated by cell walls in plant tissue. Following the establishment of infection in cells and cell-to-cell movements, the viruses expand their infected regions, spreading to the veins and moving through the vascular system and infecting the plant systemically. Some plant RNA viruses are transmitted through the seeds or via mechanical injuries, but most are transmitted from plant to plant by biological vectors such as insects, nematodes, and fungi. Previous studies have found that genetic bottlenecks occur during the transfer from lower leaves to upper leaves in systemic infections of Wheat streak mosaic virus (WSMV) (11), Tobacco mosaic virus (TMV) (24), and Cucumber mosaic virus (CMV) (18) and during the transfer from one tiller to another tiller of WSMV (11). Vector transmissions were also shown to act as genetic bottlenecks for WSMV (11), CMV (1, 3), and Potato virus Y (PVY) (20). With the exception of PVY, the typical method for detecting genetic bottlenecks has been to observe the spatial separation of closely related strains or artificial synonymous mutants inoculated as mixed populations: the narrower the genetic bottleneck, the more frequently the spatial separation should be observed. Using this idea with mathematical analyses, WSMV was estimated to infect a new tiller starting with four genomes (9), TMV was estimated to infect the upper leaves starting with 10 genomes (24), and CMV was estimated to infect a new plant starting with one to two particles after aphid transmission (3). Studies of PVY using sets of host plant cultivars with or without resistance genes and mixed strains of viruses with or without resistance-breaking abilities also estimated the number of virus particles transmitted by an aphid vector to be 0.5 to 3.2 on average (20).However, genetic bottlenecks in cell-to-cell movement of viruses have not been well characterized, although these occurrences are likely (11) and have been expected to be important for understanding the life cycle and population dynamics of plant RNA viruses. The size of genetic bottlenecks in cell-to-cell movement can be referred to as “multiplicity of infection (MOI) in plant tissue colonization,” and only a recent study showing that the estimated MOI of TMV is between 6 and 1 to 2 (10) indicates the occurrence and the size of genetic bottlenecks in cell-to-cell movement of a plant RNA virus. In this paper, we also show the occurrence of narrow genetic bottlenecks during cell-to-cell movement of a plant RNA virus, Soil-borne wheat mosaic virus (SBWMV, type species of the genus Furovirus), by observing the spatial separation of RNA2 vectors carrying different fluorescent proteins, yellow fluorescent protein (YFP) and cyan fluorescent protein (CFP). Both of the fluorescent proteins were expressed as fusion proteins to the N-terminal nuclear localization signal (NLS) peptide from Simian virus 40 (SV40) large T antigen, which enabled us to observe and count the infected cells accurately using nuclear fluorescence. Numerical data were analyzed to estimate the size of bottlenecks. We also carried out a simulation to show that, due to the narrow genetic bottlenecks, rapid selection occurs even on trans-acting elements in plant RNA virus genomes, overcoming the negative effect of complementation among adaptive and defective genomes in each intracellular population. We discuss the possible roles of the bottlenecks in the life cycle and evolution mechanisms of plant RNA viruses.     

Cell-to-cell transmission of human immunodeficiency virus type 1 in the presence of azidothymidine and neutralizing antibody.

Very few peripheral blood lymphocytes of seropositive individuals are presumably actively infected with human immunodeficiency virus type 1 (HIV-1). During coculture of lymphocytes of a seropositive individual with mitogen-stimulated normal peripheral blood lymphocytes, the number of infected cells becomes amplified such that detectable HIV-1 is produced. We report here that in addition to transmission by extracellular virus, cell-to-cell transmission is responsible for spreading HIV-1 infection from infected to uninfected cells. Azidothymidine and virus-neutralizing antibody had no effect on cell-to-cell transmission of HIV-1. Monoclonal antibodies to the CD4 receptor, but not to the CD3 receptor, prevented cell-to-cell transmission, which suggests that CD4 receptor-mediated cell fusion is involved in cell-to-cell transmission. Spread of infection in a cell-to-cell manner may be important in development of drug therapies for HIV-1 infection.     

Theoretical Conformational Analysis in the Determination of Productive Conformations of Substrates for Acetylcholinesterase and Butyrylcholinesterase

All the equilibrium conformations of 34 analogues of acetylcholine (ACh) with the general formula R-C(O)O-Alk-N⁺(CH₃)₃ are calculated by the method of molecular mechanics. In the series R-C(O)O-(CH₂)₂-N⁺(CH₃)₃, a reliable correlation is found between the molecular volume of the substrate and the rate of its hydrolysis by acetylcholinesterase (AChE); the absence of such a correlation is demonstrated for butyryl-cholinesterase (BChE). Theoretical conformational analysis confirms that the completely extended tt conformation of ACh is productive for the hydrolysis by AChE, which agrees with the results of X-ray analysis of AChE. AChE is shown to hydrolyze only those substrates that form equilibrium conformers compatible in the mutual arrangement of trimethylammonium group, carbonyl carbon, and carbonyl oxygen with the tt conformation of ACh; in this case, the rate of substrate hydrolysis depends on the total population of these conformers. A reliable correlation was found between the population of the semifolded (tg^?) conformation of the choline moiety of substrate molecules and rate of their BChE hydrolysis. In a series of CH₃-C(O)O-Alk-N⁺(CH₃)₃, the rate of BChE hydrolysis is demonstrated to depend on the total population of conformations compatible in the mutual arrangement of functionally important atoms with the tg^? conformation of ACh. The tg^? conformation of ACh is concluded to be productive for BChE hydrolysis. Similar orientations of the substrate molecules relative to the catalytic triads of both AChE and BChE are proven to coincide upon the substrate productive sorption in their active sites. It is hypothesized that the sorption stage is rate-limiting in cholinesterase hydrolysis and the enzyme hydrolyzes the ACh molecule in its energetically favorable conformation.     

Influence of the Ionic Composition of Fluid Medium on Red Cell Aggregation

The effects of ionic strength and cationic valency of the fluid medium on the surface potential and dextran-induced aggregation of red blood cells (RBC's) were investigated. The zeta potential was calculated from cell mobility in a microelectrophoresis apparatus; the degree of aggregation of normal and neuraminidase-treated RBC's in dextrans (Dx 40 and Dx 80) was quantified by microscopic observation, measurement of erythrocyte sedimentation rate, and determination of low-shear viscosity. A decrease in ionic strength caused a reduction in aggregation of normal RBC's in dextrans, but had no effect on the aggregation of neuraminidase-treated RBC's. These findings reflect an increase in electrostatic repulsive force between normal RBC's by the reduction in ionic strength due to (a) a decrease in the screening of surface charge by counter-ions and (b) an increase in the thickness of the electric double layer. Divalent cations (Ca⁺⁺, Mg⁺⁺, and Ba⁺⁺) increased aggregation of normal RBC's in dextrans, but had no effect on the aggregation of neuraminidase-treated RBC's. These effects of the divalent cations are attributable to a decrease in surface potential of normal RBC's and a shrinkage of the electric double layer. It is concluded that the surface charge of RBC's plays a significant role in cell-to-cell interactions.     

Water Transport across Maize Roots : Simultaneous Measurement of Flows at the Cell and Root Level by Double Pressure Probe Technique

A double pressure probe technique was used to measure simultaneously water flows and hydraulic parameters of individual cells and of excised roots of young seedlings of maize (Zea mays L.) in osmotic experiments. By following initial flows of water at the cell and root level and by estimating the profiles of driving forces (water potentials) across the root, the hydraulic conductivity of individual cell layers was evaluated. Since the hydraulic conductivity of the cell-to-cell path was determined separately, the hydraulic conductivity of the cell wall material could be evaluated as well (Lp_cw = 0.3 to 6.10⁻⁹ per meter per second per megapascal). Although, for radial water flow across the cortex and rhizodermis, the apoplasmic path was predominant, the contribution of the hydraulic conductance of the cell-to-cell path to the overall conductance increased significantly from the first layer of the cortex toward the inner layers from 2% to 23%. This change was mainly due to an increase of the hydraulic conductivity of the cell membranes which was Lp = 1.9.10⁻⁷ per meter per second per megapascal in the first layer and Lp = 14 to 9.10⁻⁷ per meter per second per megapascal in the inner layers of the cortex. The hydraulic conductivity of entire roots depended on whether hydrostatic or osmotic forces were used to induce water flows. Hydrostatic Lp_r was 1.2 to 2.3.10⁻⁷ per meter per second per megapascal and osmotic Lp_r = 1.6 to 2.8.10⁻⁸ per meter per second per megapascal. The apparent reflection coefficients of root cells (σ_s) of nonpermeating solutes (KCI, PEG 6000) decreased from values close to unity in the rhizodermis to about 0.7 to 0.8 in the cortex. In all cases, however, σ_s was significantly larger than the reflection coefficient of entire roots (σ_sr). For KCI and PEG 6000, σ_sr was 0.53 and 0.64, respectively. The results are discussed in terms of a composite membrane model of the root.     

Electric Impedance and Rectification of Fused Anion-Cation Membranes in Solution

At relatively high currents, fused anion-cation membranes give rise to rectifying and reactive effects. The rectification becomes less pronounced with increasing frequency. This effect results from changes in the concentration profiles of the ions during the positive and negative phases of the AC cycle. With reduction of the current, the voltage-current response becomes linear. The reactive effect can then be separated from the rectifying effect. The former effect can be attributed essentially to two factors: (a) the presence of transition regions of fixed charge and (b) the diffusion mechanism of the ions in an AC field. The first factor is largely frequency-independent and the second, frequency-dependent. A first approximation equivalent circuit is described. This circuit involves frequency-dependent elements.     

Evaluation of the Oscillatory Interference Model of Grid Cell Firing through Analysis and Measured Period Variance of Some Biological Oscillators

Models of the hexagonally arrayed spatial activity pattern of grid cell firing in the literature generally fall into two main categories: continuous attractor models or oscillatory interference models. Burak and Fiete (2009, PLoS Comput Biol) recently examined noise in two continuous attractor models, but did not consider oscillatory interference models in detail. Here we analyze an oscillatory interference model to examine the effects of noise on its stability and spatial firing properties. We show analytically that the square of the drift in encoded position due to noise is proportional to time and inversely proportional to the number of oscillators. We also show there is a relatively fixed breakdown point, independent of many parameters of the model, past which noise overwhelms the spatial signal. Based on this result, we show that a pair of oscillators are expected to maintain a stable grid for approximately t = 5µ ³ /(4πσ) ² seconds where µ is the mean period of an oscillator in seconds and σ² its variance in seconds². We apply this criterion to recordings of individual persistent spiking neurons in postsubiculum (dorsal presubiculum) and layers III and V of entorhinal cortex, to subthreshold membrane potential oscillation recordings in layer II stellate cells of medial entorhinal cortex and to values from the literature regarding medial septum theta bursting cells. All oscillators examined have expected stability times far below those seen in experimental recordings of grid cells, suggesting the examined biological oscillators are unfit as a substrate for current implementations of oscillatory interference models. However, oscillatory interference models can tolerate small amounts of noise, suggesting the utility of circuit level effects which might reduce oscillator variability. Further implications for grid cell models are discussed.     

Glycometabolism regulates hepatitis C virus release

HCV cell-culture system uses hepatoma-derived cell lines for efficient virus propagation. Tumor cells cultured in glucose undergo active aerobic glycolysis, but switch to oxidative phosphorylation for energy production when cultured in galactose. Here, we investigated whether modulation of glycolysis in hepatocytes affects HCV infection. We showed HCV release, but not entry, genome replication or virion assembly, is significantly blocked when cells are cultured in galactose, leading to accumulation of intracellular infectious virions within multivesicular body (MVB). Blockade of the MVB-lysosome fusion or treatment with pro-inflammatory cytokines promotes HCV release in galactose. Furthermore, we found this glycometabolic regulation of HCV release is mediated by MAPK-p38 phosphorylation. Finally, we showed HCV cell-to-cell transmission is not affected by glycometabolism, suggesting that HCV cell-to-supernatant release and cell-to-cell transmission are two mechanistically distinct pathways. In summary, we demonstrated glycometabolism regulates the efficiency and route of HCV release. We proposed HCV may exploit the metabolic state in hepatocytes to favor its spread through the cell-to-cell transmission in vivo to evade immune response.     

I. Novel capacitative electrode with a wide frequency range for measurements of flash-induced changes of interface potential at the oil-water interface Mechanical construction and electrical characteristics of the electrode

The mechanical construction and the electrical properties of a new type of capacitative electrode for the oil-water interface are described. The electrode is designed to detect changes of the interface potential induced by photochemical, photophysical, and photobiological reactions occurring at the interface. The construction is based on capacitative coupling of two aqueous compartments separated by a thin Teflon film. Thereby, the oil-water interface is in horizontal position and the electrode is placed with its planar bottom about 10 μm above the interface. A main feature of the electrode is the transparency to visible light which is achieved by having a clear electrolyte solution in the inner compartment of the capacitative electrode.The aqueous subphase and the inner electrolyte are connected with AglAgCl electrodes to voltage amplifiers. The capacitative electrode is best operated under open circuit conditions. The frequency range experimentally verified is 500 $\text{MHz}\text{\$}\text{?}\text{?}{}_{\mathrm{<mtext>3</mtext><mtext>dB</mtext>}}\text{\$}\text{?}\text{0.03}\text{Hz}$. The sensitivity is mainly determined by the noise of the electronic amplifiers, typical 50–100 μV.     

Machine Learning Prediction of Cancer Cell Sensitivity to Drugs Based on Genomic and Chemical Properties

Predicting the response of a specific cancer to a therapy is a major goal in modern oncology that should ultimately lead to a personalised treatment. High-throughput screenings of potentially active compounds against a panel of genomically heterogeneous cancer cell lines have unveiled multiple relationships between genomic alterations and drug responses. Various computational approaches have been proposed to predict sensitivity based on genomic features, while others have used the chemical properties of the drugs to ascertain their effect. In an effort to integrate these complementary approaches, we developed machine learning models to predict the response of cancer cell lines to drug treatment, quantified through IC₅₀ values, based on both the genomic features of the cell lines and the chemical properties of the considered drugs. Models predicted IC₅₀ values in a 8-fold cross-validation and an independent blind test with coefficient of determination R² of 0.72 and 0.64 respectively. Furthermore, models were able to predict with comparable accuracy (R² of 0.61) IC50s of cell lines from a tissue not used in the training stage. Our in silico models can be used to optimise the experimental design of drug-cell screenings by estimating a large proportion of missing IC₅₀ values rather than experimentally measuring them. The implications of our results go beyond virtual drug screening design: potentially thousands of drugs could be probed in silico to systematically test their potential efficacy as anti-tumour agents based on their structure, thus providing a computational framework to identify new drug repositioning opportunities as well as ultimately be useful for personalized medicine by linking the genomic traits of patients to drug sensitivity.