A Flood in Genomics

Nine months have passed since draft sequences of the human genome were first published.1,2 One human gestation period later, the genome, as deciphered by the International Human Genome Sequencing Consortium, still screams toward its projected Spring 2003 finish date. "The trajectory we're on for meeting that goal is precisely on target," assures Francis Collins, director, National Human Genome Research Institute (NHGRI) and spokesperson for the largest public biological science project in history. According to Collins, 55 percent of the genome is in completed form and 42 percent is in draft form. Chromo- somes 20 and Y are finished and being written up, with 6, 7, and 14 close behind. Each is sure to offer new insights to human health and life itself. But as the torrents of genomic data continue to pour in from the human genome and a myriad of other organisms, scientists are forced to valuate the data before them. As a tool for basic researchers, the human genome has been well used. Both publications on the draft genome have been cited in more than 350 papers, according to the Web of Science, ISI, Philadelphia. Yet, financial onlookers say that pharmaceutical companies may be choked by the wealth of data. And still, the flood continues as projects in proteomics and transcriptomics open up new holes in the dam. "In the rush toward other -omics," Richard Gibbs, director, Baylor College of Medicine Human Genome Sequencing Center says, "others, I think, have lost track of the real power of having the ability to decipher primary genomic information." The Business End Courtesy of Philip Ma Philip Ma Whether the availability of the human genome begets feast or famine for the biotech and pharmaceutical industries is not completely clear. Philip Ma, leader of North American pharma/medical products at McKinsey & Company, says, "While I think the definite successes of the genomics revolution has been in keying up potential targets, the industry in some sense has more targets than it knows what to do with." The foreseen possibilities in unearthing new drug targets fueled much of the hype surrounding the human genome, and that hype from the media fueled a funding boom in biotech. Glenn Giovanetti at Ernst & Young Life Sciences Industry Services, comments "You could really compare [today's situation] to a large degree with the first biotech boom in the late eighties and early nineties where the thought was, 'Hey, this is going to lead to better drugs faster,' and clearly that hasn't been the case." Having the genome in hand has brought about more drug targets, but, explains Ma, "People are getting more concerned that novel targets are going to have a higher rate of failures because there is less information on them." And when working in 10-year drug-development cycles, failures are costly. Ma points to a trend of growth in clinical informatics that would effectively garner more information from expensive clinical trials instead of simply treating them as regulatory hurdles. "People are beginning to think through to how ... to take greater advantage of that information," he adds. But increasingly, the suppliers of genomic information have been looking to do the same thing. A number of companies who formerly focused on providing and interpreting genomic information are now expanding to include drug discovery, according to Giovanetti. He lists Millennium Pharmaceuticals, Incyte Genomics, and Hyseq as a few who have begun to shift toward product end. "I think there's an evolving strategy to say that drug industry profits and margins are pretty compelling," Giovanetti says. But even for these information-generating companies the drug target bottleneck will have to be overcome. One way through that bottleneck is the use of in silico experiments that drive annotation and are amalgamating the flood of biological data into something usable. Annotation Nation Further work on the human genome has, as many expected, revealed that that the gene count may be higher than the initial 30,000 to 40,000,3 but exactly how much higher remains a point of contention. "I think people have been arguing about it for a long time, but it's sort of like medieval scholars arguing about how many angels can dance on the head of a pin," quips John Quackenbush, associate investigator at The Institute for Genomic Research (TIGR), Rockville, Md. Gene counts, and even the genes themselves, are destined to change as the human sequence nears completion. Genes currently straddling less-than-complete or erroneous areas will change, move, and realign as new versions of the human genome are released. Terry Gaasterland, assistant professor of computational genomics at Rockefeller University notes, "Tracking that history is important to the experimental biologist who may know that a particular locus of genome sequence has been associated with a particular disease. That kind of alerting and triggering and tracking doesn't exist right now." Two projects she lauds for striving to overcome these problems are the joint Sanger Centre and European Bioinformatics Institute effort, called Ensembl, and its US collaborative counterpart, Genome Browser, housed at University of California, Santa Cruz (UCSC). Genome Browser, the brainchild of graduate student James Kent, is a publicly accessible "grass-roots" annotation database that draws information from multiple sources and presents it very clearly. "We tend to show each piece of data with all its warts. You want to show a synthetic, integrated view of the data to see what the conclusions are, on the other hand you don't want to hide the raw data that underlie the reasoning," says David Haussler, Howard Hughes Medical Institute investigator and manager of Genome Browser. Scientists say that layering the annotation data with data from other sources such as other organismal genomes, cDNA libraries, expressed sequence tag (EST) data, or functional genomics information, is the only way to truly annotate every gene. "Prediction de novo is not possible," says Haussler, "There's not enough information in the raw sequence. We're not pulling all of that evidence together into a cohesive story. That's something we have to try to do-to make it more synthetic, more integrated. It approaches automated discovery." New Technologies Improvements in genomics drive down the cost twofold every 18 months, effectively paralleling Moore's law of computing, according to George Church, director of the Lipper Center for Computational Genetics at Harvard Medical School. The trends complement each other. While the digital gurus work out the new technologies and algorithms that make genomic data manageable, life scientists find new ways to flood the pool of genomic data faster than ever. "A number of times it appears that it will be necessary to use supercomputing for genomics. But by the time it seems to be articulated that you need a certain computational power, the general-purpose computers have advanced to nearly meet the challenges," Church says. The trend was anticipated and counted on by genome pioneers. Explains Department of Energy director, Ari Patrinos, "For us in the trenches it was a nail-biting experience, ... because it was never a guarantee that the technology you count on for that last-minute rush would actually materialize." Inevitably it is the trench researchers that dig up the tools necessary to quicken the pace. Gibbs of Baylor College of Medicine reminisces about the strange sense of surprise he felt when postdoc Wei-Wen Cai walked into his office and said, "'Hey, I've got an idea that's going to be vastly superior to whole genome shotgun sequencing and will revolutionize the completion of mammalian genomes.' He explained the idea, and as soon as he spilled it out, I sort of banged my head like Homer Simpson and said, 'Yeah, this is really right.'" The idea, Clone-Array Pooled Shotgun Sequencing (CAPSS), simplifies tracking tens of thousands of BAC libraries by pooling them together across a grid. More than 25,000 BACs in a 160 x 160-well plate could be pooled into 320 wells. The pooled BACs can be sequenced and the individual BACs can be reassembled and assigned to their original coordinates using less computational power than assembly via the whole genome shotgun approach.4 Gibbs reports that results from 6 x 6 and 10 x 10 clone arrays will be available for public discussion in December. Another recent genomics technique prompted the contribution of 700,000 open reading frame expressed sequence tags (ORESTES). Similar to ESTs, they may aid in the identification of human genes as well as shotgun sequencing of the human transcriptome. ORESTES are generally taken from the middle of genes using low stringency rtPCR and arbitrary primers.5 It shares a mother with most invention. "Basically it was born of necessity," explains Andrew Simpson, head of the laboratory or cancer genetics, Ludwig Institute for Cancer Research, Brazil. While working in a rudimentarily equipped lab, Simpson and his colleagues were trying to analyze expressed genes in a parasite. "We felt we didn't have the conditions to make good full-length cDNA libraries," Simpson explains. So, they relied on the low stringency techniques and later found that they had a number of advantages. Because arbitrary primers were used, they annealed to sequences in the middle of mRNA transcripts instead of from the ends as traditional EST techniques produce. The short random primers are biased toward the middle of long mRNA sequences, and when multiple copies of certain ESTs are produced they begin reannealing with themselves, thereby allowing for amplification of more poorly expressed sequences. The dataset achieved thus far may represent 60 percent of all human genes. Combined with EST techniques, says Simpson, "you can start to identify genes with some considerable precision." Patrinos says he expects new technologies in sequencing to emerge from nanotechnology. "I'm confident that we will revisit single molecule detection and sequencing, which will simplify the sequencing task and will get to the difficult-to-sequence regions, and also will ultimately make sequencing very, very cheap." Currently, a draft of a mammalian genome runs $100 million, and the push will ever be to get those costs to the point at which they are tractable and economic. Some still look forward to the day that an individual could walk into a clinic and be fully sequenced in a matter of days. That would mean an unimaginable amount of data to keep track of. Advances in genomics and all the other "-omics" will mean advances in many realms of human understanding, but first it will require fundamental advances in the ways scientists can process and use the information. Quackenbush of TIGR foretells, "What we're going to see more of in the future is approaches to integrate all of the data we have, and I think that's the coming challenge. ... Because right now, what we have are the snapshots, and the role that the genome plays-more than anything else-is providing a scaffold and a logical framework in which to do that." Brendan A. Maher can be contacted at bmaher@the-scientist.com References 1. E.S. Lander et al. "Initial sequencing and analysis of the human genome," Nature, 409:860-921, 2001. 2. J.C. Venter et al. "The Sequence of the Human Genome," Science 291:1304-51, 2001. 3. T. Hollon, "Human genes: How many?" The Scientist, 15[20]:1, Oct. 15, 2001. 4. W. Cai et al., "A clone-array pooled shotgun strategy for sequencing large genomes," Genome Research, 11:1619-23, 2001. 5. A.A. Camargo et al. "The contribution of 700,000 ORF sequence tags to the definition of the human transcriptome," Proceedings of the National Academy of Sciences, 98:12103-8, Oct. 9, 2001. Looking Beyond the Human Genome To get a full picture of the story of life, human inner workings, and the human place in evolutionary history, comparative genomics is being touted as the imperative route. And that means creating a library of genomes through which scientists can decipher all life's intricacies and differences. "The human sequence is not enough for complete interpretation of the human genome," says Eric Green, director, National Institutes of Health Intramural Sequencing Center. The continually growing sequencing infrastructure established by the International Human Genome Consortium is being freed up to churn out genomes of other organisms. Anticipating a rush by scientists to get their particular organisms of interest in the queue, the National Human Genome Research Institute set up a peer reviewed screening process that will select organisms for sequencing.1 Applications will be accepted beginning in February and scientists are already gearing up to justify the importance of their favorite species. Among the many possibilities is the call for sequencing primate genomes, which would "tremendously accelerate the progress in understanding risk factors for the disease," according to Jeffrey Rogers, group leader, genetics group, Southwest Regional Primate Research Center. "I'd like to see several primates sequenced," he says. But Rogers realizes, "It's a matter of priorities in a world of finite resources." The cost of sequencing genomes as large and complex as human or primate is prohibitive. For that reason, Rogers and a committee of supporters rally behind the rhesus as the first primate to be sequenced. Old-world monkeys, such as the rhesus and macaque, are standards in human health research, but they do not have the chimpanzee's 99-percent human homology that evolutionary genomics scientists salivate over. "My position would be that the chimpanzee is really a special case," says Ajit Varki, director, glycobiology research and training center, University of California, San Diego. "The extreme similarities to humans would make any differences of great potential value," Varki explains, adding that the chimp's differences in severity and frequency for contracting such maladies as AIDS, Alzheimer's, cancer, and malaria might make that 1-percent divergence of utmost importance to human health research. Though Rogers agrees the chimp genome would be of great value, he cites the moratorium on breeding, the cost of use in research, and the ethical considerations surrounding terminal experiments with chimpanzees as reasons for bumping the chimp genome down on the list of priorities. The Silver Project in Japan, however, has already begun sequencing the genomes of chimpanzee and other apes for comparative studies. -Brendan A. Maher 1. B.A. Maher, "NHGRI sets BAC priorities," The Scientist, 15[22]:20 Nov. 12, 2001.

A Flood in Genomics

Interested in reading more?

Become a Member of