|
|
|
Christian Schlötterer
Professor
Institute of Animal Breeding and Genetics, University of Veterinary Medicine, Vienna
|
Name: Christian Schlötterer
Age: 44
Title: Professor, Institute of Animal Breeding and Genetics, University of Veterinary Medicine, Vienna, since 2007;
Professor, University of Innsbruck, Austria, 2006-2007;
Associated professor, University of Veterinary Medicine, Vienna, 1999-2006;
Researcher, University of Veterinary Medicine, Vienna, 1995-1999;
Postdoctoral fellow, University of Munich, 1994-1995;
PhD in biology, University of Munich, 1991-1994;
Undergraduate degree, biology, University of Munich, 1985-1991.
As an evolutionary geneticist at the University of Veterinary Medicine in Vienna, Christian Schlötterer has been using microarrays and other molecular biology tools to study molecular evolution in Drosophila melanogaster and other species.
Recently, he and his colleagues tested 454’s sequencing platform for gene-expression profiling and published their results last week online in Genome Research.
According Schlötterer, this is the first published study that uses 454’s technology to sequence only 3' ends, rather than full cDNAs or ditags, which combine 5’ and 3’ ends.
In Sequence spoke with Schlötterer this week about his findings.
Why did you decide to test the 454 technology for expression profiling?
I think the limitations of microarrays for expression profiling are well known. Multi-gene families are a problem, as well as cross-species comparisons, and cross-hybridization. We wanted to test whether you can use 454 sequencing to address gene expression in multi-gene families, for example, and we are still in the process to see whether this actually fulfills our hopes.
Serial analysis of gene expression has the problem that you have to go through a cloning step, and the tag sizes that you get for SAGE are too short to reliably identify the true target, in particular if you have a very poorly annotated genome, for example, if you move away from D. melanogaster, where we have plenty of cDNAs, and we know the untranslated regions.
The problem with SAGE is that most of the tags are located in the UTRs. If you have identified your genes and your coding sequences, that’s fine, but still, since the UTR is not known, you may very easily miss the true target, or the true gene, that corresponds to the tag.
Also, it’s very difficult to do the gene mapping on a reliable basis, because with SAGE, you can’t blast it against the genome, you first have to make an in silico digest of the genome in order to reduce complexity, and then you identify the underlying gene, and if you don’t have this information, then you are a bit limited.
What are the main findings of your paper?
The first finding that we encountered was that 454 sequencing itself has certain limitations with respect to representing the number of tags in a stoichiometric manner. This is dependent on size: What we found is that tags that are too short, say, shorter than 80 bases, are not represented to the same extent as they should be. Somehow they are filtered out. The same holds for longer tags: if they are longer than a certain threshold, then they are no longer fully represented.
One has to develop a technique that allows you to circumvent those limitations of the 454 technique. That’s what we tried to do by breaking the cDNA randomly by nebulization. We then just sequenced the 3’ end of this broken transcript with the 454 technology.
As we tried to break it randomly, there are also fragments in the range that is less well represented, but it will not matter because all transcripts are affected by this to the same extent.
Of course you could try to sequence the entire transcript, but then you run into trouble of having incomplete cDNA synthesis, [and] having to adjust for the length of the transcript in order to measure expression intensity. By just sequencing the 3’ end, you focus on a much smaller region [and] you can directly count every hit on the 3’ end.
For gene-expression profiling, what are the pros and cons of 454 over other high-throughput sequencing platforms, such as those from Illumina and Applied Biosystems?
The advantage of 454 is certainly the read length, compared to the others. If your read length is too short, the unambiguous identification of the tag may become a problem.
In our paper, we made some very preliminary attempts to mimic short read lengths. When you have 20 base pairs, then you cannot unambiguously map 20 percent of the tags. But it’s very poorly investigated, and it’s one of the things that we are in the process of doing, to see whether you get biased gene expression by not being able to map all tags.
The advantage of the 454 technique is that it’s very easy to unambiguously identify the gene to which the tag belongs. The future will tell us whether 454 still remains competitive for expression profiling, depending on Roche’s pricing strategy.
Does the University of Veterinary Medicine own a 454 Genome Sequencer, or are you thinking of acquiring one?
No, this project was a collaboration with a company, Eurofins Medigenomix. We are a very small university, so we will certainly not acquire our own 454 sequencer.
Are you also thinking about trying other sequencing platforms?
We are currently in the phase of exploring other possibilities as well. Illumina and Applied Biosystems are two obvious competitors, and we have to see how things are working out.
There is a trade-off between read length and price. I anticipate that the read length from the competitors of Roche will become better, and I expect that the prices from Roche will drop. It’s very hard to make any predictions. It’s a dynamic field, and I as a customer will wait [to see] what the companies are doing, and I will be going for the best deal.
What other projects are you planning to use your approach for in the future?
We are actually planning to use this approach for projects where we know the limitations, where cross hybridization may be a problem, gene families, as well as for cross-species comparisons. But we have other ideas; it’s a very obvious thing that you can very quickly identify SNPs in transcripts. The major idea is to apply this to other Drosophila species besides melanogaster, where less well-annotated genomes are available.
Do you expect that more gene-expression studies in the future will be done by sequencing?
It’s a matter of pricing, but I would anticipate that in a couple of years, this will be the method of choice for measuring gene expression, and microarrays will be limited to many fewer applications than they are right now.
File Attachments
Christian_Schlotterer_InSeq112707.jpg
