When it comes to sequencing the human genome, “complete” has always been a relative term. The first one, deciphered 20 years ago, included most of the regions that code for proteins but left about 200 million bases of DNA—8% of the human genome—untouched. Even as additional genomes were “finished,” some stretches remained out of reach, because repetitive segments of DNA confounded the sequencing technologies of the time. Now, an international grassroots effort has sorted out those hard-to-read bases, producing the most complete human genome yet.
In six papers in Science, the Telomere-to-Telomere (T2T) Consortium—named for the chromosomes’ end caps—fills in all but five of the hundreds of remaining problem spots, leaving just 10 million bases and the Y chromosome only roughly known. And today, the T2T consortium announced in a tweet it had deposited a correct sequence assembly of the missing Y.
“I don’t think we could have imagined this even 5 years ago, certainly not 10 years ago,” says bioinformaticist Ewan Birney, deputy director of the European Molecular Biology Laboratory and part of the original Human Genome Project “It’s a tour de force.” T2T researchers say the newly sequenced stretches reveal hotspots for gene evolution and underscore the chaotic history of the human genome. It “really gives us some insight into regions of the genome that have been invisible,” says Deanna Church, a genomicist at Inscripta, a gene-editing company.
The previously indecipherable sequences of the genome that have now come into clear view include the protective telomeres and the dense knobs called centromeres, which typically reside in the middle of each chromosome and help orchestrate its replication. Also almost completely revealed are the short arms of the five chromosomes where centromeres are skewed toward one end. Those short arms were known to contain scores of genes coding for the backbone of ribosomes, the cell’s protein factories.
When Birney, Church, and their colleagues introduced that first draft of a human genome in 2001, and even after they “completed” and published it in 2004, sequencer machines and genome assembly software could not wade through areas where the DNA sequence contained very repetitive stretches of bases: The repeats could too easily be skipped or their bases linked together incorrectly. As sequencing technology got better and costs dropped, scientists reduced the number of gaps or misassembled sequences, culminating in 2017 with the release of a human genome called GRCh38. With less than 1000 gaps, it became for many the “reference” against which other human genomes are compared.
But Karen Miga and Adam Phillippy wanted to do better. Miga, a geneticist at the University of California, Santa Cruz, yearned to learn the exact sequences of the distinctive “satellite” DNA that helps form centromeres. Meanwhile, Phillippy, a bioinformatician at the National Human Genome Research Institute, was busy harnessing new sequencing technologies that could read very long stretches of DNA, reducing the need to piece together shorter sequences. After meeting at a conference, they joined forces. Then in 2019, Phillippy reported they had succeeded in sequencing the X chromosome from end to end, inspiring dozens of other researchers to join the cause. “It really took on a life of its own,” Miga says.
To simplify the task, they decided to use an anonymized cell line that was derived more than 20 years ago from an unusual growth excised from the uterus of a woman—a failed pregnancy called a mole, produced when a sperm entered an egg that lacked its own set of chromosomes. With just the sperm’s genetic material, such eggs can’t develop into an embryo, but they can still replicate, especially if the sperm delivers an X instead of Y chromosome. In a boon for the project, both members of the resulting cell line’s 23 pairs of chromosomes are identical. That “made a big difference” for eliminating gaps because sequencers didn’t have to resolve differences between the parents’ chromosomes, says Robert Waterston, a geneticist at the University of Washington, Seattle, who helped lead the Human Genome Project.
The T2T group combined sequencing technologies, including a so-called nanopore device that could read 100,000 bases at a time and another sequencer that was more accurate but only did about 10,000 bases at once. A final improvement to the latter method boosted accuracy, and together the three approaches were able to polish off all but five of the final trouble spots. “Just seeing the multiple ways they went after this [shows] these are really hard problems,” Waterston says.
The approximately 200 million bases finally in the right order and in the right place include more than 1900 genes, most of them copies of known genes. The researchers cataloged duplicated regions and mobile elements—genetic material from viruses that became incorporated into the genome. In sequencing each centromere, they learned the duplicated regions vary greatly in size, unexpected because these knobs serve the same purpose in each chromosome.
The short chromosome arms held another surprise. As expected, they included multiple copies, 400 in all, of the genes coding for the RNA that’s used to make ribosomes. “This rDNA was the last domino to fall,” as it was the hardest to sequence, Miga says.
The short arms are also “just chock-full of [other] repeats,” says Jennifer Gerton, a chromosome biologist at the Stowers Institute for Medical Research. Those include mobile elements, duplicated segments and other types of repetitive DNA, as well as many copies of genes from other parts of the genome. “It’s amazing how dynamic the human genome can be,” Church says. In five spots along these chromosomes, the resulting jumble is so long that the researchers still can’t clearly determine the order of the bases, although they have a rough idea of the sequence, Gerton says.
Short arms are likely hotspots for gene evolution, Phillippy notes, as gene copies parked there are free to mutate and take on new functions. The catalog of duplications could also shed light on neurological and developmental disorders, which have been linked to variations in the number of copies of specific sequences. Chemical modifications to the DNA in the complex repetitive areas likely play a role in disease as well, and those changes have been mapped. Because the cell line used lacked a Y chromosome, the T2T group sequenced one from a well-studied genome belonging to Harvard University systems biologist Leonid Peshkin (see sidebar, below).
Despite their latest milestone, human genome sequencers aren’t packing their bags. “There’s still some work to do,” says Human Genome Project co-leader Richard Gibbs, a geneticist at Baylor College of Medicine. He and other researchers stress that the field now needs to get similarly complete genome sequences from a greater diversity of people to look for variation in the short arms and the other tough-to-read regions, which could play a role in diseases or traits.
The T2T team has made a start by deciphering 70 more genomes, with a goal of 350 from people of diverse ancestries. These genomes, sequenced as part of the Human Pangenome Reference Consortium, are more challenging to finish because they don’t have identical pairs of chromosomes. So, for now, the team has settled for high-quality genomes that place as many of the bases as possible on their correct chromosomes. Next, the researchers plan to apply all their methods to Peshkin’s whole genome. And, eventually, Phillippy says, “We want every genome to be telomere to telomere.”
Whose DNA makes up the newly ‘complete’ human genome?
By Elizabeth Pennisi
A 51-year-old Harvard University biologist named Leonid Peshkin and an anonymous man who almost became a parent decades ago have become inextricably linked in the most complete human genome so far.
The genome’s Y chromosome came from Peshkin, and the rest of the DNA sequenced by the Telomere-to-Telomere (T2T) Consortium comes from a so-called molar pregnancy, a uterine growth that results on rare occasions when a sperm enters an egg that has no chromosomes. The fertilized cell can copy the sperm’s 23 chromosomes, creating two identical sets, and begin to replicate. As part of research into how these moles develop, Urvashi Surti, a geneticist at the University of Pittsburgh, wanted to create cell lines from such growths. With permission from her medical center’s institutional review board and removal of any information that could link these tissues to the “parents,” she succeeded in 2001, gaining access to dozens of growths excised by physicians between 1981 and 2000.
Surti and others recognized the unique chromosomal makeup of these cell lines—they only have one parent’s DNA—might make them useful for genomics studies. Within a decade, some genomic data for a cell line Surti’s team developed, CHM13 (CHM stands for complete hydatidiform mole), were in public databases, says Adam Phillippy, co-leader of the T2T Consortium and a bioinformatician at the National Human Genome Research Institute (NHGRI). So, after Tamara Potapova and Jennifer Gerton, chromosome biologists at the Stowers Institute for Medical Research, confirmed CHM13 had the right number of chromosomes and that number didn’t change through time, the consortium decided to sequence its genome.
That decision became potentially problematic in 2019, when NHGRI began to require more explicit consent from tissue and DNA donors for any “genomic data sharing.” A 2010 book about the widely studied HeLa cell line, created from tissue unknowingly provided by a woman named Henrietta Lacks, had brought to the fore the need for researchers to get adequate consent from tissue providers or their families. Today, donors are asked to permit their material to be shared broadly and to be used in future research. But no such detailed consent was requested from the woman with the molar pregnancy that led to CHM13, let alone the man whose sperm DNA actually makes up the genome.
Yet NHGRI allowed the T2T work to go forward, deciding an exception was warranted because much of CHM13’s sequence was already known and the line was so useful. The question remains open of whether the owner of CHM13’s genome could be identified using public DNA sequences in genealogy databases. Phillippy thinks not because CHM13’s genome only represents one-half of that person’s DNA. Even if it were possible, NHGRI officials argue it would be unethical to reveal him for any reason, including to get consent.
Because CHM13 has an X chromosome but no Y, the T2T Consortium added Peshkin’s DNA. He and his parents had already donated tissue for DNA research as part of the decadelong Genome in a Bottle program coordinated by the National Institute of Standards and Technology (NIST). That program provides well-validated genome sequences (and cell lines and tissue samples) for testing new technologies and for other studies involving DNA.
A few months ago, Peshkin had called NIST to propose that the genome and cell lines he and his father had provided be used for studies of aging. While on the phone, the agency told him that the T2T group was comprehensively sequencing Peshkin’s X and Y chromosome—his original consent form allows for broad-scale use of his DNA—and planned that his genome would eventually be the first full human genome to have the T2T treatment. “I’m excited to be part of this front-line, leading edge in science,” Peshkin says.