By planting 360 experimental plots of Arabidopsis thaliana across 30 climate zones and sequencing plant genomes annually over five years, we show that rapid adaptation to climate change is possible within 3–5 years when sufficient genetic diversity exists. However, populations in the hottest environments showed random genetic changes rather than adaptive evolution and subsequently went extinct — revealing an evolutionary breaking point where extreme heat pushes populations toward extinction.

Read the UC Berkeley press release: One-of-a-kind experiment tracked plant evolution in response to climate change at 30 sites worldwide.

We also walked through updates to our open bioinformatics software for variant and frequency calling and for population-genetic statistics from pool sequencing, and discussed how best to share the datasets with participants ahead of the first main resource paper.

A highlight was our cost-effective sequencing protocol, which lets us process thousands of plants affordably — making an experiment of this scale feasible. We also discussed efficient methods for analysing pooled sequences, early signals of rapid evolution across sites, and the future of the experiment.

It all started late in February 2020. Not COVID-19, but the over 1,600 samples from GrENE-net seasons 2018 and 2019 that arrived to their final destination, our lab. As all labs in Stanford were closing down mid-March, we were sorting through samples in our -80C freezer. And when I say we, it is mostly Yunru (Ru) Peng, the fearless technician that is spearheading sample processing, extraction, library preparation, and sequencing, of GrENE-net.

Unsurprisingly, verifying the correspondence of every sample with online spreadsheets filled by participants in ~45 different locations, took a while. But the records matched remarkably well the samples received (confession: I was a bit scared of samples miss-labeling or directly erased labels … phew).

We were setting up the molecular lab so there was a lot of ordering of all the reagents and equipment we needed (see first order below!).

Picture of Ru Peng (laboratory technician) and Moisés Expósito-Alonso (PI and GrENE-net co-coordinator), holding the first Illumina reagents to sequence the GrENE-net samples! (Note: picture taken pre-COVID times)

Because of the large amount of samples (more from 2020+ are expected), we wanted to use protocols that would be very cheap (<$3/sample). After grinding, we followed DNA extractions using the well-known Chlorophorm approach, which although difficult to handle, is quite efficient. On the way, we learned that this protocol (sometimes called CTAB method) had been first applied to plants in our own Carnegie department in the 80s! (before was normally used for bacterial DNA extractions).

But we needed throughput, so once it worked per sample smoothly [side story, the protocol was not working after the lab partially re-opened because buffers were likely unsterile — who knows what happened during pandemic lockdown], we tested materials and equipment to do 96 samples at a time (see plate glamour shot).

Plate glamour shot with 96 GrENE-net samples freshly extracted for DNA. Note there is some Chlorophyl giving the green color, and likely also other small amounts of yellow pigments from the flowers. (Picture courtesy of Ru Peng)

Same procedure was taken for the NGS library protocol. Like DNA extraction “kits”, library preparation kits are quite expensive for that many samples so we wanted a home-cooked method. We adopted the down-scaled Nextera protocol circulating across labs since Baym et al (2015, PLOS One), with the help of our colleagues Shreya Banerjee and Molly Schumer. Below you can see a beautiful large gel of 96 samples after library amplification (and before size cleanup)!

An example a gel of 96 Whole-Genome NGS libraries, where you can see bands around 500 bp. Variable intensity is likely due to differential success in preserving samples from field experiments (sometimes quite remote to the lab), as well as differential survival of plants across environments, leading to sometimes only 1-2 individuals to survive and sampled in a location. (Picture courtesy of Ru Peng)

All in all, as I write, there have been a fraction of samples for which we already have whole-genome sequences, so we will keep working on processing samples, and look forward to an exciting 2021 of more sequencing and data analysis to learn new insights on short-term evolutionary dynamics under different climates!

GrENE-net experiments have already grown for 1 generation, and we have exciting news: Arabidopsis is growing in many locations!

Map of sites where Arabidopsis grew and reproduced for the first generation. Total counts of flowers and individual records collected by all GrENE-net participants in 2018.

Arabidopsis growing in the campus of the University of Würzburg, Germany (left) and the Israeli Nagev desert (right). Credit: Arthur Korte and Merav Steifan.

Plants successfully fought the local climate even in some very extreme regions. For example, we did not expect them to grow at all in the Negev desert (Fig. 2). Researchers are currently monitoring the second flowering generation, and the coordinators are getting the sequencing pipeline ready.

Evolution must go on!

The site where the first flowers have been spotted is located in Mallorca, a mediterranean island of Spain (black star in the map). This is one of the 45 experimental sites (red circles in the map) that joined GrENE-net and planted the same mixture of seeds last fall.

The team in Mallorca is comprised by Carlos Lara, Gema Escribano, and Anna Traveset. They just took the first sample of flowers that will go into the sequencer in Summer.

Photo credit: Carlos Lara and Gema Escribano ——

A Friday of July 2017 we made a big step for the GrENE-net project. We mixed thousands of seeds to generate a common founder population. The seed mixture will then be distributed across the world in controlled experiments to investigate over time how, and whether populations will evolve to climate if all would start with the same genetic diversity.

The mix was generated from seeds of previously grown plants at three different greenhouses and growth chambers (at the Max Planck Institute in Tübingen, the Centre for Functional and Evolutionary Ecology in Montpellier, and the University of Tübingen). These plants were selected from the 1001 Genomes Project collection. A total of 230 genotypes, from a similar number of geographic locations across Europe, North Africa, and Asia.

The GrENE-net coordinators, Moises Exposito-Alonso, François Vasseur, and Niek Scheepens (from left to right), created the master seed mix with extreme care to avoid contamination of the seed mixes with pathogens or with other seeds, and used a high-precision balance in order to record the exact proportion of genotypes in the founder population.

After shaking the mix, we aliquoted ~ 1,500 tubes with ~ 5 thousand seeds that were later shipped to 45 locations around the globe. In fall 2017, the seeds were dispersed and the evolution experiment began!

Map of GrENE-net experimental sites (red circles) and the locations of origin of founder genotypes (black points).

The experiments started on fall 2017 and will run at least until 2020! Let’s see how much and whether they evolve in three generations. We expect very strong selection, as the experimental environments span from Israeli desserts to temperate North European forests.

Let the evolution begin!