First of hopefully many NY Area Population Genomics Workshops held today at the New York Genome Center. Tons of great talks across a wide domain of disciplines. Truly an intellectual shopping spree.

Long day of great talks, one simple post can't capture the diversity of subjects and approaches, but here are some highlights (NB: All the talks were great, these are just a few that felt most relevant to my work. Also, any misaprehension or mistakes in interpretation are fully my responsibility):

The question I should have asked:
Amy Williams (Cornell) gave a really interesting talk about non-crossover gene conversion. I don't know anything about this, but I do wonder how we know it's not just two recombination events happening really close together? She probably explained this but I just didn't get it ;)

This question does seem relevant in light of Chris Campbell's (Albert Einstein College of Medicine) talk Regulation of crossover placement in human meiotic recombination. He described work they're doing to characterize the distribution of recombination hotspots across the genome, and also to understand "crossover interference", the observation that recombination events have an antagonistic effect on subsequent recombinations that decays with increasing genetic distance. This would seem to indicate that consecutive tightly localized recombination events would be possible, just not probable.

It also made me wonder what the occurrence rate of non-recombination gene conversion events is. If it's significant it could have an impact on tools that infer population history taking into account recombination events. I would think significant non-recombination gene conversion rates would bias the inference and artificially inflate divergence times.

Michael Jordan award for most crowd pleasing performance:
Christopher Mason (Weill Cornell Medical College) "City-Scale Disease Surveillance and MIcrobial Mapping"
I feel like I should have paid for this performance, funnier and more interesting than lots of stand-up shows I've been to. Basically he swabbed every MTA turnstyle across the city and sequenced everything. Really interesting results. spoiler alert: everything is grosser than you think :p

Most precise random statistic:
Varun Aggarwala (UPenn) "A bayesian framework explains substantial variability in nucleotide substitution probabilities across the human genome"

Varun's work extends current models that assess nucleotide substitution probabilities based on 3mer analysis by including more information from surrounding neucliotides (7mers). Naturally including more surrounding information increases base substitution inference. Asked about model complexity he stated that his model includes "24,576 degrees of freedom". That's a LOT of freedom.

Mysterious gene of the day: Pdrm9
Active in promoting meiotic recombination.

Oh that's how that works
Molly Schumer(Princeton) Reproductive isolation of hybrid populations driven by genetic incompatibilities
Satisyfing mathematical proofs are often described as elegant. The nature of elegance relies on the complexity, and intractibility of the problem being addressed as much as it does on the simplicity of the solution. Molly Schumer gave us a good example of an elegant solution to a tricky problem in biology. She described a mechanism that generates isolation between hybrid and parental species that relies only on selection against genetic incompatibility. No fancyness, no complicated assumptions, just simple reasoning and genetics. This was probably my favorite talk.

Curious factoid
Drug resistance in malaria always seems to emerge in SE Asia.

A-ha moment
Niamh O'Hara (Fordham) The genomic basis of a cost of drought adaptation in a Brassica rapa plant population
Niamh's work demonstrated a rapid evolutionary shift in populations of field mustard in response to prolonged draught conditions. More than this it turns out there is a correlation between draught tolerance and fungal pathogen susceptibility, perhaps an evolutionary trade-off. When it's dry there are fewer fungal pathogens around to worry about, amirite?

Niamh's talk gave me probably the biggest A-ha moment (actually Diego gave it to me, but it was inspired by her talk). The observation that measurable genetic differentiation can take place over such short time-scales suggests that phylogeographic inference based on multiple temporally distributed samples could be biased. You can't step in the same river twice. The story you infer from a sample of a population actually could change as you sample across time.

Most satisfyingly pronouncable word: Heteroscedasticity

Curious factoid: Genetic incompatibilities bias toward piling up in the heterogametic sex.

Hickerlab report
Stephen Harris and Xander Xue (both members of the Hickerson lab at CCNY) both gave great talks at this event. Stephen is working on characterizing adaptation to urban environments in populations of Peromyscus leucopus (the white-footed mouse), and Xander is developing a multi-taxa site frequency spectrum algorithm to enable integrated explorations of shared population divergence histories.

Thanks to the New York Genome Center for hosting and to the organizers for putting it all together. Hidden details abound in these kinds of things, and the success of the event bespeaks the generosity of time and effort contributed by all involved.