ADAM: Array-based Discovery of Adaptive Mutations

Source: Goodarzi H, Hottes AK, Tavazoie, 2009. Global discovery of adaptive mutations. Nature Methods, 581 - 583.

I had to wait for a long time to actually write about a project of my own. I had been thinking about this problem for at least 2 years before finding the solution. The problem I wanted to address was a classic question: how can you map a mutation? If you do a selection and find a mutant which exerts your phenotype of interest, how would you go around finding it? The classic approach involves painstakingly going through a set of known markers around the genome and calculate the linkage between the known marker and your site of mutation. Then choose the closest marker and redo the linkage analysis with more markers close to that site, until the location becomes small enough for direct PCR amplification and sequencing. But this whole protocol, takes weeks if not months and it is quite labor-intensive. Especially, if your mutant carries multiple effective mutations, finding them becomes more complicated.

With the advent of whole-genome sequencing or re-sequencing of bacterial genomes can help us find all the mutations in the genome. But in many cases, it is not obvious whether a mutation is responsible in eliciting a certain phenotype or not. In other words, every single mutation should then be tested for its effect on the phenotype of interest.

To tackle this problem, we developed ADAM as a whole-genome approach that enables you to find all the adaptive mutations (and not silent mutations) in a matter of days. ADAM, uses parallel, genome-wide linkage analysis to simultaneously identify all mutated loci with direct fitness contributions. ADAM has three components:
1. A library of selectable markers embedded in the DNA of the parental strain.
2. A mechanism for transferring markers from the parental strain's library into the evolved strain in such a way that DNA from the parental strain adjacent to the marker replaces the corresponding DNA in the evolved strain.
3. A method for measuring the frequency of markers throughout the genome.

As the first component, we used a high coverage library of kanamycin-marked transposon insertions all across the wild-type genome. We then trasnfer the markers from the WT genome to the evolved background using P1 transduction. In this step, if a transposon is close to a site of mutation, the recombination results in the correction of this mutation back to the WT genotype. Now, in this secondary transposon library in the mutant background, the markers whose introduction to the genome has been accompanied by the loss of an adaptive mutation will be at a relative disadvantage under the selective conditions. To find the location of these transposons, we compare the frequency of transposon insertion events in each locus under selective and non-selective conditions. Adaptive mutations can then be discovered in locations in which a stretch of loci show marker depletion upon selection. For this step, we use a microarray-based whole genome footprinting.


We used ADAM to find a known CmlR cassette as a proof-of-principle. We then used this method to identify adaptive mutations in lab-evolved strains (growth on Asn and ethanol tolerance).

We're all Iranians

Source: Nature 460, 11-12 (2 July 2009) | doi:10.1038/460011a; Published online 1 July 2009

Well, I'm an Iranian and the recent unfoldings in Iran has engaged me to the extent that I don't have time to write about science anymore. But a few days ago, I stumbled upon this editorial from Nature, which truly impressed me both as a scientist and as an Iranian.

I've never been a fan of politics. Politicians hold a reactionary mindset, aiming at simplifying the matters down to meaningless words; however, a scientist is bent towards evidence and proof. There is no question that any public movement in any country should be both acknowledged and encouraged. But, what can we, in the scientific community accomplish?
1. Universities are very well equipped to speak loud and in clear... making clear that Iranians are supported by non-governmental entities across the world.
2. The recent events accompanied by massive crackdowns on academics followed by mass resignations and discontent, would urge many young Iranian academics to leave Iran. The colleges, universities and research institutions all across the world can accomplish a great deal by prioritizing these researchers for a while or at least create a more active network to find suitable jobs for them (something similar to Scholars at Risk Network).

PNAG-driven mode of biofilm formation in E. coli

Source: Amini S, Goodarzi H, Tavazoie S. (2009). Genetic dissection of an exogenously induced biofilm in laboratory and clinical isolates of E. coli. PLoS Pathog. 2009 May;5(5):e1000432.

This is the first experimental project the progress of which I've observed from conception to completion. I witnessed the myriad of challenges that Sasan had to deal with to make the case for his hypotheses. And amazingly enough, he pulled it through. In experimental biology, it's difficult to claim victory at any point. Every given experiment prompts more experiments that are both labor-intensive and time-consuming. In computational biology, it is the model, the experiment or the approach that matters... it has to be novel. In experimental biology, on the other hand, what you need to do is clear (most of the time), it's just that matter of doing it. E.g. for testing the function of a gene you need to knock it out. Every one knows that. But dong it is the difficult part. This study was one of those cases were many corroborative experiments needed to be done, from microarray-based fitness profiling of mutant libraries to peptidoglycan extraction and visualization (polyacrylamide gels) to PNAG purification and validation (mass spectrometry). And remember that one lab does not actively use all these methodologies and sometimes for doing one experiment you have to set-up the whole experiment: ordering the reagents, making the solutions, optimizing conditions and ...

At this point, I actually feel ashamed. The fact that I reduce thousands of hours of work into a couple of sentences, passing judgement from the safety of my laptop bothers me (just a little though). I'm not going to do this for this particular study. If you're interested, read the whole paper (it is open access). I would just tell you that this is an example of a complete story, from start to finish and I congratulate Sasan for accomplishing this painstaking task.

Ribosome Profiling

Source: Ingolia et al (2009). Genome-Wide Analysis in Vivo of Translation with Nucleotide
Resolution Using Ribosome Profiling. Science 324: 218-223.

Proteomics is the game, but due to technical difficulties in the direct quantification of protein types in the cell, we have instead been using transcriptome measurements as a proxy for protein expression. This is a decent proxy but far from perfect. In this paper, the authors bring us a step closer to protein quantification through measuring the portion of transcriptome that is actually being translated. They combine the ribosome-mediated protection of traslated RNA molecules with the power of deep sequencing to determine the ribosome positionings at a single nt resolution.

What the authors found?
1. There is an excess of ribosomes bound to the first 30-40 codons; a quantity which drops substantially in the later codons.
2. uORFs are quite widespread, resulting in a high ribosome presence in the 5'UTR.

Revealing Genetic Interactions in E. coli

Source: Typas et al, (2008). High-throughput, quantitative analyses of genetic interactions in E. coli. Nature Methods 5, 781 - 787.

Genetic interaction studies in bacteria are quite challenging. Before this study, we had data for hundreds of interactions in E. coli in comparison with thousands known for yeast. In this paper, the authors introduce a method termed GIANT which relies on massive Hfr conversion and double mutant generation in E. coli. The method is presented below as a figure from the original paper:
Here, we rely on conjugation to select for double markers (double mutants) and assay they growth on 384 or 1536 colony arrays. The authors make the case for their method through several validation steps. In the end, this method has the ability to generalize to other organisms for which deletion collections are available.

Proliferation-resistant biotechnology

Source: Nouri A., Chyba C., 2009. Proliferation-resistant biotechnology: an approach to improve biological security. Nature Biotechnology 27, 234 - 236.

A dear friend of mine Ali Nouri (also a AAAS congressional fellow) has published an interesting commentary in the current issue of Nature Biotech. Generally, we (the scientists) sometimes fail to grasp the security implications of science. Now, while the uninhibited progression of science is essential for our prosperity and avoiding regression back to another dark age, we should also try to find inexpensive and applivable ways to boost security.

Making an entire organism from scratch is not a dream anymore. This has been done in case of many viruses. The resurrection of the 1918 influenza virus caused a turmoil in our field. While we learned alot about the virus (e.g. how close it actually is to avian flu), many questioned whether this type of research should be prohibited. Personally, I don't think any type of basic research should be prohibited because any thing may simply revolutionize our lives, but I agree that this information should be protected against misuse and abuse.
In this commentary, the authors have simply requetsed the companies to screen their bulk requests and raise a red flag if the requested gene or genome belongs to a list of dangerous oragnsisms or toxins. Steps as simple as this are very cheap to implement. And I'm sure many of you are already coming up with solutions for potential bypass of this problem. But if we put enough obstacles in the way of misusing these technologies, the accumulative security would actually synergistically increase and may very well pass the threshold for many ill-willed individuals.

Leading into my work...

Source: Lisec, J., Meyer, R.C., Steinfath, M., Redestig, H., Becher, M., Witucka-Wall, H., Fiehn, O., Torjek, O., Selbig, J., Altmann, T., and Willmitzer, L. Identification of metabolic and biomass QTL in Arabidopsis thaliana in a parallel analysis of RIL and IL populations. 2008. The Plant Journal, 53: 960-72

The authors created a number of RIL and IL lines in Arabidopsis and then ran targeted GC-MS on them. They were able to measure 181 compounds and find QTLs for 84, for a total of 157 QTLs. The contribution of these loci was between 1.7 and 52.1%. They found that many of these metabolites co-mapped, and that in nearly all of them a good candidate gene could be found that might explain the effect. They defined a candidate gene as a gene within the support interval in the direct pathway of the metabolite. None of these metabolite linkages showed a strong ability to change biomass.



Other notes:
-permutation test for candidate gene: randomly assign linkage to metabolite, sort through interval and see if any genes overlap with metabolite in AraCyc
=most metabolites showed no significance
=only 13 metabolites showed a higher than permutation-average number of candidate genes
-near impossible to find epistasis, found it only explained 2.72% of phenotypic variation on average
-nonrandom distribution of mQTLs, does not correlate with distribution of metabolic genes