From: Jim Whitehead (ejw@cse.ucsc.edu)
Date: Tue Feb 13 2001 - 10:47:10 PST
Wow, where can I find a graduate student like this? ;-)
- Jim
http://www.nytimes.com/2001/02/13/health/13HERO.html
Grad Student Becomes Gene Effort's Unlikely Hero
February 13, 2001
By NICHOLAS WADE
A surprising hero helped the consortium of academic scientists
decoding the human genome to avoid a drubbing by its rival, the
Celera Genomics company. Scientists throughout the world are now
beating an electronic path to his Web site, where they can analyze
and download the human genome sequence. He is a graduate student at
the University of California at Santa Cruz, and his name is not
Clark but James Kent.
In four hectic weeks last spring, Mr. Kent wrote a computer
program that the consortium's leaders hadn't realized how much they
needed, one that assembles the 400,000 fragments of DNA they had
decoded into a coherent sequence. Using 100 computers that his
senior colleague, Dr. David Haussler, had persuaded the university
to buy for the purpose, Mr. Kent performed his first assembly on
the human genome on June 22, just four days before Dr. Francis S.
Collins, the consortium's informal leader, and Dr. J. Craig Venter
of Celera, announced at the White House on June 26 that each had
assembled the human genome.
Since Dr. Venter has now stated that Celera finished its computer
assembly just the day before, on the night of June 25, it turns out
that Mr. Kent's brilliant improvisation was the first assembly of
the human genome, even though one that had and still contains many
gaps.
"Without Jim Kent, the assembly of the genome into the golden path
wouldn't have happened," said Dr. Collins, referring to the
nickname for the GigAssembler, as the program is known.
Dr. Haussler, a professor of computer science, described his
student as a superstar. "He's unbelievable," Dr. Haussler said.
"This program represents an amount of work that would have taken a
team of 5 or 10 programmers at least six months or a year.
"Jim in four weeks created the GigAssembler by working night and
day," Dr. Haussler said. "He had to ice his wrists at night because
of the fury with which he created this extraordinarily complex
piece of code."
Having finished the GigAssembler, Mr. Kent then wrote another
program known as a browser that enables many other kinds of genomic
information to be aligned in tracks above the raw sequence of DNA
bases, the chemical letters in which the human hereditary
information is encoded. This extra information is essential for
making sense of the DNA sequence and in particular for discovering
where the genes are.
How did a mere private see the weakness in the order of battle
that the consortium's generals had missed? Their competitor, Dr.
Venter, had always relied on heavy-duty computation to hedge the
enormous risks he took in his sequencing strategy. In setting up
Celera he ordered a machine with vast processing power and four
terabytes of memory, creating the largest computer in civilian use.
The consortium's DNA sequencing centers had nothing to match it
because their strategy did not seem to require large-scale
computation. The consortium breaks the human genome into large
pieces known as BAC's and works from a simple map that shows how
one BAC overlaps with the next in a tiling path that extends across
the genome.
In December 1999, Dr. Eric S. Lander of the Whitehead Institute,
who heads the consortium's genome analysis group, approached Dr.
Haussler for help finding the genes in the human genome sequence,
the first step in annotating it. Dr. Haussler decided the genome
could not be properly analyzed until the pieces being generated by
the consortium were in the right order.
The pieces of DNA were so small that their average size was less
than that of the average gene.
Most of the consortium's BAC's were unfinished, each one
consisting of up to 80 different pieces of unknown orientation and
of unknown order within each BAC — hardly worthy of being called a
sequence, as Dr. Venter liked to observe of his competitors'
efforts.
Dr. Haussler believed there was enough information, some generated
inadvertently by the consortium and some available from other
sources, to orient and align the BAC fragments correctly.
He immediately started writing an assembly program. He persuaded
the university's chancellor, Dr. Marcia Greenwood, that there was a
chance to make a historic contribution and that she needed to
advance him $250,000 to buy 100 computers with Pentium III
processors. It was a previous chancellor of U.C. Santa Cruz, Dr.
Robert L. Sinsheimer, who in 1985 held the first meeting to explore
the idea of sequencing the human genome.
But progress was slow. In May 2000, when Mr. Kent sent an e-mail
message to his colleague to ask how the assembly program was going,
Dr. Haussler said he had to reply, "Jim, it's looking grim."
Mr. Kent sent an e-mail message back that said he felt he could
write an assembly program using a simpler strategy. "I sent back a
one word e-mail, `Godspeed,' " Dr. Haussler said.
Mr. Kent, who turned 41 last Saturday, is a graduate student in
his second career. In his first, which lasted more than 10 years,
he ran a computer animation programming business.
Then he decided to go back to school and become a biologist. He
started analyzing the DNA of C. elegans, the laboratory roundworm.
When Dr. Haussler accepted the human genome assignment, Mr. Kent
started adapting his worm programs to human DNA.
Mr. Kent said he offered to write a human genome assembly program
for the consortium because of his concern that the genome would be
locked up by commercial patents if an assembled sequence was not
made publicly available for all scientists to work on. "The U.S.
Patent Office is, in my mind, very irresponsible in letting people
patent a discovery rather than an invention," he said. "It's very
upsetting. So we wanted to get a public set of genes out as soon as
possible."
After receiving his supervisor's "Godspeed" message, Mr. Kent
started work on May 22. By June 22 he had written the GigAssembler
together with subsidiary programs, in 10,000 lines of code, and had
completed his first assembly of the human genome. "The two hard
things, one is the repeat structure," he said, referring to the
numerous very similar sequences of DNA letters in the human genome,
"and then for me, something that makes the engineering very
challenging is that you read DNA in two directions, so you are
merging a lot of code that has eight directions."
The program must decide which of DNA's two strands each of the
fragments belongs to because the sequences of letters in each
strand run in opposite directions.
Mr. Kent then started work on another program, called the U.C.S.C.
browser, which allows the assembled DNA sequence to be viewed in
terms of its structural features, like genes and chromosomes. The
browser allows one to search 21 tracks of information, each aligned
with the underlying DNA sequence. The extra information helps to
identify genes.
Although the consortium had deposited its BAC data in GenBank, a
public DNA database, the U.C.S.C. browser gave biologists their
first opportunity to view the assembled human genome unless they
were subscribers to Celera's database. On July 7 the browser was
posted on the World Wide Web. There was an overwhelming response.
On that day the U.C.S.C. servers put out half a terabyte of
information, Dr. Haussler said, and the browser now gets 20,000
hits a day from all over the world. "Nothing crashed, we just kept
putting it out," he said. "People wrote back it was fantastic we
had put it out with no intellectual property requirements, a gift
with no strings attached."
Mr. Kent has designed the browser so that other biologists can add
tracks of data to it, increasing the wealth of annotation. "It's
been a wonderful stone soup, where other people have contributed
bits," he said.
The consortium now lists the U.C.S.C. browser as the principal
source for viewing its human genome sequence.
The consortium's three major centers, with their large staffs of
computational biologists, did not write an assembly program of
their own because they had not needed one in their pilot project,
on the worm genome, and perhaps did not realize how useful such a
program would be. The National Center for Biotechnology Information
completed an assembly program only recently.
"It's easy for Venter to say of course you should have had this
all planned out," Dr. Haussler said. "But he had the ability to
organize an industrial-scale effort. Given the culture of the
public project, it would have been very difficult to graft on an
autocratic new software project, and if you had assigned a team to
build this whole thing over several months it would have been a
very difficult proposition. So what Jim has done is miraculous in
many ways. No one expected anyone could come in and put this
together in four weeks."
An even more telling accolade comes from Dr. Venter, who was
astonished that the consortium managed to put the human genome
sequence together, given what he regarded as the poor quality of
its data.
"They used every piece of information available," he said. "It was
really quite clever, given the quality of their data. So honestly,
we are impressed. We were truly amazed because we predicted, based
on their raw data, that it would be nonassemblable. So what
Haussler did was he came in and saved them. Haussler put it all
together."
This archive was generated by hypermail 2b29 : Tue Feb 13 2001 - 10:54:34 PST