How the Genome Ball Was Created
“One of the biggest challenges in cell biology is that it’s just so hard to figure out what’s going on inside of cells … No one’s ever been able to take a picture of what an entire genome in a single cell looks like.” – Erez Lieberman- Aiden
How do you build a model of the human genome? Few people ask that question, and even fewer try to answer it. Yet a group of talented young scientists recently did just that – creating an elegant 3-D model of how the long DNA molecules fold inside a tiny cell nucleus.
If we have to pick a starting point for the Genome Ball project, 2009 is reasonable. That year, a group that included Erez Lieberman-Aiden, a graduate student at Harvard and MIT, published a paper in Science about the folding of the human genome.2 A figure in that paper (see at right) shows how the nucleus would look if the DNA were “highly entangled” vs. “lacking knots.” The structure that lacks knots looks a lot like the Genome Ball we see today.
Miriam Huntley, a graduate student in applied physics at Harvard, was intrigued by Erez’s questions about how the long DNA molecules fold up inside the nucleus. “There are a lot of different ways you could try to answer these questions,” says Miriam. “One avenue of pursuit is to do more experiments … and one avenue is to do simulations. I started doing simulations.”
But “how should you model DNA?” turned out to be a tricky question. “We don’t actually know all the facts about DNA, even the basic physical properties like how bendable it is,” says Miriam. “Whenever anybody models DNA, you have to choose how to represent DNA. The way I chose to do it was to simplify it dramatically.”
She used an approach called lattice-based Monte Carlo. “Basically, instead of assuming that the DNA is this wiggly long polymer, we simplify it by assuming that DNA consists of a bunch of small balls that are all connected to each other in a string. And then you let the balls move around on a lattice.”
Her simulations gave the X-Y-Z coordinates of the DNA polymer in 3-D space -- that is, the location of each small ball making up the long polymer. “I had to join up all the balls and link them all together,” recalls Miriam. “Basically, that’s what I did to get it into a format that can be used for 3-D printers.”
Supermodel (you better work)
After running additional experiments on DNA folding, Erez contacted the Molecular Graphics Laboratory (MGL) at The Scripps Research Institute: Could they produce a 3-dimensional model based on Miriam’s numerical simulations?
Jon Huntoon was the first to tackle the project. But before Miriam sent the coordinates of the long polymer to the 3-D printing crew, she tweaked her results to make them look less “griddy.”
“If you look at [the Genome Ball], it looks more squishy and more bendy. The way I did this is I used something called ‘off-lattice relaxation’,” said Miriam. “If you imagine all the balls are pinned to their grid sites, you sort of let the pins go and let it relax a little bit … It’s more bendable and looks more like DNA.”
“This was a very challenging project for me because the structure has no natural supports to counteract gravity, so I relied on careful extraction and lots of super-glue to get the job done.” – Jon Huntoon
Using a dizzying array of software, Jon represented the nuclear DNA as a tube or pipe passing through a “cloud” of points. “As the pipe runs through each point, it must turn, twist, and bend to reach the next point,” he explained. Before he left Scripps in 2011, Jon had produced a model about the size of a softball. He also worked briefly with Adam Gardner, a new member of the MGL team. In 2013, when Erez wanted to make more models, Jon referred him to Adam.
“I showed Adam how I had made the models before, and then he ran with it,” Jon recalls. “Adam has developed new techniques for modeling and finishing the models … far and away better than what I had accomplished … The baton at MGL was handed to a very capable individual.”
Adam Gardner earned his bachelor’s degree in biology, but he has a strong love for the arts and was a self-taught expert in 3-D modeling and computers. “In high school, I spent all my time on computers. I wanted to be in video games back then, so I learned all the techniques that video game designers and modelers were using to create 3-D models for games. I applied that to biology and, with the new revolution in manufacturing and 3-D printing, it just kind of fit together.”
Adam is also good at explaining the complexities of 3-D printing. “You can imagine it like a normal ink jet printer, because essentially it is,” said Adam. “If you take an object and you do an MRI cross-section, cut it in one plane over and over and over, and then stack those on top of each other, that’s essentially what 3-D printers do.”
“What we’ve done before is use software and stuff and visualize it within a computer. But now we can actually print the structures out, and you can put your hand into the nooks and crannies, really get to feel the molecular structure.” – Adam Gardner
Like other printers, a 3-D printer can handle only a limited size, so he had to find a printer large enough to print the Genome Ball in eight individual sections: Imagine a globe sliced across the equator, with each hemisphere cut into four equal wedges. “Each print takes around 20 hours,” said Adam, “so you can imagine how long the entire thing took!”
The 3-D printer at the MGL used powder – normal plaster of Paris, the same stuff that sculptors have used for years. When the printer sprays food coloring and water onto the powder, it polymerizes, dries, and hardens into a layer about 0.001 millimeter thick.
Whatever isn’t printed (not sprayed with water and food coloring) remains powdery. “You can just go in and gently, gently, get the powder out, because at that point it’s super fragile,” said Adam, who used small brushes and compressed air to remove the residual powder.
The Genome Ball’s size posed additional challenges: “The bigger the product is, the heavier it is,” says Adam. “So the bigger you print it, the more you have to support it with something if there are any overhangs – and in this structure [the Genome Ball], there’s a lot of overhangs!” He even used engineering software to design a well-concealed internal support structure.
We didn’t start the fire
Sometimes the most interesting parts of an innovative project are the problems that need to be solved. For instance, infiltrating the fragile plaster of Paris model with epoxy:
“We usually use cyanoacrylate, which is super glue,” said Adam. “But for this we used epoxy because it’s a lot stronger and it holds up to UV light a lot better. Super glue will crack and chip, but epoxy kind of flexes and gives a little more strength.”
When Adam tried to infiltrate the model by submerging it in a large container of liquid epoxy, the resin hardened too quickly, leaving his Genome Ball encased in a block of solid epoxy! Although it looked impressive, a giant paperweight wasn’t quite what he’d had in mind.
After that, Adam played it safe by mixing extra hardener into the liquid epoxy and dripping it on the model. Even that caused an exothermic (heat-producing) reaction, and the more epoxy used the more exothermic it was. “I think we’re really at the limit of the size you can infiltrate with the [type of] epoxy we used,” said Adam. “We didn’t start any fires … but it did get very very hot.”
“You always make mistakes with things like this, so it’s kind of trial and error. Eventually, you get something right!” – Adam Gardner
Predictably, the printer broke several times. “We’re pushing it to the max,” Adam explained. “You’re printing as many layers as it possibly can print, and so if it’s ever going to fail, you’re giving it a lot more chances.” Considering that each section of the ball took 18-20 hours of printing, it’s surprising that the printer survived at all!
The rainbow connection
Once the challenges had been overcome, the Genome Ball now exhibited at the Smithsonian looked like an intricate sculpture, especially beautiful in living color. “We had to make some decisions about what colors we wanted to use,” Miriam had said. In fact, color information was included in the original numerical simulations she’d sent to Jon and Adam.
“I like the rainbow colors,” she added. “It’s beautiful but it also conveys information.” Colors in the model vary continuously along the length of the strand, so points further apart have more visibly different colors. “This was useful because, when the whole ball is folded up, it’s really hard to tell, without colors, how close any two points were.”
Looking ahead, Adam plans to pursue graduate work in computational or structural biology. But regarding his role in creating the Genome Ball, his response is surprising: “I don’t want people to think at all about how I made it. I want that to be kind of hidden – like I tried to make the supports as minimalistic as possible … I’d want people to be able to ask questions that trigger their imagination and enthusiasm,” he says. ”Really, I just want them to get curious about science and learning.”
In that respect, and in so many others, the Genome Ball is a resounding success.
1 Cooney, Elizabeth. “Broad researcher receives GE-Science Prize.” Broad Institute (December 1, 2011).
2 Lieberman-Aiden, Erez, et al. “Comprehensive Mapping of Long-Range Interactions Reveals Folding Principles of the Human Genome.” Science 326:289-293 (October 9, 2009).
(1) See a 9-panel color infographic of the genome ball 3D printing process.