ESF Workshop on advanced large scale gene expression profiling-
focus on miRNAs, ChIP-on-chip, and whole genome sequencing
19-22 May 2008
Turku, Finland

Organisers:

Annika Brandt, Turku Centre for Biotechnology, Finland
Riitta Lahesmaa, Turku Centre for Biotechnology, Finland
Juhani Soini, Turku Centre for Biotechnology, Finland

Draft Report

Summary

The ESFFG-TURKU 2008 workshop and minisymposium was held 19/5-22/5/2008 in Turku at Mauno Koivisto Congress Center, Tykistökatu 6A, FI-20521, Turku, Finland. The workshop was organized by Turku Centre for Biotechnology, University of Turku and Åbo
Akademi University with the title: Advanced large scale gene expression profiling - focus on miRNAs, ChIP-on-chip, and whole genome sequencing. The workshop participants (25) were selected from 60 applicants in total. Of the 25 selected workshop participants 15 were international and 10 were from Finland from total of 11 different ESF countries. The participant countries included: Belgium, Czech Republic, Finland, Germany, Ireland, Italy, Netherlands, Portugal, Spain, Turkey, and United Kingdom. The ESFFG-TURKU 2008 workshop was continuation to our previous ESF course on large scale expression analysis using DNA microarrays held in 2004. The previous course was a practical course with real wet lab experience, whereas this was a lecture course introducing excellent young investigators from world leading laboratories on ChiP-chip (David Hawkins from Bing Ren’s lab, UCSD) and data analysis (Dr. Tero Aittokallio, University of Turku; Eija Korpelainen, Jarno Tuimala and Aleksi Kallio, CSC, Finland). We had in total 28 speakers/instructors and the total number of workshop participants including speakers was 101. The relatively large number of workshop/symposium participants showed that there was an obvious need for further education in this research area. In addition, our event invited a number of speakers from European biocenters: Karolinska Institutet (Sweden), German Cancer Research Center (DKFZ, Heidelberg), Center for Genomic Research (CRG, Barcelona) and the Finnish centers FIMM/VTT (Helsinki/Turku), Viikki together with the Turku Center for Biotechnology.

Scientific Content

The ESFFG TURKU 2008 workshop consisted of three days workshop and one-day minisymposium on the topic “Advanced large scale expression profiling-focus on miRNA, ChIP-on-chip/seq, and whole genome sequencing”.

Workshop Day 1
The first workshop day was dedicated to data-analysis. The first speaker Prof Matti Poutanen (University of Turku) shared his data of a recent study on a clinical sample set and the fundamental strategies to find relevant results in this kind data. The key component in his strategy consisted of a database and the details in constructing it. The microarray expression analysis starts from a carefully planned experimental design which was strongly emphasised since that forms the basis for the subsequent success in data mining. There was short discussion over the time frames of the experiments and there seemed to be general consensus about the bottle necks in the data analysis part of the microarray experiments.

The following speaker who received “the best PhD thesis of the Year” prize in the area of Bioinformatics, Laura Elo-Uhlgrén, talked about dealing with incomplete information in gene expression studies. She showed by example the possible paths of the basic data analysis steps how the data should be considered from different “angles” before advancing further with the data analysis. All data is unique, she explained, we get to know the data in order to mine the data in a proper way. This is essential in the molecular biological research in general and the statistical data analysis has become a central part in the development of future analysis methods. The most critical part constitutes, however, of inadequate sample numbers which are much too often too limited and inaccurate for modeling of the regulatory mechanisms behind the biological phenomenon being studied. The modeling suffers from the lack of knowledge of the most critical components and the mutual relationships between these. Laura has developed in her research mathematical methods for dealing with incompleteness of the information in expression analysis studies. The mining of biologically relevant data is made more efficient by identifying the most critical regulators and through combined data mining efforts though graph theory, statistical modeling, information theory, optimisation and feature identification.

The final lecture of the day was given by Teto Aittokallio (University of Turku), who is the developer of the software “Cytoscape” for data integration. As the name indicates this software can be used for finding association between subjects (or nodes) that is genes or rather their protein products. He demonstrated how the software can be used to find or reveal integration networks. Many people wanted to get their data mined by his software, and the access, restrictions and further use was instructed. He got the most enthusiastic and happy audience to distribute the knowledge gained in his group for the use and application of cytoscape for the new potential end-users! Too short time for Tero, people were just screaming after him with questions!

The final part of the data analysis day was completed by Eija Korpelainen from CSC (Center for Scientic (bio) Computing, Finland), who made a short introduction of the functions of CSC, the programs being developed and the software and computing power available for researchers by CSC. She finally instructed shortly about the basics in the use of their new software for non-mathematicians for microarray analysis called “Chipster”, which was thereafter available for a hands-on experience in a computer class. The participants made their own data analysis with Affymetrix, Agilent or Illumina data made available by Jarno and Aleksi from CSC. The software deals with both expression analysis as well as CGH and SNP data! Participants looked were convinced about the efforts by CSC, the program combines data tracking and visualisation of the steps performed as well as reveals the R-code behind the functions accomplished!

Workshop day 2
The second, and afterwards ranked as the best workshop day by the participants, was
dedicated solely on chromatin immunoprecipitation. The topic was introduced by our star speaker David Hawkins from Bing Ren’s laboratory (UCSD, USA). He took up the most critical aspects of the method aimed for subsequent microarray analysis from the
preparation of the samples how to make the proper controls. He revealed the caveats in
making the controls, what to control and why, which has worked in his research and what aspects needed to be controlled and the difficulties in there: nobody knows exactly the actual needs for controls at this stage. The method is still in the very beginning and the knowledge is gradually gaining with the ongoing research. With an enormous amount of slides David introduced us his protocol step by step towards microarray analysis, the amounts of starting material, the fixing step, the fractionation of the DNA and all the necessary preprocessing steps not to mention the careful choice of a good antibody. Now we are ready for microarray analysis, or are we?

With a short break after David’s talk the latest news and products by the developers:
Affymetrix, Illumina and Agilent. After lunch David continued with “To chip or not to chip?” He showed that by luck they had started with general transcription factors to work with ChIP-chip. Had they chosen a more specific factor, or a bad antibody they could have missed the whole point in the assay. He showed their pioneering results and everything looks so very clear and straightforward. Now the situation is getting even more interesting with the recently developed HT semisolid sequencing methods as we heard from Illumina (and others to come by Thursday: Roche 454 technology). These new sequencing methods are currently being applied by David and Ren lab for ChIP-chip, which becomes even more powerful since no previous knowledge of the TF binding sites are required. There are, however, “gaps” to be filled in the sequence as well in the methods to “read” in the sequencing information from the new sequencing methods. The Illumina’s method creates very short reads but the amount aligned sequences can be up to 10 million. While the pyrosequencing method e.g. by the Roche 454 type of technology can be up to 250 bp, the two methods may thus be complementary. The question will be who can get access to both of these methods or obtain the necessary funding to accomplish these two complementary sequencing methods? Near future will quite soon reveal if the prices follow with the reduction of the running costs as being advertised and convinced by the manufacturers!

The Day 2 afternoon was thereafter taken over by the postdoc from Prof Lea Sistonen,
Eva Henriksson who introduced us their wonderful ChIP-chip success story “Promoter
ChIP-chip analysis in mouse testis reveals Y chromosome occupancy by HSF2”. The data analysis was performed by the bioinformatician Asta Laiho, who works at the CBT
microarray center core facility. She was at first annoyed of the data “because all the hits were strangely focused on just one chromosome: the Y-chromosome” and that “the results is very much in imbalance” and the meaning was instantly clear for Eva who was more than happy with the “Imbalanced results”. Their story has been accepted for publication in PNAS and the article is currently in press (Åkerfelt and Henriksson et al.) for more details.

Workshop day 3
The day 3 in the workshop was the minisymposium. The day started with the keynote speaker Prof Richard I. Morimoto (Northwestern University, Evanston, USA). He continued with the biological aspects of the heat shock factors that were already introduced to us by Prof Lea Sistonen’s group by Eva Henriksson and in fact Prof Sistonen is Prof Morimoto’s former postdoc. By Prof Morimoto we learnt about C. elegans and the disease models for “Integrating metabolic control, proteotoxic stress, disease and aging”. The topic was continued by Prof Lea Sistonen with “Identification of novel target genes reveals multiple functions for heat shock transcription factors”. We were told about the many functions that appear to be controlled by the HSFs, the chaperons of life. David Hawkins covered his own research on “Dynamics of chromatin modifications at cis* regulatory elements in differentiating human embryonic stem cells” where he enlighted where and how he had applied ChIP and chip/seq for basic research in the stem cell field. The director of the Turku Centre for Biotechnology, Prof Riitta Lahesmaa continued with the a related field to stem cells with the title “Gene regulatory networks involved in human T helper cell differentiation”. The combination of data from different types of analysis computed by skilled young investigators (Laura Elo) appeared to be the key to a successful understanding of the data with promising potential for future discoveries.

The afternoon of the symposium was started by VTT(Turku/Helsinki) /FIMM (Olli Kallioniemi) with Matthias Nees presenting their Canceromics program and the results in his talk “Canceromics: genome scale biology & cancer drug discovery”. Again a success story rising from combined analysis: expression analysis and CGH completed with epigenetic and HT screening of compounds. The story was later continued with more details by Matthias’ postdoc Santosh Kumar Gupta by “ Molecular function & pathways: the TMPRSS-ERG fusion oncogene in prostate cancer”.

Three European centers presented their functions, research programs and facilities:
Bernard Korn (DKFZ) by “Methylation analysis, mRNA and miRNA profiling – Lessons
learned in a central facility”; Lauro Sumoy (CRG, Barcelona) with “Cross-Platform
comparison and cross-validation of mRNA and miRNA expression data” and presenting
their work that participated in the Microarray GC Consortium (NIH): Juha Kere was
deputed by Mauro D’Amato (KI, Sweden) who presented the new genomics facility at KI.

The symposium day was finalised by three best young researchers presenting their work: Santosh Gupta (VTT, Turku); Christina Morgenstern (Cancer Research, UK): “Decoding the in vivo Notch targetome” work accomplished by ChIP and TG mice and last but definitely not least an excellent talk by Duygu Selcuklu (University College Cork, Ireland):“Identification of mir-21 (MIRN21) targets in breast cancer cells by microarray gene expression profiling”. On day 3, the circle was closed by the presentation by Prof Kaarina Sivonen (Viikki, University of Helsinki) who presented their pioneering work in sequencing of the genomes of cyanobacteria with the Sanger method and well before the new HT sequencing methods. In fact, the new methods will still face the challenge to complete the sequencing of unknown genomes! Is our future in the hands of the researchers studying cyanobacteria and who attempt to split hydrogen for future fuel and generate oxygen on the side, than those trying to find the cure for cancer?

Workshop Day 4
Workshops for miRNAs, data analysis and data integration, Agilent eArray case
studies and sequencing traditional sequencing. The last workshop day consisted of the talks by Suvi-Katri Leivonen (VTT,Turku) “miRNA expression studies and platform comparisons” with her results very much in favour for the Agilent miRNA platform over excigon LNA-oligos. Vidal Fey (VTT, Turku) presented the secrets revealed by exon array data analysis, Henri Sara (VTT, Turku) completed the picture with data integration possibilities for further data mining by the building of the databases of the results. Erica Leder (University of Turku and Elina Palonen (CBT, Turku) presented their results by using Agilent’s eArrays. Agilent has the web based service for researchers with the sequences of genomes that cannot be found on catalogue arrays or researchers with access to novel genomes. The free probe design algorithm is freely
available for probe synthesis and printing by Agilent (on Agilent multiplexed array-slides). A high graded grid that doubles the feature density will be available soon promised Ruth Burton from Agilent (September). Sini Junttila presented the Lichen sequencing project of Dr Steven Rudd (former bioinformatics group leader at CBT, now at Orion Diagnostica, Espoo, Finland). Lichens are the only food for reindeers and with the climate warming and pollution may become extinct. The sequencing project may lead to exploitation of the lichen genome for manipulation (with the symbiotic algae) that may be applied in the future food producing technologies. The new sequencing methods will open novel avenues also in this research. The last speaker of the day and in the workshop was Dr Clare Pridans who presented her work on ChIP-on-chip at the Cambridge Institute for Medical Research (UK). She had developed a working protocol for ChIP with very low starting amounts of cells. The workshop was closed with a questionnaire with the most positive feedback. Thanks to all the participants!

Assessment of the results & impact of the event

Firstly, the interest that was addressed towards our workshop on “Advanced large
scale expression profiling – focus on miRNAs, ChIP-on-chip/seq. and whole genome
sequencing” showed undeniably that there is a huge need for further education, meeting with others in the same research area, and also to up-date ones current information on the topic, not to mention the need for applying up-front research methods as such but also to develop these by oneself. This concerns particularly bioinformatics, which is the usual bottleneck for obtaining any biologically relevant results! We have shown here that through joining forces by collaboration with the top institutes such as CSC in Finland it is possible to develop new software tools such as e.g. “Chipster” that is an open source software tool for non-mathematicians. It can be further developed by the user to perform desired functions. We think this is one good example of the results of this type of workshop. The researchers will be able to meet the cutting-edge experts in the field, get educated in the secrets of the latest methods, start to change ideas and create new collaborations to develop the methods and ideas further. The novel HT sequencing methods are about to revolutionise our understanding on the genome information once again. In order to find this new information we will need to develop new methods for mining of the novel data. This will need much collaboration, funding, equipment, core facilities and sharp minds – the right people.

Secondly, it became quite clear in our workshop that there is lack of mutual
communication and change of experiences, methods, equipment and ideas between
European bio-centers in general. As already mentioned, the new methods tend to get
more and more complex, we will need better and better core facilities in the future. And
better and better ways or forums for communication. After having this workshop to
organise we feel privileged to see, hear and feel the new development taking place in this research field as was made possible for us by ESFFG.

Thirdly, it became also very clear to us that the present research would not be at this stage without the commercial companies being involved. It would be desirable if the
companies could participate atthe scientific workshops in the development phase of the their products even if this is quite hard to imagine due to patenting issues. Somehow we have to work this out!

Finally, one of the most important goals was to introduce young researchers and their
research to other researchers in the field: David Hawkins, Tero Aittokallio, Laura Elo, Eva Henriksson, Eija Korpelainen, Christina Morgenstern, Duygu Selcuklu and Santosh Gupta! I am 100% sure the entire audience of total 100 conveyors enjoyed their presentations in the workshop!