Pakistan genome study finds 34,000 ‘human knockouts’ that could transform medicine

A gene can vanish, and life can go on.

That is one of the most striking lessons from a sweeping new analysis of 173,303 genomes from Pakistan. This project is reshaping how researchers think about disease, drug safety, and the limits of animal research. The study found that about one in five people in the database carried a complete loss of function in at least one gene. This creates a rare window into what happens when human biology runs without parts long assumed to matter.

The scale alone is remarkable. Researchers identified 34,364 people carrying what the authors call homozygous loss-of-function variants. These are naturally occurring cases in which both copies of a gene are disrupted. Across the full resource, the total added up to 6,476 genes. That is about one-third of all protein-coding genes in the human genome.

For drug developers, the message is hard to miss. Human genetics has become one of the strongest guides to successful medicine. However, many early ideas still come from experiments in mice. This study shows, in case after case, that the human body does not always follow the same script.

Danish Saleheen, professor of medical sciences and director of global genomics at Columbia University Vagelos College of Physicians and Surgeons, said the better path is to study people who are born without working versions of a gene. Then, see what happens to their health.

That is especially important because South Asians remain deeply underrepresented in global genetic databases, despite making up roughly a quarter of the world’s population. In the new resource, nearly half of the coding variants identified had not been seen in existing databases.

The Pakistan Genome Resource was built from whole exomes or genomes of people recruited in 23 cities across the country. It expands earlier work by about 15-fold. In addition, it combines genetic data with medical histories, clinical measurements, and disease information for most participants.

Its unusual power comes partly from population history. Many participants came from families with close parental relatedness. This increases the chance that a child inherits the same disrupted gene copy from both parents. In the study, 30.6% of participants reported that their parents were first cousins. That pattern of shared inheritance helps reveal complete gene loss that would be much rarer in many other large databases.

The payoff is efficiency. In the Pakistan resource, at least one gene with complete loss of function appeared in 20% of sequenced participants. In non-Finnish Europeans in gnomAD, one comparable event appeared per 14.1 individuals. Within this Pakistani cohort the rate was about one per four sequenced participants.

That made the resource unusually good at uncovering genes that can be fully knocked out in humans. Of the 6,476 genes observed with complete loss of function, 3,323 were unique relative to non-South Asian individuals in gnomAD. Furthermore, 2,298 were unique even when all gnomAD populations were included.

Some of the most important findings involve genes already tied to drug development.

One example is RXFP1, a gene that has drawn interest in areas including heart failure. In mice, RXFP1 and its ligand relaxin have been linked to reproduction, fibrosis, and cardiovascular problems. That biology helped drive clinical trials of RXFP1-targeting drugs, but those trials failed to meet their primary end points.

The Pakistan data offer one reason why. Researchers identified multiple people with complete loss of RXFP1 function and found no consistent pattern of cardiovascular, kidney, or reproductive problems. The result suggests that human RXFP1 biology differs in important ways from rodent biology. This helps explain why drugs built on mouse findings did not translate well.

The same database also pointed in the opposite direction for CIDEB, a gene of growing interest in liver disease. Fourteen people with complete loss of CIDEB function had no reported history of liver disease. In burden analyses, loss of CIDEB was associated with lower alanine aminotransferase, lower aspartate aminotransferase, and lower rates of non-alcoholic fatty liver disease. Those findings support the idea that blocking CIDEB could be protective. In addition, it may be safe over time.

Another caution flag appeared around LRRK2, a major Parkinson’s disease target. Drugs that inhibit LRRK2 are already in development, so long-term safety matters. In the Pakistan resource, several individuals with complete loss of LRRK2 function showed kidney problems. These included reduced estimated glomerular filtration rate and chronic kidney disease.

The sample is small, and the authors are careful not to overstate it. Even so, the signal lines up with adverse kidney findings already seen in rodent and preclinical studies. Their conclusion is direct: kidney function should be monitored during long-term LRRK2 inhibition.

The resource is valuable not just because it is large, but because researchers can bring people back for follow-up studies and examine relatives who carry zero, one, or two disrupted copies of the same gene. That makes it possible to compare family members who often share diet, environment, and living conditions. As a result, scientists get a cleaner look at what a genotype is actually doing.

That approach helped clarify the sensory gene TRPM8, which is activated by cold. In recall-by-genotype testing, people with TRPM8 loss-of-function variants showed delayed cold-induced pain and greater tolerance of cold-related pain. The study notes that genome-wide association studies have linked TRPM8 to migraine risk. These findings suggest complete loss of the gene is tolerated, supporting the idea that inhibiting it for migraine may be safe.

The study also revisited genes long treated as biologically essential. PRDM9 is required for fertility in knockout mice, yet the researchers found three women and one man with complete loss of PRDM9 function who had children. That is a sharp reminder that even celebrated findings in model organisms may not carry over neatly to humans.

More broadly, the team found strong depletion of complete gene loss in pathways tied to core cellular work, including protein secretion, oxidative phosphorylation, unfolded protein response, DNA repair, and developmental signaling pathways such as TGFβ. In other words, some genes do appear essential in humans. However, many others are more flexible than expected.

That matters because roughly two-thirds of human genes still remain poorly understood, even decades after the Human Genome Project. Studies like this can begin to assign function by showing what happens when those genes are naturally switched off.

This study gives drug makers a more human way to test whether a target is promising, dangerous, or both. It suggests some therapies may have been pursued too confidently on the basis of mouse biology, while others deserve more attention because people already live safely without the gene in question.

It also strengthens the case for expanding genomic research beyond heavily European databases. By sampling a historically underrepresented population at scale, the Pakistan Genome Resource uncovered variants, disease clues, and drug-relevant biology that would have been easy to miss elsewhere.

For medicine, that could mean fewer expensive failures, earlier warnings about side effects, and better evidence about which targets are worth chasing. For genetics, it opens a path toward answering one of the field’s oldest unfinished questions: what all those human genes actually do.

Research findings are available online in the journal Nature.

The original story "Pakistan genome study finds 34,000 'human knockouts' that could transform medicine" is published in The Brighter Side of News.

Like these kind of feel good stories? Get The Brighter Side of News' newsletter.