BU Today

Campus Life + Science & Tech

Genomes Have Bumps

BU scientists find that shape is part of DNA’s code


A graphic of DNA topography. Courtesy of the National Human Genome Research Institute. (Below) Thomas Tullius, a CAS professor of chemistry. Photo by Vernon Doucette

In 1953, Francis Crick and James Watson discovered the double helix of DNA, the elegant simplicity of four base pairs connected in long, twisting strands that write the genetic instructions for all life.

They didn’t know the half of it.

A collaboration of Boston University and National Institutes of Health scientists has discovered that the shape of DNA – the molecule’s width and its nooks and crannies — may be as important as the base-pair sequences in translating genetic code into, well, you. Their findings, which appeared in the March 12 online edition of Science, could revolutionize genomics and help unravel the genetic underpinnings of disease.

“DNA is not a perfectly uniform double helix,” says Thomas Tullius, (right), a College of Arts & Sciences professor of chemistry, who collaborated on the research with bioinformatics doctoral students Stephen Parker (GRS’09) and Loren Hansen (GRS’10), along with researchers at the NIH’s National Human Genome Research Institute led by NHGRI investigator Elliott Margulies. “There are subtle differences in the shape of the molecule that are very important to the binding of proteins.” These proteins, known as transcription factors, allow DNA to produce yet more proteins and also regulate when, where, and how much of these vital compounds are created.

Tullius and his fellow researchers compared one tiny, well-studied part of the human genome with the same region from the genomes of 36 mammals, including mice, chimpanzees, and rabbits. They were looking for overlaps, believing that if a portion of the genome is the same across many species, it must be important.

“Mutations occur randomly,” explains Parker. “But they’re rejected in functionally important areas.”

A few years ago, other researchers did a similar cross-species comparison and found that about 6 percent of the genome section was constrained by evolution. But that study looked only at base pairs. The BU researchers incorporated the shape of each sequence in their comparison, using an algorithm developed by Parker. When shape was taken into account, the percentage of the genome that passed evolutionary muster doubled.

According to Tullius, this indicates that a lot of DNA’s function may be determined by the molecule’s shape, even if a few base pairs have been shuffled across the generations.

The researchers also found that DNA shape alters much more as a result of genetic mutations known to cause disease or biological change versus mutations that aren’t known to impact any biological function.

“What that means is that larger structural changes correspond to biological consequences,” says Parker (right).

This could impact the booming field of research into associations between genetic mutations and disease. Only a fraction of mutations known to impact biological function exist among the 2 percent of the human genome that directly codes for proteins. The vast majority of them, including more than 700 mutations studied by the BU team, exist among the remaining 98 percent of base pairs, whose role is understood less well.

“They’re out there in terra incognita,” says Tullius, “where you have a very hard time figuring out what went wrong and why this particular change led to heart disease or dementia.” A focus on DNA shape offers a new way to interpret these mutations.

The next step will be to expand this analysis to the entire genome, creating a topographical DNA map that includes identification of biologically important regions through comparisons with other mammals, transcription proteins that bind to different DNA shapes, and mutations that affect those structures.

“There’s all this stuff that’s invisible to us now,” says Tullius, “but we’re getting a clue about what’s happening out there.”

Chris Berdik can be reached at cberdik@bu.edu.


8 Comments on Genomes Have Bumps

  • Anonymous on 03.25.2009 at 1:02 pm

    Wasn't this known before?

    It goes to show you should talk to the experts. Back in the 80s when I read Scientific American for pleasure, I always assumed that transcription binding sites were linked to the shapes of specific DNA sequences. It is only natural that genetic flaws would affect transcription selection as well as completion. I am sure the discovery is much deeper here, the actual algorithm applied was a major development – still I am shocked that a concept available to me as a layman is considered ‘revolutionary’. It is a reminder … good ideas are worth sharing, sometimes they are better ideas than you thought :)

  • Anonymous on 03.25.2009 at 3:32 pm

    Link to the article

    Why do pieces like this always seem to omit a link to the journal article they’re talking about? It’s the web, folks — please make a link! This article can be found here:


  • cberdik on 03.25.2009 at 5:19 pm

    Article Link

    We provide links to articles when the actual article is available, not just an abstract. Unfortunately, Science keeps its stuff under lock and key.

  • Anonymous on 03.26.2009 at 10:24 am

    Re: Wasn't this known before?

    Good question. This is one of the reasons biology is such an exciting field right now: it is wide open for discoveries like this.

    That said, I would challenge anyone to try and find a way to measure the shape of DNA using only technology available in the 1980s. The confluence of automation, sophisticated algorithms and faster computer hardware are all important enablers for the ingenious study described here.

  • Anonymous on 03.26.2009 at 10:35 am

    Re: Article Link

    I take your point. It is frustrating that the findings of this research, much of it funded with public dollars, is not available to the public. By charging for access, Science is hampering one of the AAAS’s primary missions: to disseminate scientific discoveries.

    However, refusing to link for this reason alone only serves to reinforce that problem. It might be understandable for a mainstream news publication, but consider your audience: how many BU Today readers have access to Science? That number is probably very close to 100% because BU has a subscription to the journal.

  • cberdik on 03.26.2009 at 1:00 pm

    Re: Re: Article Link

    Sure. No problem. We can link to abstracts in the future.

  • B. Gregory Johnson on 03.26.2009 at 4:54 pm

    Very cool...

    We’ve sure come a long way from X-ray crystalography and electrophotopherisis.

    Keep up the good… and very cool… work!

  • Anonymous on 03.26.2009 at 8:54 pm

    BU Should Be Proud!

    What extraordinary and fascinating work! Thanks to the Tullius lab for helping to keep Boston University on the cutting edge of science. What a phenomenal accomplishment to publish in such a respected journal and garner so much media attention- well done to everyone who contributed to this project!

Post Your Comment

(never shown)