I have slipped in my resolution for a weekly round-up. Here is another catch-up. And there has certainly been a lot to catch up on!

Germline news

In the first ever randomized controlled trial of whole genome sequencing for healthy adults, which reported on variation across genes linked to Mendelian disease, many participants were found to have variants believed to definitively cause disease by adulthood, even though they were living disease free. STAT has a write-up.

At one point, it was thought that diseases such as obesity, or traits such as height might be polygenic in nature, i.e. affected by variation in some set of closely involved genes. Genome Wide Association Studies are premised on this idea. But it is challenged by a recent paper that puts forward the case for the omnigenic model — i.e. almost every gene effects every trait, because of how interconnected the networks are. If true, this suggests that we should be concentrating on mapping out the networks operating in different cell types. Ed Yong has a good write-up.

Veritas has launched myBabyGenome in China. For $1500 parents can find out about a series of traits (e.g. how “novelty seeking” their child may be) and disease risks. But as genetics prof Jim Evans says “You run the risk of predestination based on bad science”. They won’t be told about e.g. variation that leads to an increased risk if Alzheimers. Instead, they will be offered the chance to purchase that information later on.

A study of families in Scotland involving 20,000 individuals reports that brainier people have fewer variants that impair general health and intelligence, rather than possessing variants that make them smarter. While twin studies suggest that 50-80% of intelligence is based on genetics, studies to date had been able to attribute only about 30% of the variance to genetics — the gap is known as the “mystery of missing heritability”. This study was able to explain this gap, because it could isolate the effect of very rare variants — usually too rare to be missed, but because of their presence in families, detectable by this study. The results are support for the “mutational load” paradigm for thinking of genetic endowment.

Huntington’s sufferers have one normal and one altered copy of the gene HTT. A study that used CRISPR-Cas9 to knock out both copies of HTT in mice found no negative effects and improvements in symptoms.

Over 1000 new reference genomes of bacteria and archaea have been deposited by the JGI.

New York fertility doctor, Dr Zhang, who evaded US regulations and went to Mexico to oversee use of the three-person baby technique, has started a company to market this technique. It will market not only to couples wishing to avoid passing on mitochondrial disease, but also for age-related issues.

A study that looked at the combination of pre-implantation genetic diagnosis (PGD, performed for Mendelian diseases such as Cystic Fibrosis) and screening (PGS, performed to test for aneuploidies such as Downs), found that the combination resulted in more pregnancies given embryo transfers, but fewer transfers, as for more couples no viable embryos remained. After PGD, 56% of embryos were deemed transferable. After PGS, only 27.5% of blastocysts were deemed transferable. Standardly, about 34% of transfers result in pregnancies. Following both PGD and PGS in this study, the number was 49%.

How much of heritability comes from common (>5% frequency) variants? A new study suggests about 43%.

Why hasn’t natural selection reduced the amount of coronary heart disease? “This study provides novel evidence that CAD has been maintained in modern humans as a by-product of the fitness advantages those genes provide early in human lifecycles” — paper here. 

Analysis of mummy DNA suggests Ancient Egyptians individuals were more similar to present day people from the Near East than to populations residing in Egypt today

The Precision Medicine Initiative, launched in 2015, has just started beta enrollment in its All Of Us program, which aims to be a 1 million strong cohort. 

Age-related somatic mutations have typically been of interest because of cancer. But they are also of potential relevance for other disease types. A large-scale study finds that a particular type of somatic change in blood cells doubles the risk of coronary heart disease. Which is a good segway into…

 

Somatic related news

Encouraging results from a small cancer-vaccine study: 6 melanoma patients given a cancer vacine unique to their tumor saw no recurrence. The technique works by first sequencing the tumor, to work out which sections of the DNA make antigens that the immune system is most likely to respond to, and that are different from germline cells. Next, millions of copies of each of 13-20 of these neoantigen regions are injected.  The idea is that the patient’s immune system learns to recognize these sequences as “other”, and will then target the cancerous cells.

Evidence that PARP inhibitors perform better than chemotherapy for women with metastatic breast cancer who also have germline BRCA mutations. There are three PARP inhibitors approved for ovarian cancer on the market.

A study of a regulatory region that is well conserved between humans and mice finds when this region is deleted a) mice develop normally, and b) mice are resistant to tumor development, making it a promising drug target. 

Following Keytruda’s approval based on a bimarker indication across all cancer types, new FDA commissioner Gottleib has promised more guidance to allow more drugs to follow this path. 

Flatiron and Foundation have shown the utility of their joint molecular-clinical dataset, including the demonstration that those with an actionable variant have a survival time of 35 months, compared to 19 months for those without.

 A nice synopsis of oncologists’ thinking about genetic testing, arising out of ASCO.

Merck paused two studies of Keytruda in combination with other therapies for multiple myeloma, because of unexplained deaths.

Accurate software for distinguishing germline variants from a tumor without the need for a matched normal sample.