Human Genome Project1 By Jérémie Gallien2 and Scott A. Rosenberg3 Scott was now a couple of weeks into his MIT Leaders for Manufacturing program internship at the Whitehead Institute in Cambridge. The exciting premise for his project was that the genome finishing group at work there could benefit from a sound flow analysis of the type usually applied in manufacturing environments. While he had received great support from his supervisors all along and felt that he had already acquired a reasonably good handle of the work in his area of scope, he also knew that he still had to prove his worth: in an environment with scores of PhDs, many world-renown scientists and even Nobel prize winners walking through now and then, nobody was going to settle for small talk and unsupported recommendations. Indeed, the sequence finishing operation seemed considerably more repetitive and process-oriented than anything else at the Whitehead, an institution known for breaking new grounds in Biology through scientific experiments never attempted before. Because of the need to accurately forecast the final completion date of the genome and to plan for staffing levels, the largest complaint of the scientists overseeing the finishing group was by far the variability of weekly output. Scott suspected that the practice of bundling many tasks into a single project assigned to each finisher and the informal, on-demand policy followed when assigning new projects played no small role in this variability. What he did not know yet however was how to quantify these effects so he could convince the Whitehead managers to change their procedures. The Human Genome Project Arguably the most important undertaking in life sciences since the discovery of DNA in 1953, the Human Genome Project (HGP) began in 1990. Its simple but ambitious goal is to sequence the entire genetic makeup of the human species, which will enable decades of evolutionary and medical research on cross-genomic comparison, disease risk detection, gene therapy and possibly uncountable other applications not even imagined yet by scientists. Primary responsibility for the sequencing fell to large genome centers like the Whitehead Institute at MIT, Washington University, Baylor University, and the Sanger Center in Great Britain, with dozens of smaller centers around the world also contributing. 1 This version: July 2003. 2 Corresponding author: MIT Sloan School of Management, Cambridge, MA 021423 Most material in this case originates from Scott Rosenberg’s Leaders for Manufacturing Master’s thesis, “Managing a Data Analysis Production Line: An Example from the Whitehead/MIT Center for Genomic Research,” Sloan School of Management and Dept. of EECS, Massachusetts Institute of Technology, June 2003. Copyright 2003 © Jérémie Gallien and Scott Rosenberg 1In 2000, a draft sequence of the human genome was published. However, it contained many absent, ambiguous, or conflicting regions of DNA. In the time since the draft’s publication, genome centers like the Whitehead have concentrated their energies on systematically clarifying these problematic regions. This process is called finishing, and is both the current bottleneck of the genome project and the focus of Scott’s work. DNA Sequencing Background Deoxyribonucleic acid (DNA) is the genetic building block upon which all known life regulates its daily function and long-term evolution. Constituting the chromosomes found in the nucleus of human cells, DNA is itself comprised of long strings of just four nucleotide bases called adenine (A), guanine (G), cytosine (C), and thymine (T). Active sequences of DNA that are hundreds or thousands of base pairs long, called genes, are translated into proteins during the course of cell activity. Proteins, in turn, enable all of life’s most basic functions. Structurally speaking, DNA is a stable polymer that arranges itself into a double helical structure as shown in Figure 1. Long sequences of nucleotide bases form one half of the structure. Each base also bonds to its complementary base in the other half of the structure: A pairs with T and G pairs with C. Thus, a sequence of “ATTGC” bonds to its complementary sequence “TAACG”. All told, the human genome consists of more than three billion DNA base pairs and an estimated 30,000 genes. Figure 1. Relationship between cells, chromosomes, DNA, and proteins.4 Today’s state-of-the-art gene sequencing technology proceeds by breaking large DNA samples into small segments, determining the exact DNA sequence of those small segments, then reconstructing sequence from these segments into a composite view of 4 Source: U.S. Department of Energy, http://www.ornl.gov/TechResources/Human_Genome/publicat/primer2001/1.html. Copyright 2003 © Jérémie Gallien and Scott Rosenberg 2the original sample: DNA donated by a small set of consenting, anonymous individuals is first purified, then enzymes are used to break it down into smaller segments; From the mix that results, segments with a length of approximately 165,000 base pairs (165kbp), called BAC templates, are isolated. An engineered version of the bacteria E. coli can then be tricked into carrying and reproducing this human genetic material millions of times in just hours. To accomplish a further reduction in sample size necessary to direct sequencing, BACs are then sheared through a physical process and filtered, producing DNA segments of uniform size, usually between 4kbp and 10kbp. Once isolated, each such segment becomes known as a plasmid (see Figure 2). Figure 2. Two phase break-down of genome into BACs and then plasmids. Genome, ~ 3 bbp BAC templates, ~ 165 kbp Plasmids, ~ 4 kbp not drawn to scale After amplification through E. coli, plasmids are placed in a solution containing special DNA base pairs that are tagged with a fluorescent dye. By raising the temperature of the solution, the plasmid DNA, which normally resides in a paired helical structure, can be induced to separate. When the temperature is lowered, an enzyme in the solution reconstructs the helical structure by grabbing base pairs from the surrounding solution. Whenever the enzyme selects a dyed base pair,