My rotation project II is almost done now. Next week is finishing and write up (paper-style report which gets evaluated). It was another great rotation project. I really enjoy my time at the Sanger Institute :)

I don’t want to talk too much of this project, as there will be a big paper at some point, also incorporating all the work I have done. In short I worked on graph clustering and annotation validation, using relative entropy, Fisher’s test (which uses the Gamma function as an approximation) and a “home made” heuristic score. All this is used for mining clusters for the biological meaning. More about that later …

Besides the “rotation project Nr. 2″ all PhD students attended the graduate lecture series. This is a Sanger intern program about various relevant research topics. I have to say that I have learned more biology in the last 5 month than ever before in my life. This is good, as I am/was very computer science oriented and the lack of knowledge in (molecular biology) has to be filled. Sometimes it is quite hard to understand the bio-papers, but the more you know the easier it is to pick up new stuff and add it to the big picture. Quite nice to see that :) Also the Journal Club helps extending the own limited view of the research world and thanks to friends at Sanger or Wikipedia it is almost always possible to prepare for a paper (starting from scratch, following some medium level stuff and then understand the message of the paper – details are often too much :)

Anyway, the last month a lot more things happened. Probably I will tell you later. For today it is enough. Well, I let you know that we are currently moving house, which is EXCITING :)

The last 2 weeks were quite busy. Our first year PhD rotation projects here at the Sanger Institute are only 2 months each, which is not much time for a small project, including a new working field, a new place, new people and a lot of new experiences …

And then there is Murphy’s law, which seems to be a natural phenomenon. It is always like that, but in the end one can mostly finish on time :) It was extremely interesting and a wonderful project to combine web-lab with computer science. I have worked through the whole pipeline of an protein profiling experiment – ranging from sample preparation including “wet fingers” to the more advanced data analysis which is obviously more “theoretical”. There is a great chance, that I’ll come back to mass spectrometry for my PhD project. It is an extremely interesting and diverse field!

The project was very dynamical and I followed some side tracks, especially for the data analysis, which resulted in new ideas and approaches. Here’s a small abstract of my work:

Abstract: Since many genomes are fully sequenced, research focus is shifting towards genome function.
This relies on genome annotation which remains one of the major challenges. Mass
spectrometry can provide experimental evidence of peptides which can be mapped back onto
the genome for annotation purposes. We have established a mass spectrometry (MS) data
analysis pipeline for peptide identification and genome annotation, which can be used to verify
and elucidate novel gene products.

Now I have to catch up some sleep over the weekend :)

OK. My short programming excursion is over and I went back to do my wet-lab experiments. It is tough work (see text) and very different to computer science :) Mainly there are three main differences which I have in my mind:

1. You prepare your sample, and you need a lot of patience to process it, e.g. the slicing of my gel took me about 2 days, but destaining, digestion & drying will take even longer, probably 3 or 4 days for my 4 plates (2 parallel). Simple tasks in general, but you have to be very precise and one is not allowed to shiver or to be unfocused! On the other hand, my preparations and experiments are very short, compared to some other experiments which last for months … The incubation times are also not to neglect, so you have 2-3 hours of “free” time for paper reading; I haven’t written a program yet, which compiles for 3 hours though – smile!

2. What about if the experiment doesn’t work? This is actually a every day experience in the wet-lab. But you can’t start debugging your sample, well at least not in a way you do it with source code :) Not working programms or algorithms are still a pain, but after days hunting the bug, you’ll hopefully find it! Some experimentalists told me, they had experiments which weren’t working even after a year …

3. It is different in another way too: you prepared your sample, and after weeks you ruin it by doing the wrong thing or you accidentally hit something/someone with your samples in your hand – no way to think of it. The whole work probably would be for nothing. Therefore, all sorts of in-silico work is more “save”. You can’t screw up your sample by accident (unless you perform a $> rm -rf /* and/or don’t make backups). Imagine, after years of programming your fancy application, you wrote a class which doesn’t work (of course it has never happened before – smile) or you misconfigured your config files – but your previous work is not gone, you can simply fix it. Nice, isn’t it!

I feel more “save” in front of a computer, but I love my experience I currently gain and I also need it to understand the problems, once I have to analyse the data. All tasks in the lab seem to be easy – in theory – but practice is different. Good to know and good that even Bioinformaticians have to do at least one wet-lab rotation here during our PhD time!

Best regards!

A different day: programming!

Hagen, a Masters student who just started a PhD in Spain is visiting our group for a week. He carried out his quite interesting final year project here at Team 17 in cooperation with Richard Durbins group. Now some data has to be verified and recalculated and I have the chance to get into it and continue what was started …

For me it was a good introduction into his work today, basically very important, as it could be highly relevant for a possible PhD project. Furthermore I wasn’t in the lab, rather we were programming. I just felt much more secure and confident, compared to wet-lab work – but no wonder, that is what I learned at University ;-)

We started implementing some scripts in Perl, but as this thing turned out as a memory hungry dinosaur (we used 3GB ram in the end) we ported it to Java, which was better for me anyway, as I am a very “fluent” Java speaker + it was 10x cleaner and OO. It was fun, sitting in front of computer again …

You wonder what I did the whole last week ? OK, I try to explain a bit …

First of all, all these technologies are quite new for me. Well, I know the basics of Proteomics, but I rarely touched a wet-lab. Anyway – I learned how to prepare a cell culture to extract the proteins, I learned running a protein gel (1D SDS-PAGE), how to rinse and stain the gel, how to cut it and finally how to destain it, digest & dry the proteins. First I looked over the shoulder of Mercedes (Thanks for all your patience!), then I started my own experiment, which will take about 2 months including data analysis. Sure, I’ll have to ask quite often for help :)

I will examine proteins from nuclear and cytoplasmic samples (don’t want to tell you more here!) by LC/MS/MS, to see which proteins are abundant and which aren’t. This result can be incorporated into a bigger running experiment and will help to judge their results. But that’s not all. I will interpret the spectra from our machines and also compare the data acquired. We have the most recent machines in our lab, a LTQ-FT and a Q-TOF, and it will be interesting how they compete on the same samples (currently I prepare everything for more equal runs on these machines). Time per run, resolution, mass accuracy and sensitivity are important parameters. Anyway, I don’t want to go into details… Another issue I will cover with my experiments is a comparison between a RP HPLC and a more sophisticated technology, which is supposed to have the same resolution, but with a higher throughput. We’ll see …

Besides practical work, I read over a dozen papers, and there are still soooo many which are important to get the big picture! Within the next couple of weeks, I’ll have to go through the most important literature about my research, otherwise I won’t be able to write my first report in 2 month.

For me, the whole thing is a great opportunity. I learn the basic wet-lab techniques, get into operating the machines and also learn about the spectra & data. Considering a PhD in this group it is extremely helpful, as this would mean to implement computer aided analysis software & algorithms besides research about new approaches to mine the data. But that’s future, possibly starting in about 6-7 month ;-)

Tomorrow there is an UNIX course, which is compulsory – will be fun for me I guess! At least I hope I can support the biologists tomorrow …

Currently I am quite nervous, because tomorrow is a big day – the “real” start of my PhD ;-)

I am going to work in Team17, as I am very interested in the automated interpretation of MS data, which is a challenging undertaking. To get a feeling for this subject, I’ll do my wet-lab in MS. The project description:

Identification & characterisation of a multiprotein complex by proteomic mass spectrometry

Proteins often operate in concert with one another, organising into multisubunit complexes. The interactions between proteins modulate to a large extent several aspects of their function, including subcellular localisation, activity, specificity and regulation. Understanding the relationships between complex subunits and among complexes themselves therefore provides vital clues of molecular function and protein organisation. With recent improvements in technology and availability of genome sequences, mass spectrometry is ideally suited for the identification and characterision of proteins from complex biological samples.

The central theme of our research is to develop and apply procedures to describe native in vivo protein interactions, based on systematic targeted isolation of multiprotein complexes, followed by identification of the associated components by mass spectrometry. Our efforts are directed towards regulators of gene expression that are important in stem cell differentiation and early embryonic development.

The objective of the rotation project will be to investigate and characterise the multiprotein assembly(ies) associated to one of our candidate gene expression regulator proteins. A cell line has been constructed where the target protein is modified with a sequence tag that allows its isolation by two-step affinity purification. The tagged gene is under the control of its own promoter and therefore tagged protein levels are kept very close to physiological. The project will involve the purification of the bait protein, together with its associated proteins, from tagged cells. The isolated proteins will be processed and analysed by liquid chromatography / tandem mass spectrometry (LC/MS/MS). In instances where there is indication of post-translational modifications in the bait or interacting proteins, further mass spectrometry analysis will be carried out to confirm such features. The project will use the most modern MS technology and an array of pioneering biochemical techniques, which might include relative quantification by SILAC or ICAT.

Tomorrow I will now more about what I will learn, what I’ll have to do and which books/papers are waiting for me. Good night for now!