CSHL Biology of Genomes Wrap-Up

After five grueling but interesting days, the Cold Spring Harbor Labs Biology of Genomes has wrapped up. So where is genomics heading?

A few lessons:

– Single cell genome sequencing is here, and it’s time to think of applications

– ENCODE is still here, and is still generating lots of [insert method name here]-seq data. ENOCDE is inevitable; don’t fight ENCODE, just resign yourself to the fact that you and everyone else will be assimilated

– rare genetic variants will not explain the ‘missing heritability’… or maybe they will, or at least they can tell us interesting stories about human populations

– With unlimited sequencing power, we can now watch tumors evolve almost in real time, but it’s not clear what we do with that information

– Human Y chromosomes are no longer shrinking, and men will not go extinct

– Here’s a shocker: the trans genome is larger than the cis genome – if you randomly mutate the genome and look at the effects on gene expression, you will get more trans mutants than cis mutants.

– Using pigeons to study evolution is cool again after 150 years.

My overall thoughts: This year was very population-genetics heavy, which partly reflects the tastes of the conference organizers, but it also reflects the fact that we are now in possession of data sets of human genetic variation that are so large, that we can know apply population genetic thinking to new areas like tumor evolution.

Also, there is a tremendous amount of data collection – RNA-seq, ChIP-seq, histone marks, DNase hypersensitivity, SNPs, CNVs, etc. but most of this is accompanied by general observations, and not so much by models and ideas that researchers are testing. Few speakers posed a compelling, important question and then proceeded to definitively answer it. Our measurement technology is amazing, and we’re finding correlations between all sorts of phenomena, but we aren’t spending enough time defining our models, and taking risks with our specific ideas.

Often people tell me that all sciences have observational phases, and then theory phases, but I don’t believe that. I don’t believe genuine scientific progress is made without posing strong hypotheses and testing them. Darwin came up with natural selection because he did not come to biology simply as an observer and a classifier – he was constantly coming up with models and hypotheses and putting them up against his observations. Genomics should be more than an observational science.

I’m bushed right not, but stay tuned for some specific discussions of what happened at the meeting.