The shrinking human genome

Our understanding of the human genome has been reduced to 19,000 genes, almost all of which have ancestors prior to the appearance of primates 50 million years ago. A Spanish study has updated the number of human genes – those that code for proteins – to 19,000, well below the initial estimation of 100,000, and 1,700 below the most recent annotation. The work, published in Human Molecular Genetics, suggests that almost all of these genes have ancestors prior to the appearance of our ancestors. Researchers at the Spanish National Cancer Research Centre (CNIO) analysed proteomics experiments to determine a map of human proteins, reviewing seven large-scale mass spectrometry studies involving more than 50 human tissues to verify which genes really do product proteins. They identified 2,001 non-coding genes, based on weak conservation, lack of protein features or ambiguous annotations. Study leader Alfonso Valencia, Vice-Director of Basic Research at CNIO describes this continuous correction to the number of protein-coding genes as ‘the shrinking human genome’. “The coding part of the genome (which produces proteins) is constantly moving,” he said. “No one could have imagined a few years ago that such a small number of genes could make something so complex.” Researchers analysed thousands of genes annotated in the human genome but did not appear in the proteomics analysis. “1,700 of the genes that are supposed to produce proteins almost certainly do not for various reasons, either because they do not exhibit any coding features, or because the conservation of their reading frames does not support protein coding ability,” said Michael Tress. The study proposes that more than 90% of human genes producing proteins derived from the metazoans or multicellular organisms of the animal kingdom hundreds of millions of years ago; over 99% for those genes whose origin predates the emergence of primates. This work brings the number of human genes closer to that of other species such as the nematode worm, but Valencia prefers not to make comparisons: “The human genome is the best annotated, but we still believe that 1,700 genes may have to be re-annotated. Our work suggests that we will have to redo the calculations for all genomes, not only the human genome.” Multiple evidence strands suggest that there may be as few as 19,000 human protein-coding genes