CRUK computational biology day

Add to your list(s) Download to your calendar using vCal

• Speaker to be confirmed
• Monday 18 April 2011, 09:00-17:00
• Cancer Research UK Cambridge Research Institute, Lecture Theatre.

If you have a question about this talk, please contact Florian Markowetz.

With the continual growth of high throughput methods and technologies in biological science, there is an increasing need for computational and statistical analysis. As a result the number of bioinformaticians in CRUK is growing and will continue to grow for some time. This meeting will help to strengthen the ties between computational groups at different CRUK institutes by offering a forum to exchange ideas and skills.

We hope to see lively discussions during the meeting (and thus allocated enough time between sessions) and encourage all participants to present a poster of their research.

Registration Page (click here)

==========

PROGRAM

09:00-09:30 Welcome coffee

09:30-10:30 Session 1

• Andy Lynch [CRI] A note on a meddling mustelid
• Yinyin Yuan [CRI] A sparse regulatory network of copy-number driven gene expression reveals putative breast cancer oncogenes

10:30-11:00 break

11:00-12:00 Session 2

• Rory Stark [CRI]: Differential binding analysis for ChIP-Seq: Occupancy vs. Affinity
• Philip East [LRI]: Analysing a blend of ChIP-Seq flavours to explain a GATA2 dependent cell death phenotype.

12:00-13:00 lunch

13:00-14:00 Session 3

• Richard Mitter [LRI]: Investigating TGF -β signalling. The trials and tribulations of an RNA -seq analysis project.
• Katie Bentley [LRI]: Discovering mechanisms and dynamics in angiogenesis using an integrated simulation/experimentation approach

14:00-14:30 break

14:30-15:30 Session 4

• Gabriela Kalna [Beatson] Matrix factorisation: microarrays and chemotaxis assays
• Roger Phillips [Bradford] Impact of the tumour microenvironment on the delivery of drugs to tumours

15:30-17:00 POSTER session

17:00 END

==========

ABSTRACTS

A note on a meddling mustelid

Andy Lynch [CRI]

As the size and complexity of a genomic study increases, the probability that there will be sample-plating or sample-tracking errors in a study also increases. If you are in the business of analysing large and complex studies, then this can then be a problem. I present an overview of the ‘BADGER’ method that we have developed to try and identify such errors in a particularly complex study, and discuss the implications of such an approach for other aspects of study design and analysis.

A sparse regulatory network of copy-number driven gene expression reveals putative breast cancer oncogenes

Yinyin Yuan [CRI]

Copy number aberrations are recognized to be important in cancer as they may localize to regions harboring oncogenes or tumour suppressors. Such genomic alterations mediate phenotypic changes through their impact on expression. Both cis- and trans- acting alterations are important since they may help to elucidate putative cancer genes. However, amidst numerous passenger genes, trans-effects are less well studied due to the computational difficulty in detecting weak and sparse signals in the data, and yet may influence multiple genes on a global scale. We propose an integrative approach to learn a sparse interaction network of DNA copy-number regions with their downstream transcriptional targets in breast cancer.

Differential binding analysis for ChIP-Seq: Occupancy vs. Affinity

Rory Stark [CRI]

Next-gen sequencing has enabled protein/DNA binding to be mapped at a genomic level via ChIP-Seq. While much of the effort in computational analysis in this area has been devoted to identifying enriched regions in individual samples (peak finding), derivation of meaningful biological results requires differential analysis of many samples representing different replicates, conditions, and/or transcription factors. Such analyses typically focus on DNA sites where the presence or absence of a protein binding event changes. While these occupancy-based analyses can identify binding sites unique to a group of samples (albeit lacking reliable confidence statistics), they are less able to differentiate between sites with a positive occupancy status in both conditions, but where the affinity (proportion of cells in the sample that are bound at that site) exhibits significant changes. I present techniques (encapsulated in an R package) for identifying such binding sites using methods developed for RNA -Seq and demonstrate their efficacy in isolating sites differentially bound between groups of ChIP-Seq samples, showing how sample groups can be cleanly separated using only binding sites that are occupied in both groups but whose affinities differ significantly, as measured by FDR after multiple testing correction.

Analysing a blend of ChIP-Seq flavours to explain a GATA2 dependent cell death phenotype

Philip East [LRI]

Within Bioinformatics and Biostatistics at the LRI ChIP-Seq remains our largest NGS application. Here we highlight our efforts in this area focusing on an analysis containing data from a number of ChIP experiments. In lung cancer cell lines with KRAS or EGFR mutations GATA2 knockdown is fatal. To try and identify transcriptional control mechanisms that explain this phenotype GATA2 , RNA PolII and 3 flavours of histone methylation were chipped. We have integrated these data in an attempt to identify transcriptional control effects correlated with this phenotype.

Investigating TGF -β signalling. The trials and tribulations of an RNA -seq analysis project.

Richard Mitter [LRI]

The TGF -β signaling pathway is involved in many cellular processes in both the adult organism and the developing embryo including cell growth, cell differentiation, apoptosis and cellular homeostasis. The pathway is classically considered to be formed of two halves that operate through different ligand-receptor partnerships to activate TGF -ß target genes. The workhorses of both halves of the pathway are the Smad transcription factors that form complexes to transduce the signal from the membrane into the nucleus. RNA -seq was used to investigate the effects of TGF -ß on gene expression in the presence and absence of certain Smad factors in the hope of identifying a new branch to the TGF -ß pathway.

Discovering mechanisms and dynamics in angiogenesis using an integrated simulation/experimentation approach

Katie Bentley [LRI]

Angiogenesis (the growth of new blood vessels) is a complex, dynamic process. Multiple processes such as cell migration, cell fate selection and cell rearrangement occur concurrently to produce a regularly patterned branching blood vessel structure. To pull apart and understand the underlying mechanisms driving these interacting processes we have developed a 3D agent-based computational model; used in iteration back and forth with experimentation this has led to a continual refinement of the model, deeper understanding of important components at work in cell fate selection and indicates an entirely novel behaviour may have been discovered in pathological conditions such as cancer.

Matrix factorisation: microarrays and chemotaxis assays

Gabriela Kalna [Beatson]

The spectral biclustering (SbC) based on the singular value decomposition is a widely used matrix factorization technique for large data sets from high throughput experiments. It provides powerful unsupervised approach to identification of the main patterns as well as experimental artefacts within the data. Moreover, SbC vectors simultaneously rearrange rows and columns of the data set into distinct checkerboard patterns. In microarray analysis, coupled with pathway/gene ontology and survival/clinical data, this can identify functionally related prognostic gene sets for head and neck squamous cell carcinoma. In image analysis of chemotaxis it reveals aspects of complex cell boundary movements.

Impact of the tumour microenvironment on the delivery of drugs to tumours

Roger Phillips [Bradford]

The concentration of drug delivered to the tumour and the duration of time the drug remains in residence at the tumour site are two key factors that will determine tumour response to chemotherapy. Within tumours however, a microenvironment is created by a chaotic vascular supply and altered tumour stroma that significantly hinders the delivery of therapeutically effective drug concentrations. During this presentation, the impact of the tumour microenvironment on drug delivery will be described in the context of providing experimental pharmacological data for the development of in silico models.

This talk is part of the Seminars on Quantitative Biology @ CRUK Cambridge Institute series.

This talk is included in these lists:

• All CMS Events
• All Talks (aka the CURE list)
• Biology
• CCC talks for website
• CRUK CI Seminars
• CamBridgeSens
• Cambridge Centre for Data-Driven Discovery (C2D3)
• Cambridge talks
• Cancer Research UK Cambridge Research Institute, Lecture Theatre
• Chris Davis' list
• Computational and Systems Biology
• Graduate-Seminars
• Interested Talks
• Liam
• Life Science Interface Seminars
• Life Sciences
• Life Sciences
• ME Seminar
• PMRFPS's
• School of Physical Sciences
• Seminars on Quantitative Biology @ CRUK Cambridge Institute
• Trust & Technology Initiative - interesting events
• bld31
• custom
• my_list
• ndk22's list
• ob366-ai4er
• other talks
• rp587
• se393's list
• sfm36

Note that ex-directory lists are not shown.