On the 24th of July, 2014 PLOS Genetics published an article entitled: “8.2% of the Human Genome Is Constrained: Variation in Rates of Turnover across Functional Element Classes in the Human Lineage”. Two weeks later the article had been viewed 12,307 times, it already had one citation, and had been shared on social media 436 times. This means that, on average, 36 people had been viewing the article online every hour.
Unlike many other papers that might attract this level of interest through traditional press coverage, this article was primarily picked up by social media. This flurry of activity was in part driven by coverage in ifls, which has been shared on Facebook 12,200 times. However, it wasn’t just a social media piece, it has also been featured in Nature, Sci-news, the Guardian, the French Tribune, and the International Business Times.
“It is pleasing to see that the article has received lots of views of course. I think article metrics like views and shares on social media (such as twitter and facebook) are important metrics because they are nearly instantaneous measures of impact (compared to citations) and they imply there is a public interest in the research. I believe science communication is important, particularly for publicly funded research such as this. The open access publishing at PLOS helps promote science communication.” – Chris Rands (author)
What made this article such a hit? The paper explores how much of the human genome can be considered ‘functional’. While scientists were able to successfully sequence the human genome over a decade ago, it has been unclear how much of the genome is doing something useful. In 2012 some members of the ENCODE consortium suggested that as much as 80% of the genome might be considered functional. Chris Rands and colleagues argue that this might be as little as 8.2%; ten times less than the ENCODE estimate.
The researchers used a computational approach to compare the whole DNA sequences of various mammals, including human, horse, bush baby, panda, and mouse. Image Credit: George Lu CC BY Flickr
Natural selection is expected to preserve important DNA, removing insertion and deletion (indel) mutations from these genomic regions. In order to find the sequences with evidence of functionality, Chris Rands and colleagues identified areas of the human genome that have been unusually evolutionarily stable with respect to these indel mutations. The researchers used a computational approach to compare the complete DNA sequences of various mammals, including human, horse, bush baby, panda, and mouse.
Using this method they estimate that between 7.1-9.2% of the human genome can be considered ‘functional’. If this is the case, then the majority of the human genome can experience mutations without affecting fitness. Coding exons, bound motifs and DNase1 footprints make up around 9% of the human genome. These are areas that we would expect to be considered functional, based on what we know about their roles within the genome. While these areas will not encompass all of the functional sections and there may be sequences outside of these sections that are functional, it could be taken as an indication that the estimate that Chris Rands and colleagues suggest is in the right ballpark.
Mouse as a model organism may miss many subtle aspects of human noncoding biology. Image Credit: Duncan Hull CC BY Flickr
As well as an indication of DNA functionality Chris Rands tells PLOS Genetics that “We find that only 2.2% the human genome is ‘functional’ and shared with mouse too; the remaining 6% has gained ‘functionality’ in the human lineage since humans diverged from their common ancestor with mice 80 million years ago.”
Their findings could therefore provide a more quantitative basis for assessing the relevance of model organisms to specific questions of human biology.
“Although the main media message about the paper has been about the proportion of the human genome that is ‘functional’, the main novel research contribution is about the turnover of ‘functional’ sequence as it is lost and gained over time. Given we estimate that approximately 3 times as much sequence is ‘functional’ in the human genome as is ‘functional’ and shared with mouse, this implies that mouse as a model organism may miss many subtle aspects of human noncoding biology, although the real practical implications of this tentative conclusion remain to be fully explored.”