SIMC 2017

Permanent URI for this collection

http://hdl.handle.net/11071/10447

Browse

Now showing 1 - 20 of 85

Adaptation of a RAS pathway activation signature from FF to FFPE tissues in colorectal cancer
(Strathmore University, 2017) Omolo, Bernard; Yang, Mingli; Lo, Fang Yin; Schell, Michael J.; Austin, Sharon; Howard, Kellie; Madan, Anup; Yeatman, Timothy J.
Background: The KRAS gene is mutated in about 40 % of colorectal cancer (CRC) cases, which has been clinically validated as a predictive mutational marker of intrinsic resistance to anti-EGFR inhibitor (EGFRi) therapy. Since nearly 60 % of patients with a wild type KRAS fail to respond to EGFRicombination therapies, there is a need to develop more reliable molecular signatures to better predict response. Here we address the challenge of adapting a gene expression signature predictive of RAS pathway activation, created using fresh frozen (FF) tissues, for use with more widely available formalinfixed paraffin-embedded (FFPE) tissues. Methods: In this study, we evaluated the translation of an 18-gene RAS pathway signature score from FF to FFPE in 54 CRC cases, using a head-to- head comparison of five technology platforms. FFPE-based technologies included the Affymetrix Gene Chip (Affy), NanoString nCounter™ (NanoS),Illumina whole genome RNASeq (RNA-Acc), Illumina targeted RNASeq (t-RNA), and Illuminastranded Total RNA-rRNA- depletion (rRNA). Results: Using Affy_FF as the “gold” standard, initial analysis of the 18-gene RAS scores on all 54samples shows varying pairwise Spearman correlations, with (1) Affy_FFPE (r= 0.233, p = 0.090); (2)NanoS_FFPE (r= 0.608, p < 0.0001); (3) RNA-Acc_FFPE (r= 0.175, p = 0.21); (4) t-RNA_FFPE (r=−0.237, p = 0.085); (5) and t-RNA (r= −0.012, p = 0.93). These results suggest that only NanoString has successful FF to FFPE translation. The subsequentremoval of identified “problematic” samples (n= 15) and genes (n= 2) further improves the correlations of Affy_FF with three of the five technologies: Affy_FFPE (r= 0.672, p < 0.0001); NanoS_FFPE (r= 0.738, p < 0.0001); and RNA-Acc_FFPE (r= 0.483, p = 0.002). Conclusions: Of the five technology platforms tested, Nano String technology provides a more faithful translation of the RAS pathway gene expression signature from FF to FFPE than the Affymetrix GeneChip and multiple RNASeq technologies. Moreover, Nano String was the most forgiving technology in the analysis of samples with presumably poor RNA quality. Using this approach, the RAS signature score may now be reasonably applied to FFPE clinical samples.
Analysis of a mathematical model of influenza dynamics with drug resistance aspect
(Strathmore University, 2017) Kanyiri, Caroline W.; Kimathi, Mark
Influenza has posed a terrific public health concern. It has led to unacceptably high mortality rates especially to immune compromised persons worldwide. Efforts to effectively treat and combat the spread of influenza can be put in place if its dynamics are well understood. Numerous challenges have been faced in the event of controlling the spread and eradicating this pandemic, a major impediment being the rise of drug resistance. In light of this, a deterministic model is formulated and used to analyze the transmission dynamics of influenza having incorporated the aspect of drug resistance. A system of differential equations that models the transmission dynamics of influenza is developed. The basic reproduction number (R0) is calculated and stability of the equilibrium points analyzed. Results of the analysis show that there exists a locally stable disease free equilibrium point, E0 when R0 < 1 and a unique endemic equilibrium E*, when R0 > 1. The effect of drug resistance and transmission rate of the resistant virus on the infected and the recovered is discussed.
Analysis of categorical data in presence of latent random effects using Structural Equation Modeling: an application
(Strathmore University, 2017) Keli, Robert; Mwambi, Henry; Okango, Elphas
In most medical research, of interest is to establish the causal relationships that exist between variables which may be direct or indirect. This research intends to use structural equation modeling technique to analyze the effect of categorical latent variable(s) when assumed to follow a normal random effect model. The statistical inference is carried out under the Mplus statistical software and the developed models validated using empirical data from Kenya Aids Indicator Survey (2007).
Banana Xanthomonas Wilt dynamics with mixed cultivars in a periodic environment
(Strathmore University, 2017) Nakakawa, Juliet; Mugisha, Joseph; Shaw, Michael; Karamura, Eldad
In this study, a non-autonomous model for the spread of Banana Xanthomonas Wilt disease (BXW) in a seasonally fluctuating environment is considered. Two categories of cultivars with different susceptible levels for inflorescence infection (AAA-genome and ABB-genome (highly susceptible)) were considered. Through mathematical and numerical analysis, threshold condition for existence and stability of both the disease-free equilibrium and periodic solution were obtained. From the sensitivity analysis of key parameters with respect to time-averaged basic reproduction number (R0), it was noted that R0 increases linearly with transmission parameters and declines exponentially with roguing parameters. It was also noted that the critical roguing rate of AAA-genome cultivars was less than that of ABB genome cultivar. The peaks in disease prevalence indicate the importance of effective implementation of controls during the rainy season. Controlling same cultivars via roguing and debudding led to much lower values of R0 as compared to between different cultivars. We conclude that highly susceptible cultivars play an important role in the spread of BXW and control measures should be effectively implemented during the rainy season if BXW is to be eradicated.
Bayesian analysis of Multivariate Stochastic Volatility models
(Strathmore University, 2017) Agasa, Lameck; Shitandi, Anakalo
Multivariate stochastic volatility (MSV) models have gained applicability in Time Series (TS) data for analyzing multivariate financial and economic time series because they capture the volatility dynamics. Bayesian prior works allow analysis of MSV models to provide parsimonious skew structure and to easily scale up for high-dimensional problem. Bayesian MCMC estimation are used for high dimensional problems because it’s a very efficient estimation method, however, it is associated with a considerable computational burden when the dimensionality of the data is moderate to large. Forward-filtering backward-sampling (FFBS) algorithm by sampling is used as it considers reparameterizations. This is applied directly to heteroscedasticity estimation for latent variables. To show the effectiveness of this approach, we apply the model to a vector of daily exchange rate data from Central Bank of Kenya.
Bayesian estimation of Multivariate Stochastic Volatility by applying state space models
(Strathmore University, 2017) Agasa, Lameck; Ombasa, Kiame
This work seeks to apply a Bayesian analysis in estimating multivariate stochastic volatility (MSV) using state space models. A multiplicative model based on inverted Wishart and multivariate singular beta distributions is proposed for the evolution of the volatility, and a flexible sequential volatility updating is employed. Being computationally fast, the resulting estimation procedure is particularly suitable for on-line forecasting. Bayesian MCMC is applied to estimate high dimensional problems. Three test are conducted on estimates: the log likelihood criterion, the mean of standardized one-step forecast errors, and sequential Bayes factors. The test and procedure are applied in real data set that will comprise ten exchange rate Kenyan shillings versus other currencies in Nairobi stock exchange.
Bernoulli collocation method for solving linear multi-dimensional diusion and wave equations with Dirichlet boundary conditions
(Strathmore University, 2017) Zogheib, Bashar; Tohidi, Emran; Shateyi, Stanford
In this paper, a numerical approach is proposed for solving multi-dimensional parabolic diusion and hyperbolic wave equations subject to the appropriate initial and boundary conditions. The considered numerical solutions of these equations are considered as linear combinations of the shifted Bernoulli polynomials with unknown coecients. By collocating the main equations together with the initial and boundary conditions at some special points, equations will be transformed into the associated systems of linear algebraic equations which can be solved by robust Krylov subspace iterative methods such as GMRES. Operational matrices of dierentiation are implemented for speeding up the operations. In both of the one-dimensional and two-dimensional diffusion and wave equations, the geometrical distributions of the collocation points are depicted for clarity of presentation. Several numerical examples are provided to show the eciency and spectral (exponential) accuracy of the proposed method.
Blood perfusion flow of an electro-kinetic fluid through a porous medium with viscous dissipation
(Strathmore University, 2017) B.A., Peter; O.D, .Makinde; A.W., Ogunsola
In this work, we considered a mathematical model of an electro-kinetic fluid flow through a porous medium with blood perfusion and viscous dissipation. The fluid is assumed to poses temperature-dependent variable viscosity and thermal conductivity. The nonlinear governing partial differential equations were obtained and solved numerically using Garlekin weighted residue method coupled with fourth order Runge-Kutta technique. The results obtained were presented graphically and discussed.
Cancer cases in Kenya; forecasting incidents using box & amp; Jenkins Arima Model
(Strathmore University, 2017) Langat, Amos; Orwa, George; Koima, Joel
The aim of the study was to fit appropriate time series models in assessing the accuracy of the Box Jenkins and ARIMA model in forecasting of Cancer case admissions for all people of any age from different health facilities across the country. Box-Jenkins was selected for evaluation because it has the potential of producing a point forecast within a given population, it provides a forecast interval, and is based upon a proven model. Forecast results and their associated forecast intervals may help Health facilities and health practitioners make informed decisions about whether the number of observed cancer reports in a given timeframe represents a potential incidence or is a function of random variation. Data management and analysis were done in SPSS Software. The data was segmented into two sets: Training Set (from 2000 to 2015) and the Test Set (from 2016 to 2018). The hold out set (test) provides the gold standard for measuring the model‘s true prediction error which refers to how well the model forecasts for new data. To note, the test data were only be used after a definitive model has been selected. This was to ensure unbiased estimates of the true forecast error. The results were presented in form of tables, graphs and context. In this study, the developed model for cancer case incidents in Kenya was found to be an ARIMA (2,1,0). From the forecast available by using the developed model, it can be seen that forecasted incidents for the year 2015-16 is higher than2014-15 and in later years the incidents increases. The model can be used by researchers for forecasting of cancer incidents in Kenya.
Certain properties of the essential spectra in Banach spaces
(Strathmore University, 2017) Muhati, L. N.; Bonyo, J. O.; Agure, J. O.
Let X be an arbitrary Banach space, and L (X) be the set of all bounded linear operators on X. We apply closed range theorem to determine some duality properties of the essential spectrum on L(X). Specifically, we introduce different parts of the essential spectrum and establish their duality relations. Moreover, we look at certain algebraic properties of the essential spectra on Banach spaces. Mathematics Subject Classification (2010). Primary 47A10; Secondary47A05, 47A53.
Comparison of binary diagnostic predictors using entropy
(Strathmore University, 2017) Kathare, Alfred; Otieno, Argwings
The use of gold standard procedures in screening may be costly, risky or even unethical. It is usually therefore, not admissible for large scale application. In this case, a more acceptable diagnostic predictor is applied to a sample of subjects alongside a gold standard procedure. The performance of the predictor is then evaluated using Receiver Operating Characteristic curve. The area under the curve provide a summative measure of the performance of the predictor. The Receiver Operating Characteristic curve is a trade-off between sensitivity and specificity which in most cases are of different clinical significance. Also, the areas under the curve is criticized for lack of coherent interpretation. In this study, we proposed the use of entropy as a summary index measure of uncertainty to compare diagnostic predictors. Noting that a diseased subject who is truly identified with the disease at a lower cut-off will also be identified at a higher cut-off, we substituted time variable in survival analysis for cut-offs in a binary predictor. We then derived the entropy of the functions of diagnostic predictors. Application of the procedure to real data showed that entropy was a strong measure for quantifying the amount of uncertainty engulfed in a set of cut-offs of binary diagnostic predictor.
A Comparison of the Bayesian regression and the Ordinary Least Squares regression
(Strathmore University, 2017) Okango, Ayubu
The ordinary least squares regression model assumes that there are enough data to make inference about the parameters. For Bayesian regression however, the data are supplemented with additional information in the form of a prior probability distribution. The prior belief about the parameters is combined with the data's likelihood function according to Bayes theorem to yield the posterior belief about the parameters. The prior can take different functional forms depending on the domain and the information that is available a priori. This study uses simulated data to compare models fitted using the classical regression approach and those that obtained by the Bayesian regression technique.
A Computational investigation on the dynamics of shallow water waves
(Strathmore University, 2017) Turgut, AK
The dynamics of shallow water waves has been an active research area for the past several decades. In this context, there are several models that govern this wave flow. In this study, finite element method based on B-spline interpolation functions are successfully applied to Korteweg-de Vries equation with power law nonlinearity to examine the motion of a single solitary wave whose analytical solution is known. The stability analysis is also carried out for these waves. Also, evolution of solitons is studied with Gaussian and undular bore initial conditions.
Correlation between Malaria and Pneumonia in Kwale County, Kenya
(Strathmore University, 2017) Mungai, Esther
Malaria and Pneumonia have been the major causes of morbidities and mortilities in Kwale county in the past years. Majority of the people in the county who get pneumonia have a higher incidence of Malaria. Thus, even though Malaria is prevalent in Kwale, individuals whose Malaria incidence is lower rarely get Pneumonia. In this project we try to find whether there is any correlation between the two diseases. We investigate the factors that may contribute to this correlation. We derive that control of these factors would lead to the reduction of both diseases.
Drought prediction numerical model in North - Eastern region of Kenya
(Strathmore University, 2017) Bulinda, Vincent
Numerical models are essential tools applied in predicting drought on day to day basis that involves taking current observations of weather and processing the data with the models to forecast the future state of weather at the North-Eastern Region which includes Mandera, Marsabit, Turkana and Wajir counties. The poor long rains in 2016 affected more than 1.3 million Kenyans, which extended to early 2017, according to the government of Kenya which distributed relief food and cash on training, vaccination, animal feed and encouraging people to sell animals before they fall sick. The current weather observations serve as input to the numerical models through data assimilation to produce outputs of temperature, precipitation, and other meteorological elements. The model is focused to improve the techniques for predicting such droughts with some measure of accuracy. The research is aimed at aiding at reducing poverty and hunger in line with Millennium Development Goal 1 (MDG 1) and building a more sustainable and competitive agricultural system that will contribute to the Government of Kenya (GoK) goal of building a food secure and prosperous Kenya through a commercially-oriented and competitive agricultural sector. The results are represented on the graph using MATLAB software which showed that the application of numerical algorithms on past meteorological data can lead to accurate predictions of future agricultural drought so that future work can be based on designing a solution for multiple regions.
Dynamic TCP pacing for Delay Intolerant Cloud Communications
(Strathmore University, 2017) Nyangaresi, Vincent O.; Abeka, Silvance O.; Ogara, Solomon
In the recent years, many organizations have turned to cloud technology to support their information technology services. The cloud servers are therefore increasingly holding huge and sensitive information belonging to diverse groups of individuals and companies. Additionally, some organizations employ the cloud to provide them with online backup services. One of the most outstanding requirements for cloud customers is availability – the customers must be able to access their information and other resources stored in the cloud any time and from anywhere on the globe. This means that there should be efficient network design such that any delays are averted. The connection between the customer and the cloud can therefore be regarded as delay intolerant. Network congestions often lead to delays and packet losses. Transmission control protocol employs four congestion control algorithms– slow start, congestion avoidance, fast retransmit and fast recovery, all of which fail to meet the requirements of delay intolerance. Transmission control protocol pacing has been suggested as a possible solution to delays and packet dropping in computer networks. However, the current pacing is static in nature, meaning that constant pauses are introduced between packet transmissions to prevent bursty transmissions which can lead to delays at the receiver buffers. This paper therefore presents a dynamic pacing where the delay period is hinged on the prevailing network conditions. This dynamic pacing algorithm was designed and implemented in Spyder using Python programming language. It employed probe signals to gather network intelligence such as the applicable round trip times of the network. Thereafter, this network intelligence was employed to tailor the paces to these network conditions. The results obtained showed that this algorithm introduced longer paces when more packets are transmitted and shorter paces when few packets are transmitted. In so doing, this new algorithm gives enough time for large packets to be delivered and smaller paces when few packets are sent. The analysis was done in terms of bandwidth utilization efficiency, round trip times and congestion window size adjustments. The congestion window – time graphs and throughput – time graphs showed that the developed dynamic pacing algorithm adjusted quickly to network congestions hence ensuring that the network is efficiently utilized by averting delays.
Effect of collaborative peer learning on learner attitude and performance in Mathematics in Kenya
(Strathmore University, 2017) Amukohe, Gabriel
Learners learn best when they are actively engaged in the processing of learning information. One method of involving them in active learning is to have them learn from each other in small groups or teams. Research shows that students working in small groups tend to learn more of what is taught and retain it longer than when the same content is presented exclusively by the teacher, and to an extent, appear more satisfied with their classes (Davis 1993, Barkley, 2005). Proof of this has not been done in Kenya and Africa as a whole. There is need to do so scientifically and that is what this paper entails. A model of study will involve purposeful selection of a heterogeneous high school that incorporates the sets of mixed gender boarding and day school to widen the applicability in all the types of schools in the country. These are: • Boarding Girls High School. • Boarding Boys High School • Day School Girls High School • Day School Boys High School • Day School Gender Mixed High • Boarding Gender Mixed High The treatment group shall closely be guided and monitored to make group learning effective. After instruction the mid-term and end term exam results in all the classes shall be compared. A higher mean score of the treatment group shall mean we advocate for peer learning while a lower comparative mean would imply that peer collaborative learning is not practical in Kenya and Africa as a continent. Pre-project and post instruction survey tools shall also be analyzed to get the attitude and ability change in the learner as a result of the collaborative peer teacher guided instruction.
Empirical density estimation and back-testing of Value at Risk (VaR) from parametric volatility models
(Strathmore University, 2017) Kimundi, Gillian
This paper forecasts one-day-ahead foreign exchange volatility using parametric models and compares their empirical forecasting performance of Value at Risk of five spot exchange rates; namely, the Kenyan Shilling versus the Euro, U.S. Dollar, Japanese Yen, Great British Pound and the South Africa Rand. Univariate GARCH family models (GARCH, E-GARCH, GARCH-M and FI-GARCH) are compared against the Discrete-time Stochastic Volatility Model. The daily mean exchange rates from January 2007 to December 2016 are used. Comparison analysis is divided into in-sample and out-of-sample forecasting performance which is evaluated using exceedance-based back-testing methods of conditional coverage, independence and unconditional coverage.
An Empirical performances comparison of meta-heuristic algorithms for school bus routing problem
(Strathmore University, 2017) Semba, Sherehe; Mujuni, Egbert
School Bus Routing Problem is an NP-hard Combinatorial Optimization problem, and hence solving the School Bus Routing Problem, requires the application of one or more of the metaheuristic algorithms. This work presents a model of the School Bus Routing Problem and empirical performances comparison between three meta- heuristic algorithms namely, Simulated Annealing, Tabu Search and Ant Colony for solving a real-life School Bus Routing Problem. We have analyzed their performances in terms of computation time, efficiency and solution quality. All the three algorithms have effectively demonstrated the ability to solve the School Bus Routing Problem. The computational results show that better solution quality and fastest execution time of the Meta-heuristic algorithms depends on the number of buses and stops. The results also show that Ant Colony Algorithm produces better solution, followed by Simulated Annealing, then Tabu Search for those schools with a large number of buses and stops.
Estimating covariance matrix with high frequency data for Currency Portfolio Optimization
(Strathmore University, 2017) Chepchirchir, Rancy
The covariation between asset returns plays a crucial role in modern finance both in portfolio optimization and risk management. The paper is an effort towards estimating a covariance matrix using high-dimensional, high-frequency data (quadratic covariation) from the perspective of portfolio selection. An investors’ or a portfolio managers’ objective is to solve the portfolio optimization problem i.e. ensure risk is minimized for any given return through huse of covariance matrices and subsequent rebalancing. Besides significant increase in the sample size for estimation of the covariance matrix, use of HFD also enables for better adaptation to the local volatilities and local correlations among a vast number of assets thus an improved estimation of portfolio variance. One-minute forex data were used across the six major world currencies; i.e. EUR/USD, EUR/JPY, EUR/CHF, EUR/GBP, EUR/AUD and EUR/CAD for the period 03/01/2016 17:00 to 26/12/2016 01:38 with a robust check using out-of- sample 30-minute EUR/USD data as from 08/12/2016 01:00 to 07/02/2017 11:30. The paper employ sa quasi- maximum likelihood estimator that is rate efficient as well as consistent on the basis of time synchronization approach and compares volatility estimates to that from traditional models in which returns follow the geometric Brownian Motion. Covariance matrix estimates based on intraday returns and daily returns are constructed and evaluated. The results tend to be in favor of the high-frequency data. However, these results could be biased towards an investor or portfolio manager with a shorter rebalancing interval.

Browse

Browsing SIMC 2017 by Title

Results Per Page

Sort Options