Title - Prediction of metastasis site in primary breast cancer
By Kristian Wennmalm (PhD 2007) [lab link] [PhD thesis] [paper 3] [paper 4]
ABSTRACT
Background
Breast cancer is clinically known to develop different metastatic patterns. Recent preclinical model-system results indicate that the site of distant metastasis can be predicted. Methods In two population based cohorts with 402 primary breast tumors we defined skeletal, lung and liver metastasis signatures on the basis of Affymetrix expression microarrays, and validated proposed cell-line signatures. A multiple random validation approach was applied to avoid overestimation of the predictive capacity.
Results We found that proposed cell-line signatures have little ability to predict lung and skeletal metastasis in our clinical cohort. In contrast, signatures defined in our primary breast cancers could robustly predict lung and liver metastasis. Prediction was further characterized by poor sensitivity, numerous false positives, and a strong connection to biology underpinning histopathological grade and HER-2/neu status.
Conclusions The utility of metastasis-site gene expression signatures, in primary breast cancers, should be interpreted with considerable caution. We find that they can predict tumors at increased risk of metastasis, whereas prediction is confounded by grade and HER-2/neu related gene expression differences.
----------- Some words on "Seminars in Systems Medicine"---------------------------------------
Place and Time: The Tegnér-Björkegren laboratory for Computational Medicine hosts these Seminars in Systems Medicine (SSM) held at GV - King Gustaf V Research Institute (GV) (KS hospital Map pdf) (M1 behind the main hospital building; hitta.se link - at the blue spot) seminar room, first floor, every fourth thursday, between 3-4 PM. During this semester we will host seminars the following dates; 31/1, 6/3, 27/3, 24/4, 22/5, 19/6.
Some demarcations: By systems medicine we include research which targets complex multifactorial diseases. This is to be viewed as the medical part of systems biology whereby we exclude work on E-Coli, Yeast, and similar systems of less relevance to the clinic. Topics and work suitable for SSM is most likely integrative across discplines but regardless if it is mainly experimental, clinical or computational, the research should target either central molecular/cellular disease processes, diagnostics or methodology (measurement techniques, computational analysis, animal/cell models) relevant for addressing multifactorial and complex diseases such as MS, cancer, diabetes, neurological conditions, inflammatory processes, cardiovascular, ... The systems part in the seminar series refers to the fact that for complex diseases there are as rule several non-homogenous components (SNPs, mRNA, proteins, metabolites, cells, organs) which interact across several spatial and temporal scales. This is the major reason why single-target approaches are less likely to succeed in bridging the genotype-phenotype scale which also is convoluted by different life-style factors.
Our aim is to support and create a community which includes relevant staff personnel and researchers ranging from PhD students to Professors who all have a shared mission of employing integrative systems approaches for understanding complex diseases. By understanding we mean finding groups of predictive markers (diagnostics) as well as mechanistic identification of disease processes.
Johan, Josefin and Jesper, on behalf of the Computational Medicine Team.
----------------------------------------------------------------------------------------------------------------------------
From cataloging glomerular genes towards systematic understanding of kidney glomeruli
By Liqun He (PhD 2007) [lab link] [paper link]
ABSTRACT
Kidney glomeruli play a crucial role in the maintenance of body homeostasis. Many diseases attack the kidney function by primarily affecting glomeruli. However, the transcriptome profiles and the function of the glomerulus are poorly understood.
In order to explore the transcriptome of kidney glomeruli, several genomics and bioinformatics approaches were applied. First, we constructed and large scale sequenced four Express Sequence Tag (EST) libraries generated from pure preparations of newborn and adult mouse glomeruli. By comparing the transcript abundance profiles in the glomerulus EST libraries with public whole kidney libraries, we identified 497 glomerulus-enriched mouse transcripts in the newborn and/or adult mouse glomerulus. The glomerular ESTs were printed on glass slides in order to generate cDNA microarrays with broad representation of glomerulus-expressed genes (GlomChip). Subsequently, by using GlomChip to compare the RNA samples from the glomerulus with non-glomerulus kidney tissues, we identified 357 mouse genes as glomerulus-upregulated. Further, by combing the result from Affymetrix whole genome array study and published SAGE and Stanford cDNA array results, we did a meta analysis and merged the data into a catalogue of 1407 glomerulus-enriched genes. Based on this, a protein-protein interaction network in the glomerulus (GlomNet) was constructed.
The transcript catalogue that we have generated provides information about the transcriptome profiles of the kidney glomerulus, and contributes new information about their function, physiology and disease. Also, GlomNet will contribute an integrated systematic understanding of the kidney glomerulus.