Genocide of One: A Thriller (15 page)

“But does perfect software like that really exist?”

“No,” Jeong-hoon said simply.

Kento felt deflated. “So even considered logically, it’s pointless?”

“But fathers’ wishes should be respected,” Jeong-hoon said, smiling, and reached out
for the larger of the laptops. “Let’s take a look at this GIFT.”

They waited a moment as the software booted up, and when the computer-generated image
of mutant GPR769 appeared on the screen, Jeong-hoon shouted in surprise.
“What the—?!”

“So it looks weird? Even to an expert?”

“I’ve never seen such realistic graphics before, but it’s—how should I put it? Very
convincing.” Jeong-hoon stared intently for a time at the model of the receptor, the
ropelike structure that looped GIFT. He let out little interjections as he went along,
punctuated with an occasional chuckle. When he finished checking it out, he turned
to Kento. “This software shouldn’t exist. It’s fifty years ahead of present-day science.
No human could create this kind of software.”

“In other words, it goes beyond human intellect.”

“That’s exactly what it does. First of all, all you need to do is enter the gene sequence
and you can see the three-dimensional structure of the protein. Plus you can design
the structure of the drug that will bind with it de novo—from zero. It also predicts
the complex structure after docking. And what’s this?”

An item on the display menu said
ADMET
. Kento recognized that acronym. “It’s an index for how a drug will react within the
body. It stands for absorption, distribution, metabolism, excretion, and toxicity.”

“Right.” Jeong-hoon got it. “Pharmacokinetics and toxicity.”

“Can you even investigate ADMET with this software?”

“It’s not so unusual that it would have that function. There are other kinds of software
that predict pharmacokinetics and toxicity. But GIFT allows you to determine this
by species—human or mouse, for instance—and there’s a column for inputting genome
information so you could create custom-made treatment.”

It finally dawned on Kento how extraordinary GIFT was. “If this software is perfect,
then you wouldn’t need to do clinical trials.”

“Exactly. This software alone would allow you to run the entire drug-development process.
All people would have to do is take care of the actual process of synthesizing the
drug and checking it.”

Kento and Jeong-hoon looked at each other and laughed.

“Okay, let’s move on,” Jeong-hoon said, and turned back to the laptop. He seemed to
be enjoying playing with this unbelievable software. “Let’s find proof that the software
isn’t perfect. There’s got to be a good way.”

“Would this help?” Kento asked, and pulled down a sheaf of printouts from a shelf,
the reports on PAECS that his friend Dr. Yoshihara had downloaded for him. “A Portuguese
researcher has reported on the three-dimensional structure of that receptor.”

Jeong-hoon looked through the report. “Homology modeling. Nice,” he murmured, switching
the display on the laptop a few times. The computer-generated graphics on the screen
changed to an abstract model, a combination of spheres and ribbons. When he enlarged
the active site of the receptor, the screen showed, at the atomic level, the part
that bound with the ligand.

“Just as I thought,” Jeong-hoon said. “The two models are quite different. The atomic
coordinate values are different.”

“So GIFT’s a fake?”

Jeong-hoon frowned. “Too early to say. Logically, either the Portuguese researcher
is correct or GIFT is correct, or both of them are wrong.”

Kento was impressed by the fact that Jeong-hoon refused to give an easy answer. Tenacious
logic was a researcher’s sole weapon.

“Computer-aided drug design has run up against a wall, in fact,” he went on. “Even
using cutting-edge software it’s difficult to predict the three-dimensional structure
of membrane proteins. Most likely this Portuguese researcher is using an incorrect
model.”

Jeong-hoon booted up his own laptop, copied the DNA base sequence, and inputted it
into GIFT. “Since we know the structure of the protein is accurate, let’s compare
the responses.”

He hit the enter key and a message came on the screen, in English.
PLEASE CONNECT TO THE INTERNET
, it said.

“Why?” Jeong-hoon wondered.

Kento plugged the high-speed Internet line in his apartment into the white laptop.
Once the machine was connected to cyberspace, the display on GIFT changed:
REMAINING TIME: 00:03:11.

The numbers decreased with each passing second.

“Just three minutes?” Jeong-hoon muttered.

Three minutes later GIFT gave its response. On the screen was the three-dimensional
structure of the protein Jeong-hoon had specified. As he checked various aspects of
it, his face grew solemn. “This is really odd. This software has correctly depicted
the structure of a protein that is made up of a hundred linked amino acids.”

Kento was surprised. Didn’t this mean that GIFT was indeed perfect?

But Jeong-hoon didn’t rush to judgment. “The first thing we should consider is that
this software is fake.”

“Then how could it model the structure of the protein?”

“When it was calculating, it directed us to connect to the Internet, right?”

“Yeah.”

“Maybe it searched out an existing structure on the Internet and made it appear as
if it had calculated this itself. If you access the protein database you can find
all sorts of information like that.”

“I see,” Kento said, but soon a new question occurred to him. “If that’s true, though,
then we can’t decide whether GIFT is real or a fake.”

“Exactly. We can’t distinguish between a correct calculation GIFT’s made on its own
and a discovery that it’s copied from somebody else. Because there’s only one correct
structure. And if we make it calculate an unknown structure, no one could tell which
is correct—GIFT or some other model.”

Kento felt taken in by a cunning trick. But if GIFT really were a hoax, who would
have come up with such an elaborate prank? And why?

“Was your father familiar with programming?” Jeong-hoon asked.

“No. Not at all.”

“Then where could he have gotten hold of this software?”

“I have no idea.”

The name of the software—GIFT—was starting to take on an ominous tone.

“The other possibility is the hypothesis that GIFT is actually perfect. Though it’s
just a hypothesis,” Jeong-hoon emphasized. “This program might have distributed computing
software in it.”

“Distributed computing?”

“That’s the method used by the SETI-at-home project in their search for extraterrestrial
life. By analyzing all the radio waves that reach Earth, they try to detect any artificial
signals, all of which takes tremendous computing power. So what they did was connect
the personal computers of volunteers and use part of their CPUs for the calculations.
If you link tens of thousands of personal computers like that to do the calculations,
you get more computing power than a supercomputer.”

“Could I change the subject for a moment?”

“Sure.”

“Did they find evidence of aliens?”

“Not yet.”

Kento felt a bit let down.

“Six times, though, they’ve detected unknown radio waves coming from the center of
the Milky Way. The waves remain a mystery to this day. Astronomers from all over the
world have already decided on a protocol for announcing this news if they do find
evidence of extraterrestrials.”

“Wow.”

“Let’s get back on topic.”

“Go ahead.”

“Let’s consider one of the functions of this program—to calculate the complex of receptors
and ligands. There are two important factors that determine how well the software
will perform this function: the machine’s computing power and the algorithm.”

“By algorithm you mean computational procedure?”

“Right. The method that will arrive at the correct answer through the least number
of computational procedures, omitting any unnecessary calculations.”

Kento tried his best to keep up. This was outside his area of expertise.

“First of all, if this computer is letting other computers do the calculations through
distributed computing, it has plenty of computing power. But even if you link a hundred
million computers together, it’s impossible to do perfect molecular dynamics calculations.”

Kento was aware of this, too. It was precisely because perfect calculations were impossible
that after you made a drug you collected structure-activity relationship data and
investigated a more optimal chemical structure. Technologically advanced nations competed
so ruthlessly to develop supercomputers precisely because they were living in an age
when computing power and science and technological advances were directly linked.

“And labor-saving computational procedures—algorithms, in other words—make up for
the lack of computing power. Researchers have tried many different methods, but never
found one that’s perfect. No two algorithms will give you the same answer. That’s
the limit of present-day science. In other words, computing power is insufficient,
and nobody’s discovered a perfect computational procedure.”

Kento said, “So GIFT can’t be perfect is what you’re saying.”

“Logically, that’s the case,” Jeong-hoon replied. His expression remained unconvinced.

“Is there something else bothering you?”

“I’m not sure how to put it…Would you call it touch, maybe? Feeling?”

“You mean like a sense?”

“That’s it. You get a strange sense when you use this software.”

“What, exactly?”

“I don’t know how to express it,” Jeong-hoon said, scratching his head and trying
to translate the sense of discomfort he felt into Japanese. “It’s like, when you use
it, you sense the program really
is
all-powerful.”

This was a feeling that only Jeong-hoon, well versed in all sorts of software, could
understand.

“Whatever researcher created this software is amazing. He made it appear to reveal
tremendously complex biological activity down to the level of molecules and electrons.
If you really can use this software, it deserves a Nobel Prize. A bunch of them.”

Kento felt the same way.

“But I can’t find any direct proof that it’s fake. It’s really an amazing program.”

“Why would somebody create this software?”

“It’s a mystery. A specialist would recognize right away it’s impossible, while an
amateur wouldn’t know what kind of software it was.”

Kento was startled. “So a researcher in another field might be fooled by it.”

“What do you mean?”

“Someone like my father, who was in virology. He might have been told that it was
a program that could do anything related to drug development and be totally taken
in by it.”

Yuri Sakai came to mind. It wasn’t clear what sort of relationship she and his father
had had, but maybe she was the one who brought him this program and suggested it could
cure a chronic disease. She’d reeled him in with the promise of all the royalties
he could earn. She planned to embezzle all the private research money his father had
sunk into it and then disappear. The bank account under another person’s name with
the large amount of research funds in it had to be money Yuri was going to have him
transfer to her.

But then why, after his father’s death, did Yuri risk seeing Kento? It made sense
if the small laptop she wanted to get back had some electronic data—e-mail exchanges,
for instance—that might prove her fraudulent activities. She could ignore the computer
with GIFT installed on it, because that alone wouldn’t reveal where it originated.

“How gullible can you get?” Kento said angrily.

“But your father must have had some good points, too,” Jeong-hoon said, trying to
smooth things over.

The only thing that didn’t fit this scenario, though, was the Heisman Report. What
was in the fifth section? Kento still hadn’t heard back from Mr. Sugai. Maybe he should
ask him not just about the Heisman Report but also about this drug development fraud
as well. Depending on how it played out, he had to be prepared for the police to get
involved.

“Can I check out this laptop?” Jeong-hoon asked, pointing to the computer with GIFT
installed. “I want to fool around with it a little more.”

“Uh, sure.”

“Thanks.”

They talked for another hour, exchanged cell phone numbers, and Jeong-hoon left before
midnight. During their visit Kento found out more about this Korean exchange student.
He’d skipped a grade in high school and entered college at seventeen, so he must be
unusually bright. His excellent Japanese was something he’d mastered just through
lessons at school. And when he’d had to take time out from college to serve in the
military, he’d worked at a US military base and had picked up English as well. Skipping
grades, being drafted into the military—the situation with students in Korea certainly
was different from that in Japan.

After Jeong-hoon left, Kento felt the special kind of joy that comes from making a
new friend, someone he could really get along with. He took a shower in the tiny bathroom,
brushed his teeth, and got ready to go to sleep. As he lay down in bed, he came to
a final conclusion. Since he knew they couldn’t use GIFT, he decided that the research
his father had left him was impossible. He had to give up the idea of ever developing
a drug to treat PAECS. Kento, used to disappointment, told himself that it was absurd,
that trying to save the lives of one hundred thousand children had been totally beyond
him from the beginning.

Still, he couldn’t forget one tiny girl, fighting to breathe, her mouth bloody.

Maika Kobayashi.

The girl was only a twenty-minute walk from his apartment. Lying in a hospital bed,
gasping for air, unable to get enough oxygen. The girl was alive now, but in a month
she will have vanished from the face of the earth.