Round-up Feb 6th – Feb 12th

When is a variant too common to be pathogenic? >5% is a hard cut-off according to the ACMG variant interpretation guidelines, but a frequency “greater then expected for the disorder” is also strong evidence for a benign interpretation, and a rate of 0,01% is suggested by this recent paper, provided that known founder mutations are taken into account.

This paper reports on core promoter strength across the human genome.

Most association work between genotype and phenotype is done on a variant by variant level. This paper reports on a haplotype based correlation between essential hypertension and a particular haplotype occurring with a particular haplotype.

DNA methylation patterns from single cells is here.

study that reiterates the importance of understanding the dosage pathogenicity of genes in interpreting the consequences of copy number variants.

Genomic medicine is constantly evolving. Who has the responsibility to ensure patients who have had genetic testing are kept up to date? A UK group recommends that re-contacting preferences be built into the initial consult.

The genetic underpinnings of the third largest cause of death in the US, ,Chronic obstructive pulmonary disease (COPD), are being uncovered.

When it comes to BRCA testing, a large majority of women who are at risk are not being recommended testing by their doctors.

How to separate the effects of genotype and phenotype? Subjecting those with identical genotype (identical twins) to very different environments (one in space, one not) can yield interesting results, and not just for those planning colonialization of Mars. The telomeres of the DNA of an identical twin sent to the International Space Station ended up longer than this stay at home twin.

Tom Price is the new leader of the Department of Health and Human Services. An advocate for the repeal of Obamacare, and an opponent of Medicare, will be sure to shake things up.


Round up Jan 30th – Feb 5th

After a couple of months off, I am restarting the weekly round-up with the aim of truly being weekly — Monday nights to be precise. Welcome back.

Let’s face it – this week it has been nigh on impossible to focus on anything other than the new administration. There have been some genomics related happenings:

  • Berkeley geneticist Michael Eisen has announced he is running for the senate, likely as an independent, because the new administration’s attitude of “basic rejection of the fundamental principles upon which science is based”
  • Trump announcedto pharma execs that his main focus for the FDA will be on improving time to market through decreased regulation. This has many in biotech concerned that hard won consumer protections may be lost, and that startups will not be able to compete alongside the pharma giants. Trump’s pick for the top job at the agency should be announced soon.
  • Meanwhile, although Biotech executives were quick to condemn the immigration ban, pharma execs were mostly quiet— perhaps because Trump thinks they’re “getting away with murder” on pricing, and are hence not wanting to rock the boot.

Despite being dropped by Roche last year, PacBio claims to be doing well. They improved the chemistry to the point at which de novo assemblies (i.e. with no reference genome) became feasible on their Sequel machines. Meanwhile BGI has announced that 2017 will be the year that the tech it bought from Complete Genomics really starts delivering. They will first deliver quantitative RNA-seq on their BGISEQ-500, with plans later in the year to deliver a $600 whole genome.

The MedSeq project, who performed WGS on 100 healthy people and 100 cardiomyopathy patients, has reported that 6 months after sequencing, those who had received WGS results has not cost their health care systems much more than those who had not. This is one of the very few studies providing data on the cost and benefits of WES/WGS. A major worry that payors have is that NGS testing will spark a flurry of expensive follow on tests.

Heidi Rehm weighs in on how low cost sequencing will impact health care. The usual emphasis on the importance of data sharing and models to encourage that. This sentence was thought provoking, “Although genetic counsellors will undoubtedly be critical in this space, it will also become increasingly important to facilitate knowledge-building through online tools and for individuals to take an active role in educating themselves about genetics.” What shape should these online tools take, who should build them, and how to encourage people to take an active role in educating themselves?

Over the last couple of weeks there have been a couple of events to show how seriously gene editing is being taken:

  • The ACMG published a Statement on Jan 26th 2017: “Genome editing is an area of very rapid technological change, so what is not possible today could well become a reality in the very near future… The potential for rapid advance of this approach, and the pressure to apply it clinically, should not be underestimated. The ACMG Board of Directors strongly encourages broad public debate regarding the clinical application of genomic editing.”  They also call out application to nondisease traits.
  • The FDA publishedits intent to regulate genetically modified animals in the same way they do drugs. This has not been popular.


Science – mostly association studies

Height has high heritability — 60-80% of variance in height is accounted for by genetics. We are moving from the world of genetic association studies that could only investigate common variants to the ability to probe rarer variants throughout the genome. When it comes to height, the work of the Genetic Investigation of Anthropometric Traits (GIANT) Consortium spans these approaches

  • In the first iteration of work in 2014, GIANT had focused on polymorphisms — variants that are common (>5%) in the population. They found 700 such variants associated with height, the effect size for any given variant was ~1mm. This work lead to ~20% of heritability of height being explained.
  • In their just publishedsecond iteration, GIANT focused on rarer variants (<5%, >0.01%) in protein coding genes, and found 83 variants, some of which contributed >2cm (nearly 1 inch) if height. Rarer variants like this are harder to study because you need larger sample sizes (this study used >700,000 people), but the rewards pay off. Now ~27% of heritability of height is explained.

This approach still used a chip design (rather than NGS). Explaining more of the heritability will come from larger sample sizes on the one hand, and looking at more of the genome on the other.

Meanwhile, a Chinese team has reported the first general population genome wide significant SNPs associated with mathematical ability, using a chip based assay to get at variants down to 2% frequency. Four SNPs in the SPOCK1 gene each effect match scores by 2.33-2.43 points. (The name of the gene seems very apt, though the high-functioning Vulcan was not the inspiration for the name). The motivation for getting at the genetic underpinnings of intelligence according to the authors: “Understanding mathematics ability is an essential step to improve children’s numeracy skills and academic achievements and could also provide novel insights into human brain functions.” This study is one of many in China focused on the genetic underpinnings of intelligence, something we in the West have far more “hang-ups” about studying.

Another large association study was published this week, this one on the genetic underpinnings of blood pressure. Part of the UK BioBank, this study used an exome chip (for variants down to 0.01%) as well as a SNP array (for variants down to 1%) and looked at over >140,000 individuals.

study of 1463 whole genomes from Finnish individuals compared to the same number of British individuals reveals some of the hallmarks of population bottlenecks, including more loss of function variants in the Finnish population.

The Dana-Farber Cancer Institute reports on their experience of WES for cancer patients, looking at both germline and somatic variants. “The variant review and decision-making processes were effective when the process was changed from that of a Molecular Tumor Board to a protocol-based approach.”

CIViC, a community knowledgebase for expert crowdsourcing the clinical interpretation of variants in cancer, has published a summary of their efforts to date: there are 1,678 clinically relevant curated interpretations of 713 variants affecting 283 genes, using 1,077 publications, and performed by 58 curators. They are dedicated to openness (cough cough HGMD).

Move aside genomics: a review of blood-based proteomics in setting cancer treatment

Round-up Sept 30th – Nov 8th

It has been too long since the last round-up, so here is a birds eye view of some of the more eye catching developments.

  • ASHG 2016 happened
  • The 4th Global Alliance for Genomics and Health happened
  • CMS has updated CPT codes, which is relatively good news for reimbursement potential of NGS tests.
  • Illumina and PacBio shares fall on Q3 earnings reports
  • PacBio suing Oxford Nanapore
  • When genetic testing goes wrong – the misdiagnosis of a family of 20

I am going to try to shift to weekly updates as of now!

ASHG 2016, some of my highlights (which you would have seen if you’d been following me on Twitter…):

  • Daniel MacArthur announces Genome Aggregation Database, contains 120k+ exomes and 15k+ genomes, gnomAD. Wow. #ASHG16
  • gnomAD, initial focus on analysis of protein-truncating regions; pilot studies involving recontact of individuals with “extreme genotypes” #ASHG16
  • Your age written in your genome – age predicted mean error 11 years from NGS data, using telomere length and mosaic chromosome loss. #ASHG16
  • Method to integrate 10k+ functional genomics datasets, can predict about half of conserved non-exonic bases #ASHG16

Global Alliance for Genomics and Health, 4th Plenary Session, some of my highlights:

  • The guiding principle of data sharing not protection from harm, but activating the right to benefit from science (UDHR 27) – B Knoppers @ #GA4GH2016
  • Improving care of individuals and families through “share, compare, resolve” framework for variant curation, BRCA Exchange #GA4GH2016
  • Match Maker Exchange as “a dating service for lonely exomes”, a federated network of databases for finding similar patients #GA4GH2016
  • We should call out those not sharing data, and be on the watch for those “sharewashing”. Voices of industry at #GA4GH2016
  • Data Working group mantra “collaborate on interface; compete on implementation”; data schema and API now available #GA4GH2016
  • #GA4GH2016incoming head Ewan Birney says we have to raise our game, got to make stuff that works practically, will bribe volunteers with chocolate


Market news:

CMS has upped the reimbursement for certain CPT codes covering NGS tests, though not enough for some:

On Oct 11th Illumina announced lower than expected Q3 earnings, driven by a “larger than anticipated year-over-year decline in high-throughput sequencing instruments”. Shares fell by 25% and have not recovered since.

PacBio’s shares also fell on its Q3 financial reports, despite revenue jumping 80%. It has launched a patent battle with Oxford Nanapore, seeking to ban ONT’s presence in the US:

PierianDX bought Tute Genomics. Pierian had been focusing its efforts on software in daignostic somatic testing, and Tute was primarily focused on germline.

23andMe have abandoned NGS testing, instead focusing on their chip technology – NGS was too complex and the consumer price point not achievable.

Veritas, who have a $999 genome on the market, have raised a $30m Series B:

Illumina have launched a new 170 gene panel that uses a hybrid capture technique for DNA and RNA simultaneously. First only for research use, Illumina anticipate its use as a companion diagnostic by pharma companies:

Invitae will offer the ACMG Incidental Findings 56 gene panel on the Helix platform:

The UK government has made NIPT for trisomies 13, 18 and 21 standard of care. Although abortions are expected to rise, this number will be smaller than the number of miscarriages prevented because the use of amnios will go down:

You can now order a genomics test on Amazon:


HLI have published the results of deep sequencing of 10,000 individuals. They find that 84% of human genome can be reliably sequenced via Illumina technology. Each individual contributes ~8,500 novel variants.

Using Complete Genomics long reads technology, researchers at the Personal Genomes Project have published 100 haplotyped genomes, with associated phenotype data:

A study from Ambry makes the case for the continued need for Sanger confirmation, because achieving a satisfactory Specificity/Sensitivity for clinical applications with NGS is impossible, particularly in certain regions. They argue that training pipelines to assess variants in these regions is impossible except with vast validation sets, which labs are unlikely to be able to do.

The BabySeq project, which aims to sequence both healthy and sick newborns, is having surprisingly low uptake, even in the sick baby cohort:

CRISPR potential for sickle cell disease. “What we have right now, if we can scale it up and make sure it works well, is already enough to form the basis of a clinical trial to cure sickle cell disease with gene editing,” first author Mark DeWitt told the LA Times.:

A survey of the practices of clinical diagnostic labs finds large discrepancies particularly in “use of phenotypic data to inform case analysis and interpretation and reporting of case-specific quality metrics and methods”:

An accessible discussion of the graph based genome.

— The first open access graph of ~1000 people will be available by the end of the year

— Seven Bridges unveiled a beta proprietary graph earlier this year – the NCI and the VA are both down to use it

— A competition is being set up by the Global Alliance to “find out who has got the best graph”

— Suspicion of proprietary tools and patents in the space

— Likes of MacArthur (of ExAC fame) thinks long term it will be a thing, but not in a hurry to make the transition away from tried and tested reference based tools

Also in the news:

A family of 20 was mis-diagnosed with a cardiac issue, with at least one very invasive procedure carried out. “When the pursuit of the genotype gets in front of the establishment of the phenotype, bad things happen.”

There are over a dozen companies doing DTC genetic analysis of sports performance. This article is from a woman who ordered several of them and was very unimpressed by both the lack of concordance between tests, and the type of data that she received:

J Lo will star in a new NBC series based in a world where CRISPR’s use is rampant:


Round-up August 23rd – September 29th



Europe is launching a “Million Genomes initiative’, partly in response to the US’s Precision Medicine Initiative, and partly to combat the fragmentation of research along national lines currently seen in Europe:

Zuckerberg and Chan have committed $3bn to curing all the world’s diseases. As some have pointed out, the sum pales in comparison to what is spent on biomedical research, but the ambition is admirable:

Meanwhile, Microsoft have set their intention to “solve cancer” within the decade, using AI. I’m including this blog post from someone skeptical partly because it has such an on point XKCD comic:

Should babies be sequenced at birth? $25m of new projects announced by the NIH will study the practicalities:

A “population-based health intervention effort ” in Nevada had thousands of registrants within the first 24 hours of launch, and all registrants will receive free 23andMe reports. The project is a colloboration of Renown Health (an ACO, with rick phenotype data), the Desert Research Initiative (who have environmental data) and 23andMe.

Interesting papers

How rare is rare enough to be concerning for a variant? A new paper by the ExAC team suggests some thresholds for the “maximum credible population AF” for different conditions. Spoiler: much less than 5%

Using a genomic test to identify which breast cancer patients can skip chemo:  “The main goal of the study (of 6693 early stage breast cancer patients) was to find out whether women with a high clinical risk but a low genomic risk (23%) could safely forgo chemo… After five years, among those who did not receive chemotherapy, 94.4 percent had no distant spread. Those who received chemo fared slightly better: 95.9 percent had no distant spread.”

The future of baby making became even more exotic with the report of the unnecessary nature of the egg cell (in mice at least):

A good illustration of the importance of gene-environment interactions: children with a particular variant were not more likely to develop asthma than average. But children with the variant who had a viral respiratory infection when an infant are 12 times as likely to develop asthma.

Genomiser, an extension of Exomiser, has been published. It prioritizes non-coding variation in the context of phenotype.

The Middle East is an interesting population from a genetic point of view because of high levels of consanguinity (100 times higher than the US) and large families. A study looking at 1111 exomes, found e.g. “rare homozygous putative loss-of-function variants in 301 genes, of which a substantial proportion were novel,”:

A delightful paper that dwells on some of the inconsistent uses of language in genomics. e.g. “It is the phenotype or trait that exhibits a given inheritance pattern, not the pathogenic variant. Thus, it is incorrect to say that a variant “is” autosomal recessive (or dominant, etc.) but correct to say it is associated with an autosomal recessive (or dominant) pattern of inheritance.” And calling it Next Generation Sequencing is so yesteryear, now we’re calling it “Massively Parallel Sequencing”.

A study of how direct to consumer companies handled data privacy and sharing showed most came up short. The study is also of interest for listing the breadth of concerns that such a company needs to address:

5.6% of people who took a direct to consumer pharmacogenomics test have had prescriptions altered based on the results:

You’ve heard of GWAS, you’ve probably heard of PheWAS, but what about EWAS? Epigenome Wide Association Studies are starting to be a thing, such as this one on Schizophrenia, which discovered hundreds of sites differentially methylated in those with the condition, some overlapping with sites previously identified via GWAS:

And in another win for omics, a Metabolomic study has shown clear differences in those suffering from Chronic Fatugue Syndrome – their metabolomes resemble those who have infections. The signal is clear enough to suggest a much needed diagnostic method:

The ACMG are planning on updating their list of 56 genes for incidental findings by year end. Subsequent iterations will likely involve pharmacogenomics genes.

A new database for DNA modification:

Industry News

Health Nucleus, “powered by Human Lengevity Inc” (Craig Venter’s latest project) – an “interactive health platform” completely personalized to you, using your genome.

Tute Genomics recently launched a campaign via Kickstarter offering DTC genomic services. The FDA swiftly sent them a letter and yesterday the campaign was suspended. The whole exome (~$400) or whole genome (~$1000) service would have not only given customers access to their raw data, but also reports on carrier status, ACMG incidental findings, pharmacogenomics, and “information on well-characterized variant-disease risks for different common or complex diseases”: The note on cancellation:

IDbyDNA, a company that shares a significant overlap of brain power with us at Omicia, has raised a $9m Series A. Their tech enables rapid and exact pathogen detection.

Sophia Genetics and Illumina have started a partnership to address the CLinical Diagnostics European market:

GenomOncology will work with Genomics England on their cancer samples:

Color Genomics, who offer a $250 hereditary cancer test, have raised a $45m Series B.

A fairly good overview of what Illumina is up to these days, including Helix: “In the case of Flatley’s own malignant hypothermia, which can result in sudden death while under general anesthesia, a user might query this before deciding on surgery for instance. Or a family might check for an MT-RNR1 m.1555A>G mutation before their child is being treated with gentamicin saving the 1:500 kids with this particular variant from going deaf while in the ICU.”


Round-up August 1st – 22nd

Thermo apparently made a $30bn offer for Illumina, though neither party has commented on the speculations, and analysts argue it would make no sense for either party:

The ExAC paper is out,, and something more accessible from STAT: I already covered some of the results of this >60,000 exomes analysis when the article appeared in preprint, but its important enough to recap:

  • Each ExAC subject has ~54 variants classified as causing a genetic disorder, almost all of which are false positives. i.e. ExAC highlights how bad existing databases are — because of their reliance of studies of very small sample size
  • 99% have a frequency of <1%; 54% are singletons (variants seen only once in the data set); 72% are absent from both 1000G and ESP data sets
  • It covers: one variant for every 8 base pairs (bp) within the exome intervals; 7.5% of all possible synonymous variants; 63.1% of possible CpG transitions (C to T variants, in which the adjacent base is G); 3% of possible transversions; 9.2% of other possible transitions
  • 72% of LoF-intolerant genes have not yet been assigned a human disease phenotype despite clear evidence for extreme selective constraint
  • Filtering on the highest allele frequency in any one population (‘popmax’) rather than the average (‘global’) allele frequency was best for identifying pathogenic variants
  • “The abundance of rare functional variation in many disease genes in ExAC is a reminder that such variants should not be assumed to be causal or highly penetrant without careful segregation or case-control analysis”
  • Average individual has 85 heterozygous and 35 homozygous protein truncating variants (PTVs), only ~2 singletons
  • “discovery of homozygous PTVs is markedly enhanced in the South Asian samples, which come primarily from a Pakistani cohort with 38.3% of individuals self-reporting as having closely related parents, emphasizing the extreme value of consanguineous cohorts for human knockout discovery”

Related stories

  • Many variants previously reported pathogenic are now considered benign, particularly as the result of large scale analyses that show some of these variants are fairly frequent in the population. This misclassification has particularly affected those of African descent, a study on cardiac variants reports. This is largely because of the lack of individuals of African descent in large scale population databases.
  • A study of mutational rates across cardiac genes makes the case for domain by domain knowledge in variant interpretation, and points to much misclassification in databases:
  • In a paper entitled “Using Genetic Technologies To Reduce, Rather Than Widen, Health Disparities”, a group of academics make the case for policy to ensure non-hispanic whites do not get the majority of the benefits from genomics:


Quest/Athena have replied to Amy William’s in the latest step in this genetic-testing drama. Recap: Athena, now owned by Quest, reported a mutation in the plaintiff’s son, Christian, as a Variant of Uncertain Significance. The variant is now reported as pathogenic for Dravet syndrome, a diagnosis which, if made at the time, would have altered Christian’s course of treatment and possibly saved his life (he died age 2). The report, issued in 2007, was sent to Christian’s clinical geneticist, with the recommendation that the parents be tested to assess whether or not the variant was de novo. Ms Williams says she never saw the report. Quest/Athena have asked for the case to be dismissed, and that decision is pending. “If allowed to go forward, legal experts believe that Williams v Quest/Athena could define what constitutes the standard of care for a genetic testing lab.”

Oxford Nanopore and Illumina have reached a settlement over the latter’s suing of the former. The case settled on a particular protein pore, which is not used in the latest ONT product anyway:

Foundation Medicine has submitted its FoundationOne product to parallel review by the FDA and CMS: “If approved, FoundationOne could be the first FDA-approved comprehensive genomic profiling assay to incorporate multiple companion diagnostics to support precision medicine in oncology, including an indication for use as a companion diagnostic across a diverse range of solid tumors.”

A documentary about the Undiagnosed ( has spawned a consortium dedicated to extending genomic analysis of six undiagnosed individuals to contextual info beyond the genome. The participants (in academe and industry) met on the 16th August to assess their results, though the rest of us may have to wait until the film airs in early 2017 to hear how they did. This endeavor is a follow on from the CLARITY challenge.

George Church’s lab is nearing completion of an E-coli genome engineered to use a different genetic code – one that eliminates 7 ‘redundant’ codons, thus freeing them up for potential reuse:

A systematic review of the potential integration of WES into clinical care found interpretation of variants of unknown significance, incidental findings, and the cost and reimbursement of WES-based tests as the most commonly reported challenges.

A blog article making the case for the relevance of genomics to public health (from the Director of the Office of Public Health Genomics, Centers for Disease Control and Prevention):

Blood forming stem cells have been extracted from patients with sickle cell disease, and modified using CRISPR to produce high (and healthy) levels of foetal haemoglobin:

Some cancer-treatment combos work better if the tumor has a higher mutational load:

The Human Functional Genomes Project, which aims to large-scale map genetic variation to functional consequence, has published data on 500 individuals related to immune response, and has more similar projects in the pipeline:

Weaver, a new graph based variant caller from U of Illinois, focused on somatic samples and structural variation.

The UK’s project to do WES or WGS pre-natally, after abnormal ultrasound, is starting to return results to clinical teams:

Counsyl have made the scientific case for expanded carrier screening over existing guidelines (which are more targeted). Their business model rests on such screening.

Google are continuing their push into genomics, with a big partnership with Stanford Health announced:

Oxford Nanopore have announced they can now do direct sequencing of RNA:

How did life on earth begin? Before the modern relationship between DNA and RNA, the “RNA world” hypothesis is that RNA acted both as storer of information and as cellular catalyst. That story just garnered extra weight by the creation of an RNA molecule that can make copies from other RNA molecules (it can’t copy itself, thats a hope of future research). The study used in vitro evolution.

New study of bacterial evolution shows that most new persisted mutations are beneficial:


Round-up, July 12th – 31st

Research highlights

Type 2 Diabetes is known to have a strong genetic component. GWAS analyses have found several associated, relatively common variants (>5% frequency). A huge study looking at WES, low coverage WGS, and SNP data has found very little else. There had been the hope of finding “synthetic association”, i.e. the idea that common GWAS SNP signals come from nearby rare causal variants in linkage disequilibrium. But they found very limited evidence for the role of rare variants in Type 2 Diabetes. An accessible write-up: The original paper:

Money from the ice bucket challenge has helped find some genetic underpinnings to ALS, otherwise known as Motor Neuron disease, and with Steven Hawking being a patient. Two papers:, and NYT write-up:

Circulating tumor DNA is ~20bp shorter than other circulating DNA, which may help the sensitivity of assays:

A new reference genome for the Qatari population (called QTRG), which produces ~750,000 fewer variants compared to the standard reference:

In genetics we have the notion of heritability — how much of a trait is due to genetics, as supposed to environment. For most traits, the amount of heritability we can explain is lower than the estimates of heritability from e.g. twin studies, leading to the notion of “missing heritability” — to be hunted for in NGS data. A new study suggests there may be less missing heritability to be found, because estimates of overall heritability have tended to be inflated, as common elements in environments have not been taken into account:

Some standardized terms for clinical reporting of pharmacogenomic variants have been proposed by an international consortium, in the hope that this will help ease PGx data into clinical care. Table :

A review on the promise of re-evaluating WES data over time

Market News

Cornell have the first NY State DoH approved somatic exome test. They present their computational framework, and claim that in 82% of cases they found something actionable. They play up their abilities to capture structural variants and recently characterized variants as compared to hotspot panel testing. They have an in house knowledge base for assessing actionability:

Software called OncoTrack quantifies the amount of DNA per gene (SPKMG, Sequence Per Kilobase of exon, per Megabase of the mappable Genome), aimed at CNV and LOH analyses. The measure is proposed to be helpful for looking for phenotype associations, and for tracking a cancer over time. It operates on BAM files.

Pathway Emerge will be using the software of GenomOncology, who have a solution for labs reporting out on molecular tests (NGS, FISH, karyotyping, RT-PCR, IHC, and others):

Sanford Medicine will be using Translational Medicine’s software for reporting, initial focus on pharmacogenomics:

N-of-One, who provide curated info on somatic variants, raised a $7m Series B

LabCorp are purchasing Sequenom, who produce non-invasive prenatal tests, for $302m.

Phosphorous, a new computational genomics company, has raised $10m Series A. They have a test for predicting cardiac risk, and a test for familial hypercholesterolemia, with plans for more. They aim to be the “hub of a new computational biology network”:

Commentary and Context

James Watson is not impressed by the cancer moonshot, calling it “crap”, and “the same old people with the same old ideas”:

The ACMG have said non-invasive prenatal screening can replace conventional testing for trisomies 21, 18, and 13 in most women:

Carl Zimmer is keeping on with his “Game of Genomes”: season 1 here,, season 2 here,

The WSJ on healthy people and genome sequencing, giving pilot results from the PeopleSeq project, investigating healthy human attitudes to genome sequencing, e.g. reasons given. Predictably one of the comments is the one word “GATTACA”:


Round-up May 23rd – July 11th


I’ve been away for a while, so a higher level catch-up below.


The FDA has introduced two new sets of guideline docs for NGS testing

1) how test developers could demonstrate the analytical validity of NGS tests for germline diseases

2) guidance on how they might establish the clinical validity of genetic variants gauged by such tests using public databases recognized by the agency

The reaction has not, in general, been positive, e.g.. ‘”They don’t know what they’re saying,” she said, adding that test developers already have good guidelines from CAP and NYSDOH. “What does the FDA think it’s going to add to this?” posited Bale, who is a founding member of the American College of Medical Genetics and Genomics, and an author on NGS standards issued by the group in 2013. “From a technical aspect, they are way in over their heads.”‘

The FDA are trying to distant these guidelines from their proposal to regulate Laboratory Developed Tests, but they’re not convincing many.

The French have launched a national $780m genomes project (cf $215 for the US’s Precision Medicine Initiative), “The first objective is to make France a leader among the countries already laying the groundwork for genomic medicine. Second, is to prepare for the integration of genomic medicine into the normal course of patient care in the country, which means sequencing about 235,000 genomes per year by 2020. And the third goal is to establish a national genomic medicine industry to serve as a source of innovation and economic growth.”

The NIH has a new data system for sharing of genomic data, including clinical data, and focused on cancer data:


“CMS Preliminary pricing plan would reduce payments for several molecular tests“:

InsideDNA will launch in the Fall a Google cloud service that allows users to plug and play various different bioinformatics tools, for e.g. variant calling, but also discovery tools:

Genos, a consumer genomics company arising out of the ashes of Complete Genomics, will give consumers a VCF of their exome, and put them in touch with genetic counsellors if requested. They will make their money by allowing researchers to ask users if they can pay to access their genetic data.

Illumina is looking to make big bucks of software, including variant interpretation:

IBM have partnered with the DVA, with Watson to recommend treatment options based on genetic analysis of veteran’s tumors:

Counsyl move into Oncology with a new business unit:

IDbyDNA, out of Mark Yandell’s lab, is still in stealth mode, but it is based on an exclusive license to the Taxonomer software, details of which were recently published. Taxonomer is a metagenomics tool to identify which pathogens are present.

Hortonworks, who focus on open and connected data platforms, is leading a Big Data/Spark/Hadoop open data consortium, involving Baylor, Mayo, and others:


The LawSeq project has been launched, which will look at: liability, quality of data and variant interpretation, privacy, and frameworks for research-clinical translation.

A study of consumer attitudes to WGS found a lot of heterogeneity, including that 38% of people would not pay for anything for actionable variants, 7% would pay more than $400, and that 3% would pay more than $1000.

Genomics helps enable precision medicine, but its also relevant in public health. The CDC has launched this new database on the relevance of genomics to public health, which will be updated at least weekly:

Scientists behind the “Human Genome Project – Write” published on some of their pilot project goals, and the overall vision:

Mount Sinai reports on using WES and RNA-Seq for tumor-normal analysis

A GenomeWeb write-up from some of the results presented at ASCO includes the report that Foundation’s panel test can be used to assess overall Tumor Mutational Burden. Some studies suggest that high TMB can give better response to therapy.

Craig Venter reports on deep sequencing of 10,000 genomes. One nugget: each genome contributes about ~8,500 novel variants.

A score based on several SNPs helps predict the age at which you start reading, which in turn predicts some socioeconomic outcomes: The score only helps predict ~2% of the variance, but will surely be an early example of whats to come…

A pilot study of 20 sick babies at CHEO identified genetic causes in 40% of them using the TruSightOne panel.

A three year Uganda based study will look for genetic markers that dictate TB and HIV progression in children across three countries: