Welcome! The seminars are held at Rice and we meet usually on Fridays from 2:00 pm to 3:00 pm in Duncan Hall. All times and dates are given below. A Google Calendar of events has also been included below to see future events that might not be posted yet. If you have any questions, please feel free to contact Stephanie Hicks at sch1@rice.edu .
20 November 2009
2:15pm
in 1042 DH
Dennis Cox
Title:
Abstract:
For more information, contact:
dcox@rice.edu
6 November 2009
2:15pm
in 1042 DH
Roberto Bertolusso
Title:
Abstract:
For more information, contact:
roberto.bertolusso@rice.edu
16 October 2009
2:15pm
in 1042 DH
Bo Peng
Title:
Abstract:
For more information, contact:
bpeng@mdanderson.org
18 September 2009
2pm
in 1049 DH
Billy Hered
Title: Using Celltracker
Abstract: Celltracker is a program designed to extract intensity data from a series of cell images. Although the program is able to provide useful information about the relative intensities
of the nucleus and cytoplasm over time, gathering the data is a very time intensive task. This is due mainly to problems with cell boundary detection and declaring the Region of Interest. When the relative
intensities of the nucleus and the cytoplasm were plotted against time, a clear periodic oscillation could be observed in most of the trials. However, a fair amount of noise was also observed in the data.
For more information, contact:
wah1@rice.edu
4 September 2009
2:15pm
in 1042 DH
Marta Iwanaszko
Title:
Evolutionary Conservation of Transcription Factor Binding Site in
the Promoter Regions of NF-KB-dependent Genes
Abstract:
N-FKB family plays a prominent role in innate (early) immune
response and have impact on other processes as cell cycle activation
or cell apoptosis. Upon stimulation by pathogens such as viral RNA a
kinase cascade is activated, which eventually strips the NF-KB of
its inhibitor IKBα molecule and allows it to translocate into the
nucleus. Once in the nucleus, it activates transcription of
approximately 90genes, some of which trigger further stages of the
immune response. NF-KB-dependent genes can be categorized, based on
the timing of their activation counted from N-FKB translocation into
the nucleus, as early, middle and late genes.. It is not obvious
what mechanism is responsible for segregation of the genes’ timing
of transcriptional response. One likely hypothesis might be that the
later the gene is, the more co-factors are required to activate it,
but late response maybe also connected with need of chromatin
remodeling. It is likely that the differences in timing are
reflected in differences in the structure of promoter regions of
genes in different categories. Specifically, this might concern
differences in number and type of transcription factor binding
motifs, required for NF-KB itself as well as for the putative
co-factors. This issue is best considered in the evolutionary
framework, first, since functional binding sites are likely to be
conserved in evolution, and second, since the patterns of
evolutionary change of promoter regions are not very well-known and
are of serious interest.
For more information, contact:
miwanaszko@gmail.com
28 August 2009
2:15
pm in 1042 DH
Dajiang Liu
Title:
Statistical Genomics of Rare Variants
Abstract: There is
strong support that quantitative phenotypes can be due to rare
variants. Next generation sequencing makes it possible to identify
rare variants, which is the first step to carry-out direct mapping
association studies of quantitative trait loci (QTL). In designing
association studies to identify rare variant QTL (rvQTL), sampling
individuals with extreme quantitative trait values (QTVs) can be
used to enrich for rare causal variants. In particular rare causal
variants are more likely to be aggregated in related individuals
with extreme phenotypes. Although extreme trait ascertainment in
related individuals has had only limited success in linkage and
association studies, this design is highly advantageous in the
analysis of rare variants, because the success of rvQTL mapping is
highly dependent on causal rare variant enrichment. Collecting a
specific family type (e.g. discordant sib-pairs) can be difficult
therefore it is desirable to be able to combine different family
structures as well as unrelated individuals in the analysis of
rvQTL. Although methods do exist to analyze QTL due to common
variants, these methods are underpowered when applied to the
analysis of rvQTL data, and inflexible for modeling relative
phenotype correlations due to multiple shared rare variants.
Additionally these methods cannot combine the analysis of QTL data
from related and unrelated individuals. We propose a flexible
likelihood framework with mixed effects for modeling extreme trait
genetic associations with rvQTL (MEGA-rvQTL) for the analysis of
related and unrelated individuals with extreme QTVs. MEGA-rvQTL
detects associations with rvQTL through likelihood ratio tests. A
unique feature of MEGA-rvQTL is that parameters of genetic interests
such as heritability and sibling residual correlation can be
efficiently estimated. We investigated the power of the MEGA-rvQTL
method, for 7 commonly used prospective selective sampling
strategies. Simulation was carried out via coalescence theory using
parameters estimated from population genetic data and heritabilities
from clinically relevant quantitative traits. We demonstrate that
analyzing sibpairs with extreme QTVs or using one sib per sibpair
with extreme QTVs are consistently more powerful than using
unrelated individuals with extreme QTVs. In conclusion, MEGA-rvQTL
is a powerful approach to analyze next generation sequence data to
map QTL due to rare variants by combining data from related and
unrelated individuals with extreme phenotypes.
For more
information, contact: dajiang.liu@gmail.com
21 August 2009
2pm
in 1044 DH
Stephanie Hicks
Title:
Predicting Functionality of Missense Mutations with Varying Levels
of Evolutionary Depth
Abstract: We investigate three
computational algorithms, SIFT, Polyphen, and Align-GVGD, that
predict the impact of missense mutations on protein structure and
function. Past research has shown a high predictive value for
methods that use evolutionary sequence conservation, surprisingly
with or without protein structural information. The goal of the
study is to investigate the prediction accuracy of functionality of
missense mutations by using both curated and uncurated alignments of
varying evolutionary depth.
For more information, contact:
sch1@rice.edu
14 August 2009
3pm
in 3076 DH
Kevin Dehoff
Title:Image
Processing and Live Cell Tracking
Abstract: An
introduction to image processing techniques, object tracking, and
how these methods are implemented in the Cell Profiler and Cell
Tracker programs.
For more information, contact:
kdhoff@rice.edu
7 August 2009
2pm
in 3076 DH
Biao Li
Title:
Abstract: Theory of multi-type age dependent branching
process with immigration. Review of techniques in stathmokinetic
experiments and discussion of its application in estimate of cell
cycle length parameters in neurogenic system.
For more
information, contact: lb4@rice.edu
3 August 2009
11am
in 1044 DH
Rosa Banuelos
Title:
Abstract:
For more information, contact:
banuelos@rice.edu
13 July 2009
2pm
in 1044 DH
Deborah Goldwasser
Title:
Stochastic Modeling of Lung Cancer Screening
Abstract: We
evaluate data on tumor size and stage from two data-sets on lung
cancer screening, namely the Mayo Lung Project and the Mayo CT
study, in order to examine evidence for and against the bipartite
model of lung cancer progression. An initial homogeneity analysis of
the two data-sets suggests the presence of two or more clusters
having distinct expected mean size at stage transitions. We use a
model that parameterizes tumor growth and stage transition by the
evolutionary parameters of branching fraction (f) and the cell
mutation rate (u) in order to simulate likelihoods for cancers from
both the MLP and the Mayo CT study. Clustering of MLP cancers based
on a likelihood-derived similarity matrix indicates the presence of
two distinct clusters (A and B) having expected mean size at stage
transition of 94 mm (CI: 64,100) and 6.1 mm (CI: 5,20) respectively.
The tumor stage and size distribution evident in the Mayo CT results
are consistent with a reduction in representation of Cluster A
cancers due to higher sensitivity of the prevalence CT screen and
detection of Cluster B cancers at smaller sizes. These results
indicate that a reduction in lung cancer mortality attributable to
CT may require a lower detection threshold than previously believed.
For more information, contact: dlg2004@rice.edu
19 June 2009
2pm
in 3110 DH
Xing Chen
Title:
Correlation of tumor size and epidemiological characteristics in
patients with non-small cell lung cancer: a preliminary study of
tumor size data
Abstract: Background: Tumor size at
diagnosis is not only an indicator for the treatment and prognosis
of lung cancer, but also an important predictor for the tumor
growth, metastasis and survival in patients. To our knowledge, the
distribution of tumor size at diagnosis may reveal the nature
history of lung cancer. So it’s of great importance to find out
what factors would influence the tumor size at diagnosis.
Material
and methods: Recently, we reviewed the electronic medical records of
about 1400 patients in M.D. Anderson Cancer Center to collect the
information about tumor characteristics (with the tumor size of
radiology and pathology) and patient characteristics (including
smoking history and vital status with last contact date). We used
T-test and Kruskal-Wallis test method to explore the relationship
between the tumor size at diagnosis and the epidemiological
factors.
Results: Smoking may affect the tumor size at the time
of diagnosis which leads to larger tumor at presentation (P<0.05).
There was a difference in tumor size by gender and age (Male 4.40cm
VS female 3.61cm, mean, P<0.01; Age>65 3.83cm VS age<65
4.26cm, mean, P<0.01). Besides, lung cancer with emphysema always
have smaller tumor at presentation than those whom don’t have this
condition (3.62cm of emphysema VS 4.15cm of non-emphysema, mean,
P<0.01). The tumor histological results such as cancer cell type
(Adenocarcinoma 3.40cm VS squamous cell carcinoma 4.45cm, mean,
P<0.01) and degrees of differentiation (Poorly to not
differentiated 4.28cm VS moderately to well 3.71cm, mean, P<0.01)
have relationship with tumor size at presentation. With the
limitation of ethnic groups (only 147 black compared to 974 white),
the Africa American have significant larger tumor (P<0.01) at
diagnose.
Conclusions: According to the data set, tumor size at
the time of diagnosis is associated with smoking, gender, age,
ethnicity, tumor histology, and presence of emphysema.
For
more information, contact: XChen6@mdanderson.org
8 June 2009
2pm
in 1044 DH
Xiaowei Wu
Title:
Bayesian Analysis of Array CGH Data
Abstract: We propose
a Bayesian approach to analyze array comparative genomic
hybridization (CGH) data. Different from most currently available
methods, this new approach builds a Bayesian hierarchical model for
the data and use a reversible jump Markov chain Monte Carlo (MCMC)
algorithm for posterior sampling. The estimated parameters are used
to identify copy number changes (gains/losses) in the target DNA
sequence. This method is computationally efficient and is flexible
to model various covariance structures of the data. Moreover, it
also has great advantage in analyzing recurrent copy number
alterations (CNAs) in multiple arrays. Simulation study shows a good
performance of the proposed method. As a real data analysis example,
we apply the method to publicly available Corriel cell lines data
and obtain satisfying results.
For more information,
contact: xwwu@rice.edu
22 May 2009
3:15
in 1044 DH
Pawel Paszek
Title:
Oscillations and feedback regulation in the NF-KB signaling
Abstract: Feedback regulation is a common structural
motif controlling dynamics of genetic networks. Here we discuss how
negative and positive feedbacks contribute to the control of
temporal and spatial activity of transcription factor NF-KappaB in
immune response and inflammation. We theoretically explore how the
NF-KB system achieves an intricate level of control by using
redundant negative feedback loops due to IKBalpha and IKBesilon
synthesis. We show that the transcriptional time-delay between
activation of IKBalpha and IKBepsilon feedback loops increases the
total level of intrinsic noise in the system and thus modulates the
heterogeneity of NF-KB oscillation timing between individual cells.
Redundancy in IKB feedback was predicted to allow the system
achieving more robust responses, through decreasing the sensitivity
to parameter variation. These data suggest a general mechanism for
signalling system robustness and control which may also be important
in other signalling networks.
For more information, contact:
paszek@liv.ac.uk
Centre for Cell Imaging, School of
Biological Sciences, The Biosciences Building, University of
Liverpool, Crown St., Liverpool L69 7ZB, UK, Tel: 0151 795 4029
15 May 2009
2pm
in 1044 DH
Biao Li
Title:
Review of Branching stochastic processes with immigration in
analysis of renewing cell populations by Dr. Andrei Yakovlev
Abstract: The talk will present and review a paper,
Branching stochastic processes with immigration in analysis of
renewing cell populations, which is written by Dr. Andrei Yakovlev.
The abstract is available on this webpage.
I will also summarize recent work on computation and programming of
model of differentiating neuron stem cells.
Presentation
For more information, contact: lb4@rice.edu
8 May 2009
10am
in 1075 DH
Millena Foy
Title:
Lung Carcinogenesis Modeling and it's Application
Abstract:
Lung cancer is the second leading cancer in terms of incidence for
both men and women. However, because of its serious health
implications, lung cancer is the leading cancer killer for both men
and women. It is well known that smoking is a major risk factor for
lung cancer. The two-stage clonal expansion (TSCE) will be fit to
data to determine the effects of smoking on the risk of developing
lung cancer. This model is then used in the Smoking Base Case
project, and in an analysis of the effectiveness of CT screening in
the reduction of lung cancer mortality.
The TSCE model is
traditionally fit to prospective cohort data. A new method is
introduced that allows for the reconstruction of cohort data from
the combination of risk factor data from a case-control data, and
tabled incidence/mortality rate data. Simulation study reveals that
the results of fitting the model using this new method are
reasonable. The method is used to fit a TSCE model based on smoking
history.
The model is then applied to the smoking base case
project in conjunction with the Smoking Base Case project of the
CISNET lung group. The results of simulations show that the model
reasonably reconstructs US mortality.
Lung carcinogenesis
modeling provides a framework for simulating lung cancer mortality
to serve as a surrogate for a control arm in the analyses of the
effectiveness of lung cancer screening in studies missing a control
comparison. A previously fit model was used to simulate lung cancer
mortality in individuals in the absence of screening given their
particular smoking histories, ages, and genders. The model was used
to create a pseudo-control arm that is missing from the lung cancer
screening trial and allowed for a comparison of lung cancer
mortality. The model shows that CT screening does reduce LC
mortality. (Due to confidentiality, no actual data will be shown on
the screening study.)
For more information, contact:
mfoy@rice.edu
13 April 2009
1pm
in 107 Keck Hall
Roberto Bertolusso
Title:
Abstract:
For more information, contact:
Roberto.Bertolusso@rice.edu
6 April 2009
1pm
in 119 HZ
Xiaowei Wu
Title:
Markovian paths to extinction
Abstract: Subcritical and
critical Markov branching processes die out sooner or later. This
talk summarizes some known results about the survival probability,
the limiting distribution conditional on non-extinction, and
demonstrates the asymptotic behavior of the time to extinction and
the path to extinction. These results are verified by simulations.
For more information, contact: xwwu@rice.edu
27 March 2009
1pm
in 1042 DH
Elizabeth Jones
Title:
Single-cell characterization of heterogeneity of Androgen Receptor
Activity
Abstract: Single-cell based studies on nuclear
receptor (NR) mediated gene activation and regulation demonstrate a
range of responses to environmental and physiological stimuli not
previously appreciated by population-based studies. Utilizing
imaging modalities, including deconvolution, confocal, and automated
microscopy we can visualize single cells and measure distinct
cellular responses in an androgen receptor (AR) system. During the
transcriptional activation process, AR changes spatial organization,
a microscopically visible process, including nuclear translocation
and organization into subnuclear speckles. To examine the
ligand-based population shifts and the variation within the AR
subcellular trafficking and synthesis of target gene mRNA we
developed a quantitative biological approach utilizing a
stably-expressing GFP-AR HeLa cell line. We identify and categorize
subpopulations of AR-responsive cells during a time- and
concentration-based series defining measurements that are linked to
responsive vs. non-responsive cells, and may indicate factors, for
example cell cycle stage, that could influence sensitivity to
ligands. Further analysis, both experimentally and mathematically,
of this data will allow us to quantify individual cellular responses
and the variation between these responses from a seemingly
homogenous population.
For more information, contact:
edjones@bcm.tmc.edu
20 March 2009
1:30pm
in 3092 DH
Krzysztof Fujarewicz
Title: Planning identification experiments for cell
signaling pathways
Abstract: Mathematical modeling of
cell signaling pathways becomes very important and challenging
problem in recent years. Fitting of the parameters of the model is
not a trivial problem because of the nature of measurements. The
blotting techniques usually give only semi-quantitative observations
which are very noisy and they are collected only at limited number
of time moments. The accuracy of parameter estimation may be
significantly improved by proper experiment design. The talk will
concern the optimal sampling design - one of the aspects of optimal
experiment design.
For more information, contact:
Krzysztof.Fujarewicz@polsl.pl
Google Calendar of talks:
[Home] [Research Group] [Biomathematical Seminar] [Software] [Publications]