October 10, 2013

Separating the wheat from the chaff is a lot easier than figuring out which variants matter in a GWAS.
Cheap and easy genome sequencing has been both a blessing and a curse. We are able to find an incredible wealth of variation, but for the most part we have no easy way to tell whether a difference might contribute to a disease or not.
The poster child for this problem is autism. Lots of genome wide association studies (GWAS) have been done and lots of rare variants in lots of different genes have been found – unfortunately, way too many to pick out the ones that really matter.
Luckily our friend yeast can help. Various researchers have identified a number of variants in the human cation/proton antiporter gene NHE9 that associate with autism. In a new study, Kondapalli and coworkers used the NHE9 ortholog NHX1 from S. cerevisiae as an initial screen to identify which variants impact the activity of the NHE9 protein. They found that two of the three mutations they looked at compromised the activity of yeast Nhx1p.
They then set out to confirm these results in mammalian cells. When they looked at protein activity in glial cells, they found that all three mutations compromised the activity of NHE9. This is obviously different from what they found in yeast.
Now this doesn’t mean that yeast is useless for this approach (God forbid!). No, instead it means that it is probably only useful for a subset of autism mutations. Kondapalli and coworkers had suspected this, but apparently the subset is smaller than they initially thought.
The first thing they did was to generate a rough three dimensional map of the NHE9 protein in order to see which parts the two proteins shared. The idea is that they could then do a quick screen in yeast with mutations that affect the shared structure.
While the structure of NHE9 has not been solved, we do have the structure of its distant bacterial relative, NhaA. Kondapalli and coworkers aligned the two along with the yeast ortholog Nhx1p and identified conserved regions.
Three of the NHE9 mutations associated with autism—V176I, L236S, and S438P—were all predicted to be in shared, membrane-spanning parts of the protein. The researchers introduced the equivalent mutations into NHX1—V167I, I222S, and A438P.
A yeast deleted for NHX1 grows poorly in high salt and low pH and also has increased sensitivity to hygromycin B, as compared to a yeast with a functioning NHX1. Two of the mutant genes, carrying A438P or I222S, failed to rescue these growth defects. The other mutant gene, with the V167I change, worked as well as wild type NHX1 at rescuing the yeast. So at least in yeast, two of the three mutations appear to impact protein activity.
The next step was to see if the same was true in mammals. Easier said than done! Ideally they would want to investigate whether these mutations affected the protein in the cells where NHE9 is usually active. Too bad no one knows this protein’s natural habitat. This is why the researchers starting slicing mouse brains to figure out when and where the protein is expressed.
While we don’t have time or space to go into all the details here, Kondapalli and coworkers found that when and where in the brain NHE9 was expressed made sense as far as a possible contribution to autism. They also found that glial cells had about 1.2 fold more NHE9 transcripts than did neuronal cells. They therefore did their assays of protein activity in a type of glial cells called astrocytes.
While they couldn’t completely knock out NHE9 in mouse astrocytes, they were able to knock down its expression by over 80%. When they added back the mutant NHE9 genes, they found that all three failed to mimic the effect of adding back wild type NHE9 to these cells. This is different than what they found in yeast, where only two of the mutations impacted protein activity.
When they went back to their 3D model, they saw that the mutation that differed, V167I, affected a less defined part of the structure. This points to the fact that for the quick yeast screen to work, they need to be looking at parts of the protein where the structure is shared between the yeast and the human version. In a perfect world they would have had crystal structures of each to work off of instead of having to kludge together a model.
In any event, this is the first step towards validating yeast as a quick screen for identifying mutations that can impact protein activity and so are good candidates for being involved in disease. Yeast may help scientists separate the wheat from the chaff of GWAS and so help figure out how diseases happen and maybe help find treatments or even cures. Well done yeast.
by D. Barry Starr, Ph.D., Director of Outreach Activities, Stanford Genetics
Categories: Research Spotlight, Yeast and Human Disease
Tags: autism, Saccharomyces cerevisiae, yeast model for human disease
October 07, 2013
Congratulations to Randy Schekman, James Rothman, and Thomas Südhof, who have been awarded the 2013 Nobel Prize in Physiology or Medicine for their work in understanding how the cell organizes its transport system. Randy Schekman used the awesome power of yeast to identify and characterize genes required for vesicle traffic. James Rothman characterized these and other proteins in mammalian cells, and Thomas Südhof showed the critical role of vesicle trafficking in nerve cells. You can read summaries of their Nobel winning work at Nature, The Scientist, and The New York Times, or search SGD to see how each of these researchers has used our model organism in their research: Randy Schekman, James Rothman, Thomas Südhof.
Categories: News and Views
October 02, 2013

At least yeast only had to deal with a lack of tryptophan!
Remember in Dune when Paul Muad’Dib took a sip of the “Water of Life” and needed weeks in a coma to turn it into something that let him survive and emerge even more powerful than before? Turns out yeast sometimes have to do something similar.
Now of course the yeast aren’t consciously moving molecules around to deal with a poison like Paul did. No, instead they sometimes need to transcribe low levels of a mutated gene over a long period of time to survive in a new environment.
This process is called retromutagenesis. The idea is that a cell gets a mutation that would allow it to survive and prosper in a new environment if only it could replicate its DNA. Unfortunately the new environment is so unforgiving that the cell can’t replicate.
The cell escapes this catch-22 by transcribing the gene with the mutation so that the mutant protein can get made. Once enough of this protein is made, the cell manages to get up enough steam to power through a cell cycle. Now the mutation is established and the yeast can make lots of mutant protein and happily chug along.
In a new study in the latest issue of GENETICS, Shockley and coworkers hypothesize that something like this is happening in their experiments. They were studying oxidative damage to DNA and found that some of their mutants required many days before they could grow in the absence of tryptophan (trp). They argue that these late arising revertants were due to the cells having to wait until retromutagenesis allowed enough functional Trp5p to be made so the cell could replicate.
The authors have created strains of yeast with various mutations in the TRP5 gene that cause the yeast to be unable to grow in the absence of trp. What makes these strains so useful is that they are set up in such a way that six different, specific point reversions can result in a functional TRP5 gene. They can then analyze any Trp+ revertants to see what types of damage lead to which type of mutations.
One of the first things the authors discovered was that oxidative damage caused all six different reversions. While this was interesting, the specific mutation they wanted to focus on was a G to T transversion which occurs when G is converted to 8-oxoguanine. This is why they focused on the trp5-A149C strain.
The main way that yeast cells deal with 8-oxoguanine is by removing it with the Ogg1 protein, a DNA glycosylase. When Shockley and coworkers deleted this gene in their strain, the number of revertants increased by 20-fold. From this they concluded that most of the revertants were the result of the misreplication of an 8-oxoguanine.
This is where the yeast run into a problem. In the absence of trp, the trp5 mutants do not replicate at all…they do not go through even one cell cycle. But to revert to a functional TRP5 gene, this strain needs to go through a cell cycle. This is why the authors think that the first step towards reversion is a mutation in the TRP5 transcript.
Consistent with this idea is the fact that the mutated G in this strain is on the transcribed strand and that this is important for high revertant frequencies. It also helps to explain why revertants took so long to appear. Basically there had to be a buildup of enough functional Trp5p to allow a single cell cycle to happen. Then the G could be converted to a T and the yeast could happily grow. In this specialized case, it looks like reversion is dependent on retromutagenesis.
But retromutagenesis, also called transcriptional mutagenesis, doesn’t happen only in yeast cells. It’s being studied as a possible way that all kinds of quiescent cells, such those in the process of becoming tumor cells, or bacteria whose growth has been stopped by an antibiotic, can mutate and escape the conditions that are restricting them. Our little friend may not save the human race from destruction like Paul did, but once again yeast is proving pretty darn useful in getting results that make a difference for human health.
by D. Barry Starr, Ph.D., Director of Outreach Activities, Stanford Genetics
Categories: Research Spotlight, Yeast and Human Disease
September 26, 2013
Yeast has been responsible for a lot of hook ups in its day (think beer goggles and margaritas on the beach). Now it is payback time. In a new study, Giraldo-Perez and Goddard have figured out how to make yeast more promiscuous.
No, they don’t get the yeast drunk. Instead, they found that strains containing VDE, a homing endonuclease gene (HEG), entered meiosis more often than genetically identical strains that lacked VDE. The yeast that contained this “selfish” gene (well, actually intein) were ready to go haploid more often than those that didn’t.
VDE and its ilk are said to be selfish because they end up getting passed down to more offspring than a certain Austrian monk might have predicted. When a diploid is heterozygous for an HEG, the homing endonuclease cuts the sister chromosome at the equivalent spot. Then, when the diploid undergoes meiosis, the sister is repaired through recombination causing both chromosomes to contain the VDE gene. Now instead of two spores containing VDE, all four will.
Giraldo-Perez and Goddard monitored the percentage of sporulating cells over a 30 day period and found that after five days, a higher percentage of diploids homozygous for VDE sporulated compared to diploids heterozygous for or lacking VDE. The authors contend that under the right conditions, this increased sporulation would allow VDE to spread through a population 20 times faster than it might otherwise. And the authors found that VDE needs something like this or it might disappear.
Like alcohol, VDE isn’t all lowered inhibitions and good times. For example, yeast homozygous for VDE grow significantly more slowly than do yeast lacking VDE in YPD, grape juice, vineyard soil, vine bark (heterozygotes fall in between). This obviously puts yeast carrying VDE at a disadvantage, meaning that if it didn’t have another trick up its sleeve, it would dwindle away to nothing. That trick is speeding up sporulation.
The authors weren’t able to determine why this little bit of DNA can have such a profound effect on the growth rate of yeast. It is almost certainly too little DNA to affect the time it takes the yeast to copy its DNA. And the endonuclease itself is probably not randomly nicking the chromosomal DNA in the mitotic state, since it is kept out of the nucleus by host encoded karyopherins.
So VDE is a truly a parasitic selfish gene. It is parasitic because it sucks a little of the life out of a yeast cell. And it is selfish because way more daughters end up with it than might be predicted. Sounds like a nice description for many drunk people…
by D. Barry Starr, Ph.D., Director of Outreach Activities, Stanford Genetics
Categories: Research Spotlight
Tags: homing endonuclease, intein, Saccharomyces cerevisiae, selfish gene, VDE
September 19, 2013
Back in 2008 and 2011 there were huge spikes in the cost of food that caused riots in various parts of the world. These things were pretty bad and one of our favorite beast’s best products, ethanol, may have been at least partly to blame. In an attempt to deal with global warming, governments had created incentives that made it more lucrative to turn food into ethanol to power cars rather than keeping it as food to feed people. The law of unintended consequences reared its ugly head and caused food prices to rise high enough to be unaffordable by the very poor.

Getting yeast to turn more of this into ethanol is good for us and the environment.
This situation arose because right now, pretty much the only commercially viable way to make ethanol is to use sugars like those found in sugar cane or starches like those found in corn. Ultimately this won’t be a problem once scientists learn to coax yeast or other microorganisms to make ethanol out of agricultural waste. Until then, though, one way to lessen the impact of ethanol production on food supplies might be to engineer a yeast strain that can more efficiently turn sugars into ethanol.
One of the most inefficient parts of yeast fermentation is that the silly thing converts anywhere from 4-10% of the sugars it gets into glycerol instead of ethanol. In a new study, Guadalupe-Medina and coworkers have engineered a strain of yeast that produces 60% less glycerol and 8% more ethanol than other commercial strains. If they can scale this up, it might help us feed both the world’s population and our cars.
It has been known for some time that yeast end up making glycerol during fermentation because of redox-cofactor balancing issues. In essence, the excess NADH that is made in fermentation reactions is reoxidized by converting part of the sugar into glycerol. One obvious way to get less glycerol would be to give the yeast some other way to reoxidize its NADH.
Guadalupe-Medina and coworkers decided to persuade yeast to use carbon dioxide instead of sugars. Not only would this make sugar use more efficient, but their particular plan would also convert that carbon dioxide into a precursor that could be shunted into the ethanol producing pathway. Theoretically the yeast should now increase its ethanol production both by wasting less sugar on glycerol and by turning carbon dioxide into ethanol. And it turns out that this idea actually worked in practice.
The first step was to introduce the Rubisco enzyme into the yeast. Rubisco (ribulose-1,5-bisphosphate carboxylase oxygenase) is really one of the key enzymes in life…it provides the foundation for almost all life on the planet by fixing carbon dioxide from the air into ribulose-1,5 phosphate. But that isn’t the important point here. No, the key point for this work is that in the process of doing this, the enzyme oxidizes NADH. By putting Rubisco in yeast, the yeast should now be able to reoxidize its NADH without making useless glycerol.
Of course this is easier said than done! Rubisco is multi-subunit in most beasts and persnickety to boot. But with a bit of work, they managed to get Saccharomyces cerevisiae to express a working copy of Rubisco.
So they would only have to introduce a single gene, the authors used the single subunit enzyme from T. dentrificans. As expected, this gene alone was not enough. They knew from previous work that Rubisco would not work in yeast without the help of a couple of E. coli chaperones, groEL and groES. When they expressed all three genes at the same time, they got Rubisco to fix carbon dioxide in Saccharomyces cerevisiae.
The next step was to introduce the enzyme phosphoribulokinase (PRK) so that the ribulose-1,5 phosphate could be converted into 3-phosphoglycerate, a precursor in the ethanol pathway. Luckily this was much easier than Rubisco and worked on the first try. They had now engineered a Frankenyeast that should be able to make more ethanol and less glycerol.
When they tested the new strain, Guadalupe-Medina and coworkers found they had indeed engineered a more efficient yeast. In anaerobic chemostat conditions, this yeast made 68% less glycerol and 11% more ethanol than the usual commercial strain. They obtained similar results, 60% less glycerol and 8% more ethanol, in batch fermentations. They had succeeded in improving an already awesome beast.
If this strain works on an industrial scale and if commercial producers all used this strain instead of the ones they currently use, the authors calculate we could get an extra 5 billion liters of ethanol added to the 110 billion we are already making. That might just be enough to tide us over until scientists come up with a way to make ethanol commercially from non-food sources.
by D. Barry Starr, Ph.D., Director of Outreach Activities, Stanford Genetics
Categories: Research Spotlight
Tags: biofuel, Saccharomyces cerevisiae
September 18, 2013

Dr. Fred Sherman
The yeast community mourns the loss of Dr. Fred Sherman, who passed away on September 16, 2013. Dr. Sherman was a member of the faculty at the University of Rochester from 1961 until his death. He served as Chair of the Department of Biochemistry and then Chair of the combined Department of Biochemistry and Biophysics from 1982-1999. He performed ground-breaking research on the structure and regulation of genes and the effects of genetic mutations on proteins and was a proponent of the use of baker’s yeast as a genetic model system – a system that is now used at virtually all research centers worldwide, largely due to Dr. Sherman’s efforts and his teaching of many leaders in the field. The importance of his work has been recognized by his appointment to the prestigious National Academy of Sciences in 1985, by his receipt of an Honorary Doctorate from the University of Minnesota in 2002, by his election as a Fellow of the American Association for the Advancement of Science in 2006, and by his receipt of both the George W. Beadle Award and the Lifetime Achievement Award from the Genetics Society of America in 2006. He was continuously funded by NIH for over 45 years.
Dr. Sherman’s family will receive friends on FRIDAY September 20, from 3-7 PM at Michael R. Yackiw Funeral Home, 1650 Empire Blvd., Webster. On SATURDAY, friends may join his family for a graveside service gathering at the Mt. Hope Ave. entrance of Mt. Hope Cemetery at 11 AM. In lieu of flowers, contributions may be directed to a fund to support an annual lecture in Fred’s memory. To donate please mail donations to: Fred Sherman Lecture of the University of Rochester, Box 712, University of Rochester Medical Center, 601 Elmwood Ave., Rochester NY, 14642.
Plans for a future memorial service will be announced at a later date.
Categories: News and Views
September 10, 2013

Style is everything…and yeast has both style and substance.
Parthenogenesis is one of the cooler things in biology. When a female Komodo dragon can’t find a mate, her eggs simply double their DNA and voila, a whole litter of female Komodo dragons is born. (Interestingly, they aren’t clones of mom…)
Now, this doesn’t work in mammals like us (curse you imprinting!), but something similar can happen in yeast. Given the right conditions and the right mutations, yeast can go from haploid to diploid without all that messy mating.
In a new study out in GENETICS, Schladebeck and Mösch uncover the newest mutation to be shown to cause whole genome duplication (WGD) in haploid Saccharomyces cerevisiae: the whi3 deletion. And this mutant is no slouch…the haploid will go diploid in no time flat if given the right conditions.
Schladebeck and Mösch looked at the stability of the haploid state of the whi3 mutant in both minimal and rich media, either in liquid culture or on solid agar. They generated fresh whi3 deletion strains and then followed them in each of these growth conditions for 72 days, passaging them every two days.
What they found was that the haploid state was actually pretty stable in liquid culture using minimal media. They found very few diploid cells after 72 days. The same was not true for the other growth conditions.
On solid minimal media and liquid rich medium, there was a complete switch after 72 days. And on solid rich medium, the cells were all diploid after only 14 days. Genome duplication appeared to stop at the diploid level though. Even after 72 days on solid rich media there was no sign of tetraploids.
The authors next set out to figure out why deleting WHI3 had such a profound impact on haploid stability. They have not yet figured out everything that is going on, but they did uncover some interesting clues.
First they looked at the protein Nip100p. They already knew that NIP100 interacted genetically with WHI3, and that a nip100 deletion mutation affected chromatid separation. They found that Nip100p levels were significantly reduced when WHI3 was deleted, and even more so when the whi3 mutant strain was grown on solid rich medium. These are the conditions that most favored the transition from haploid to diploid. This suggests that NIP100 might be a key player in maintaining the haploid state.
The authors also compared transcriptional profiles of the wild type haploid strain, the whi3 deletion in a haploid background, and the whi3 homozygous mutant diploid. One of the findings from these experiments was that most of the genes involved in the yeast cohesion complex were upregulated in the absence of WHI3. Since this complex is required for sister chromatid cohesion, the idea would be that inefficient separation of chromatids in the whi3 mutant would increase the rate of whole genome duplication.
One of the as yet unexplained aspects of all of this is why the diploid state remains stable. There was no difference between the haploid and diploid deletion strains with regard to either Nip100p levels or transcription of cohesion-relate4d genes – the cohesins were upregulated in both and Nip100p was reduced in both.
One idea Schladebeck and Mösch put forth is that the diploid state isn’t inherently stable in this mutant. Instead, they do not see tetraploids simply because tetraploids have decreased viability. They appear but are quickly outcompeted by their diploid sisters.
The discovery about WHI3’s role in controlling ploidy is just one aspect of this new study. The authors also found important new information about the central regulatory role of WHI3 in cell division and biofilm formation.
The finding about ploidy control is important because maintaining haploid and diploid status is obviously a big deal: you don’t want to switch willy nilly from one to the other. And many pathogenic fungi, such as Candida albicans, change the organization of their genomes to adapt to changing growth conditions in their human hosts. They have WHI3 homologs, so these results could lead to better ways to cure fungal infections. Just one more example of how basic research can lead scientists to stumble on unexpected but ultimately important results…
by D. Barry Starr, Ph.D., Director of Outreach Activities, Stanford Genetics
Categories: Research Spotlight
Tags: ploidy, Saccharomyces cerevisiae
September 05, 2013
Huntington’s disease (HD) is a truly awful, inherited and ultimately fatal genetic disease. People with this neurodegenerative disorder typically start having trouble with their coordination and displaying mild cognitive and psychiatric problems in mid-adulthood. Their symptoms continue to worsen, with most of these folks passing away within 20 years of their diagnosis. This disease strikes down adults in their prime.
Scientists have known for decades what causes HD—too many CAG repeats in the huntingtin (htt) gene. What they haven’t been able to figure out is what to do about this misfolded protein. To date, the treatment options are very limited.

I don’t always do an unbiased screen but when I do, I prefer yeast.
A new study out by Mason and coworkers has a chance to change all of that. Using an unbiased screen in Saccharomyces cerevisiae, these authors were able to identify a class of proteins, the glutathione peroxidases, that when overexpressed protected yeast from the harmful effects of the mutant htt protein. They then followed up and showed that these proteins had a similar effect in fruit fly and mouse HD cell models as well as in a whole fruit fly model. And this isn’t even the exciting part.
There are druggable small molecules that when added to cells (or whole animals) can upregulate the activity of glutathione peroxidases. The authors used one of these molecules, ebselen, and showed that it mimicked the effects of overexpressing various glutathione peroxidases in cells and, more importantly, in whole fruit flies. When these flies were fed ebselen, their neurons degenerated at a much slower rate. Mason and coworkers have identified a small molecule that can mitigate the effects of the mutant htt protein in model systems.
While we shouldn’t get ahead of ourselves here, this is all pretty exciting news. How cool would it be if one day people with HD lived longer, happier lives because of a drug identified using our favorite model organism? (Pay attention NIH!)
Mason and coworkers looked in S. cerevisiae for open reading frames that, when overexpressed, would lower the toxicity of the mutant htt protein. They identified 317 of these, and used a variety of bioinformatics tools to group them into different pathways and gene networks.
In the end, they decided to focus on two powerful suppressors, the glutathione peroxidases Gpx1p and Hyr1p (also known as Gpx3p), for a variety of different reasons. These proteins are powerful antioxidants, and oxidative stress is known to contribute to HD symptoms. Also, these proteins aren’t already upregulated in patients with Huntington’s disease, suggesting that it might be possible to increase their activity using drug therapy.
Now of course yeast aren’t mammals, so Mason and coworkers needed to show that having extra glutathione peroxidase activity would help in mammalian cells too. And this is just what they did: adding a mouse version of glutathione peroxidase, mGPx1, suppressed cellular toxicity in mouse cells that overexpressed the mutant form of htt.
Next they tested whether activating glutathione peroxidases would have the same effect. They focused specifically on a selenocysteine-containing molecule called ebselen because it is highly bioavailable, can cross the blood-brain barrier (critical for HD) and has been used in treating stroke and noise induced hearing loss. When added to the mouse HD model cell system, ebselen had very similar effects to overexpressing mGPx1.
So upregulating glutathione peroxidase activity by either overexpressing mGPx1 or adding the small molecule ebselen appears to help in a couple of different model cell systems. But what about a whole animal? Looks like it can help there too.
Mason and coworkers looked at HD in a fruit fly. When they added mGPx1 to this model fly, various neurons in these flies were protected from the effects of HD. And they got similar results when they fed these flies the molecule ebselen.
As a final experiment, the authors wanted to figure out whether glutathione peroxidases were really having their effect because of their antioxidant activity. In one way it makes sense that this activity is why they are so effective at mitigating the effects of the mutant HD—scientists have known for a while that oxidative stress is a major contributor to symptoms of HD. But on the other hand no antioxidant therapies have worked to date for HD. In fact, if anything they made matters worse. So one thought was that there was something special about the antioxidant activity of these proteins. For these experiments, they needed to go back to yeast.
The authors looked at a variety of antioxidant proteins, including superoxide dismutase, catalases, and glutathione reductases, and none protected the yeast from the effects of the mutant htt protein. They then checked the effects of catalase and superoxide dismutase in the HD mouse cells, and again saw no effect.
It is well known that antioxidants negatively affect autophagy and that disrupting this process can make HD symptoms worse. From this the authors reasoned that glutathione peroxidases were special because they were antioxidants that did not affect autophagy. They provided support for their idea by showing that ebselen did not affect autophagy in yeast while a control antioxidant, N-acetylcysteine, did.
Once again, yeast shows why it is such an important tool in finding potential new treatments for human disease. Without the unbiased screen, it’s difficult to imagine how scientists would have found this target. You can really only do this easily in a beast like yeast.
Symptoms like these may one day be delayed because of the awesome power of yeast genetics.
by D. Barry Starr, Ph.D., Director of Outreach Activities, Stanford Genetics
Categories: Research Spotlight, Yeast and Human Disease
August 27, 2013
SGD has compiled a selection of seminal yeast literature, comprising landmark papers in yeast biology. The list is available on the SGD Wiki and includes important publications on cell biology, early genetic maps and genome surveys, and the original S288C sequencing consortium. Also listed are key papers describing the genomes of other sequenced strains of S. cerevisiae.
This new page is just one of the many resources already available on the SGD Wiki, such as What are Yeast?, Protocols, and Job listings. We encourage you to add additional information to any of the SGD Wiki pages. If you don’t already have an SGD Wiki account, please contact the SGD Help Desk to request one.
Categories: New Data
August 26, 2013
SGD staff will be attending the 26th International Conference on Yeast Genetics and Molecular Biology, which will be held at Frankfurt University, Germany, from Thursday August 29th – Tuesday September 3rd. We will be available at a table at the poster sessions to answer all your questions about SGD, give one-on-one tutorials, listen to your ideas about what you’d like to see in SGD, and demonstrate some exciting new features. We will also be presenting talks and having an open discussion at the SGD Workshop (#7) on Saturday from 4:30 – 7 PM. Please stop by to say hello!

Categories: Conferences