He pointed to the image. "This is a typical example, because you see the DNA has an error, down here in line 1201. Much of the DNA we extract is fragmented or incomplete. So the first thing we have to do is repair it-or rather, the computer has to. It'll cut the DNA, using what are called restriction enzymes. The computer will select a variety of enzymes that might do the job."
1 GCGTTGCTGGCGTTTTTCCATAGGGTCCGCCCCCCTGACGAGCATCACAAAAATCGACGC
61 GGTGGCGAAACCCGACAGGACTFITAAAGATACCAGGCGTTTCCCCCTGGAAGCTCCCTCG
NspO4
121 TGTTCCGACCCTGCCGCTTACCGGATACCTGTCCGCCTTTCTCCCTTCGGGAAGCGTGGC
181 TGCTCACGCTGTAGGTATCTCAGTTCGGTGTAGGTCGTTCGCTCCASGCTGGGCTGTGTG
BrontIV
241 CCGTTCAGCCCGACCGCTGCGCCTTATCCGGTAACTATCGTCTTGAGTCCAACCCGGTAA
301 AGTAGGACAGGTGCCGGCAGCGCTCTGGGTCATTTTCGGCGAGGACCGCTTTCGCTGGAG
434 DnxTl AoliBn
361 ATCGGCCTGTCGCTTGCGGTATTCGCAATCTTGCACGCCCTCGCTCAAGCCTTCGTCACT
421 CCAAACGTTTCGGCGAGAAGCAGGCCATAATCGCCGGCATGGCGGCCGACGCGCTGGGCT
481 GGCGTTCGCGACGCGAGGCTGGATGGCCTTCCCCATTATGATTCTTCTCGCTTCCGGCGG
541 CCCGCGTTGCAGGCCATGCTGTCCAGGCAGGTAGATGACGHCCATCAGGGACAGCTTCAA
601 CGGCTCTTACCAGCCTAACTTCGATCACTGGACCGCTGATCGTCACGGCGATTTATGCCG
Nsp04
661 CACATGGACCCGTTGCTGGCGTTTTTCCATAGGCTCCGCCCCCCTGACGAGCATCACAAA
721 CAAGTCAGAGGTGGCGAAACCCOACAGOACTATAAAGATACCAOOCOTTTCCCCCTGGAA
924 Caoll I DinoLdn
781 GCGCTCTCCTOTTCCOACCCTOCCOCTTACCOGATACCTOTCCOCCTTTCTCCCTTCGGG
841 CTTTCTCAATOCTCACOCTGTABGTATCTCAGTTCGGTOTAGGTCGTTCOCTCCAAOCTO
901 ACGAACCCCCCOTTCAGCCCGACCGCTGCGCCTTATCCGGTAACTATCGTCTTGAOTCCA
961 ACACOACTTAACCOOTTOOCATGGATTGTAGGCGCCGCCCTATACCTTGTCTOCCTCCCC
1021 GCGGTGCATGOAOCCOGOCCACCTCGACCTGAATOGAAGCCGOCGOCACCTCOCTAACOG
1081 CCAAGAATTGGAGCCAATCAATTCTTGCGGAGAACTGTGAATGCGCAAACCAACCCTTGG
1141 CCATCGCGTCCGCCATCTCCAGCAGCCGCACGCGGCGCATCTCGGGCAGCGTTGGGTCCT
1416 DnxTI
SSpd4
1201 GCGCATGATCGTGCT:+=:CCTGTCGTTGAGGACCCGGCTAGGCTGGCGGGGTTGCCTTACT
1281 ATGAATCACCGATACGCGAGCGAACGTGAAGCGACTGCTGCTGCAAAACGTCTGCGACCT
"Here is the same section of DNA, with the points of the restriction enzymes located. As you can see in line 1201, two enzymes will cut on either side of the damaged point. Ordinarily we let the computers decide which to use. But we also need to know what base pairs we should insert to repair the injury. For that, we have to align various cut fragments, like so."
[picture]
"Now we are finding a fragment of DNA that overlaps the injury area, and will tell us what is missing. And you can see we can find it, and go ahead and make the repair. The dark bars you see arc restriction fragments-small sections of dinosaur DNA, broken by enzymes and then analyzed. The computer is now recombining them, by searching for overlapping sections of code. It's a little bit like putting a puzzle together. The computer can do it very rapidly."
1 GCGTTGCTGGCGTTTTTCCATAGGCTCCGCCCCCCTGACGAGCATCACAAAAATCGACGC
61 GGTGGCGAAACCCGACAGGACTATAAAGATACCAGGCGTTTCCCCCTGGAAGCTCCCTCG
121 TGTTCCGACCCTGCCGCTTACCGGATACCTGTCCGCCTTTCTCCCTTCGGGAAGCCTGGC
181 TGCTCACGCTGTAGGTATCTCAGTTCGGTGTAGGTCGTTCGCTCCAAGCTGGGCTGTGTG
241 CCGTTCAGCCCGACCGCTGCGCCTTATCCGGTAACTATCGTCTTGAGTCCAACCCGGTAA
301 AGTAGGACAGGTGCCGGCAGCGCTCTGGGTCATTTTCGGCGAGAACCGCTTTCGCTGGAG
361 ATCGGCCTGTCGCTTGCGGTATTCGGAATCTTGCACGCCCTCGCTCAAGCCTTCGTCACT
421 CCAAACGTTTCGGCGAGAAGCAGGCCATTATCGCCGGCATGGCGGCCGACGCGCTGGGCT
481 GGCGTTCGCGACGCGAGGCTGGATGGCCTTCCCCATTATGATTCTTCTCGCTTCCGGCGG
541 CCCGCGTTGCAGGCCATGCTGTCCAGGCAGGTAGATGACGACCATCAGGGACAGCTTCAA
601 CGGCTCTTACCAGCCTAACTTCGATCACTGGACCGCTGATCGTCACGGCGATTTATGCCG
661 CACATGGACGCGTTGCTGGCGTTTTTCCATAGGCTCCGCCCCCCTGACGAGCATCACAAA
721 CAAGTCAGAGGTGGCGAAACCCGACAGGACTATAAAGATA CCAGGCGTTTCCCCCTGGAA
781 GCGCTCTCCTGTTCCGACCCTGCCGCTTACCGGATACCTGTCCGCCTTTCTCCCTTCGGG
841 CTTTCTCAATGCTCACGCTGTAGGTATCTC AGTTCGGTGTAGGTCGTTCGCTCCAAGCTG
901 ACGAACCCCCCGTTCAGCCCGACCGCTGCGCCTTATCCGGTAACTATCGTCTTGAGTCCA
961 ACACGACTTAACGGGTTGGCATGGATTGTAGGCGCCGCCCTATACCTTGTCTGCCTCCCC
1021 GCGGTGCATGGAGCCGGGCCACCTCGACCTGAATGGAAGCCGGCGGCACCTCGCTAACGG
1081 CCAAGAATTGGAGCCAATCAATTCTTGCGGAGAACTGTGAATGCGCAAACCAACCCTTGG
1141 CCATCGCGTCCGCCATCTCCAGCAGCCGCACGCGGCGCATCTCGGGCAGCGTTGGGTCCT
1201 GCGCATGATCGTGCTAGCCTGTCGTTGAGGACCCGGCTAGGCTGGCGGGGTTGCCTT
1281 AGAATGAATCACCGATACGCGAGCGAACGTGAAGCGACTG CTGCTGCAAAACGTCTGCGA
1341 AACATGAATGGTCTTCGGTTTCCGTGTTTC GTAAAGTCTGGAAACGCGGAAGTCAGCGCC
"And here is the revised DNA strand, repaired by the computer. The operation you've witnessed would have taken months in a conventional lab, but we can do it in seconds."
"Then are you working with the entire DNA strand?" Grant asked.
"Oh no," Wu said. "That's impossible. We've come a long way from the sixties, when it took a whole laboratory four years to decode a screen like this. Now the computers can do it in a couple of hours. But, even so, the DNA molecule is too big. We look only at the sections of the strand that differ from animal to animal, or from contemporary DNA. Only a few percent of the nucleotides differ from one species to the next. That's what we analyze, and it's still a big job."
Dennis Nedry yawned. He'd long ago concluded that InGen must be doing something like this. A couple of years earlier, when InGen had hired Nedry to design the park control systems, one of the initial design parameters called for data records with 3 X 109 fields. Nedry just assumed that was a mistake, and had called Palo Alto to verify it. But they had told him the Spec was correct. Three billion fields.
Nedry had worked on a lot of large systems. He'd made a name for himself setting up worldwide telephone communications for multinational corporations. Often those systems had millions of records. He was used to that. But InGen wanted something so much larger…
Puzzled, Nedry had gone to see Barney Fellows over at Symbolics, near the M.I.T. campus in Cambridge. "What kind of a database has three billion records, Barney?"
"A mistake," Barney said, laughing. "They put in an extra zero or two."
"It's not a mistake. I checked. It's what they want."
"But that's crazy," Barney said. "It's not workable. Even if you had the fastest processors and blindingly fast algorithms, a search would still take days. Maybe weeks."
"Yeah," Nedry said. "I know. Fortunately I'm not being asked to do algorithms. I'm just being asked to reserve storage and memory for the overall system. But still… what could the database be for?"
Barney frowned. "You operating under an ND?"
"Yes," Nedry said. Most of his jobs required nondisclosure agreements.
"Can you tell me anything?"
"It's a bioengineering firm."
"Bioengineering," Barney said. "Well, there's the obvious…"
"Which is?"
"A DNA molecule."
"Oh, come on," Nedry said. "Nobody could be analyzing a DNA molecule." He knew biologists were talking about the Human Genome Project, to analyze a complete human DNA strand. But that would take ten years of coordinated effort, involving laboratories around the world. It was an enormous undertaking, as big as the Manhattan Project, which made the atomic bomb. "This is a private company," Nedry said.
"With three billion records," Barney said. "I don't know what else it could be. Maybe they're being optimistic designing their system."
"Very optimistic," Nedry said.
"Or maybe they're just analyzing DNA fragments, but they've got RAM-intensive algorithms."
That made more sense. Certain database search techniques ate up a lot of memory.
"You know who did their algorithms?"