I’m back and nearly recovered from the American Society of Human Genetics meeting in Orlando, Florida. The conference was well-covered on the #ASHG17 hashtag as usual, but I also compiled detailed notes that I’ll share here. The obvious theme of this meeting, from my point of view, was the rise of clinical genome sequencing in patients with inherited disease.

Genomes for breakfast with Karl Stefansson from deCODE

I heard a great talk by Karl Stefansson at one of the early-morning exhibitor events. He said that DeCODE has genotyped 425k individuals, performed WGS on 50k, and RNA-seq on 10k. Of particular interest was their work on de novo mutations in families. Some background:

• Each one of us born with ~70 de novo mutations
• 1/10 children born with LoF mutation in a gene
• 1/20 children born with LoF mutation in a gene expressed in brain

De novo mutations shared by siblings

Decode examined 1,010 sib pairs from 253 families:

• 447 autosomal de novo mutations shared by two or more siblings. (out of 17,710 per fam?, 2.5%)
• But notes use of strict filters: no reads showing mutation in parent selects against mosaic
• Age effect: the older the father, the fewer de novo mutations shared
• 70% of the diversity in de novo mutation rate is explained by paternal age. 2x mutations if 40 yo father vs 20 yo

Stefansson also relayed some interesting work on non-transmitted alleles in parents of children, finding that even when not passed on, many were significantly associated with the kids’ educational attainment, socioeconomic status, and other factors. He called it “genetic nurturing” and I love the concept. He also offered a few memorable quotes during his talk; my favorite was, “In human genetics, your competitor always becomes your collaborator.”

Clinical WGS: One Test to Rule Them All

Much of the ASHG conversation of course, was driven by Illumina. At their lunch symposium, CSO Ryan Taft touted clinical WGS as “one test to rule them all” — a single assay for detecting SNPs, indels, repeat expansions (an area of active software development at the company, i.e. ExpansionHunter), and large structural/copy number variants. Although I’m personally dubious about the ultimate ability of short-read sequencing to fully characterize large repeat expansions like fragile X and C9orf72, Taft sounded optimistic. He also shared some vignettes from the iHope network, an initiative to provide clinical WGS pro bono to families who can’t afford genetic testing. Their 2017 cohort included 81 cases; I took some quick numbers down (note, not

• 62 trios, 14 duos, 5 quads
• 7% were already likely positives (VUS that later got bumped)
• 7% reached a partial explanation (some but not all of disease)
• 21 cases with likely causal variants; many of these involved CNVs
• Clinically signif. variant types: 48% SNVs, 4% indels 1-4 bp, 31% CNVs 19kb-15mbp, 4% gross abnormalities, 11% multiple variant types, 2% LOH/UPD

Taft described one case of a de novo deletion on 19q13 (KMT2B) in 9-yo male. Array results were “not accessible” at time of referral. Obviously such an alteration would have been detected by an array at a significantly lower cost. Taft said that the patient had gotten an array, but the results were “not available” when they decided to proceed with sequencing. Even so, WGS certainly does have an advantage over clinical arrays for detecting kilobase-scale CNVs. The challenge, of course, is the cost.

The Clinical benefit of WGS in 300 families

Peter Bauer from Centogene gave a nice summary of clinical WGS in 300 families. Their interpretation strategy, it must be said, is nevertheless very coding-centric: they focus on coding region +/- 10bp, examining deep intronic variants only if already described as pathogenic, in an established candidate gene, or a region of homozygosity. He also highlighted the importance of previous testing on clinical WGS yield: they saw a 26% diagnostic rate for cases without previous WES testing, compared to 18% for cases that had been through WES. Interestingly, 55% of their diagnosed cases had a positive family history. I interpreted that as an illustration of the power of family-based sequencing with multiple affected relatives. A breakdown of pathogenic variants:

• 29% splice site (wow!)
• 25% missense variants
• 22% nonsense variants
• 22% indels (20% frameshift, 2% inframe)
• 2% large deletions

Note the contrast with Illumina’s numbers, which featured CNVs much more prominently.

 

Mendelian Disease Updates

Here are some highlights across a number of talks about clinical/research sequencing in rare inherited disorders.

 

Posey on behalf of CMGs on clinical impact of gene discovery

Jennifer Posey gave a nice report on progress at the Centers for Mendelian Genomics, from the Baylor perspective. She reminded us that the CMGs post their candidate genes on a website, many before publication. Baylor alone has discovered 300+ disease genes; 14% of these have already led to a diagnosis in their clinical lab (accounting for 3-4% of positive reports). She also mentioned that clinically negative WES cases are referred to the research lab for further evaluation, and 51% (yes, HALF) of these ultimately get a candidate diagnosis. I asked a question about this, because the rescue rate seemed almost too good to be true. She said that many of those rescues were achieved after additional family samples sequenced, which allowed a more definitive interpretation of a VUS. However,  sometimes new disease gene publications were the source.

 

Rare and Undiagnosed Diseases in Pediatrics Initiative in Japan

Japan’s initiative has a 32% diagnostic rate, 77.4% autosomal dominant (presumably de novo), 12% X-linked, 9.1% autosomal recessive. I think it’s valuable to include these numbers because they show similar trends to what we’ve seen in predominantly European ancestry cohorts: a 1-in-3 solve rate, and a majority of positive clinical reports due to de novo mutations.

The contribution of rare recessive coding variants to severe developmental disorders in DDD

Hilary Martin gave a nice talk about the DDD project, which has enrolled over 10,000 cases and already yielded some valuable insights into the genetics of developmental disorders. For example, de novo mutations account for 40-45% of cases. Their analysis framework for recessive variants suggests that parents show lower frequency of biallelic variants than expected by chance … because presumably lethal or disease-causing biallelic pairs are not in healthy parents. There are twice as many recessive genes as dominant ones, but even so, recessive genes account for only a fraction (~5%) of their mostly-European-ancestry cohort.

A Review of Clinical Exome and Genome Sequencing

Stephen Kingsmore from the Rady Institute presented a meta-analysis of exome/genome sequencing in 19,715 infants/children from 75 studies. It’s an impressive dataset, and I think the results are thus very compelling:

• Diagnostic sensitivity: 0.40 for WGS, 0.35 for WES, 0.09 for chromosomal microarray.
• Between 2013 and 2017, the rate of diagnosis increased 16% each year.
• Rate of consanguinity in study is inversely proportional to rate of pathogenic de novo mutations
• Odds of Dx by trio testing were TWICE that for singleton testing
• The diagnostic rate was higher at hospital labs (0.41) than reference labs (0.28?)
• In six 2017 studies, odds of Dx by WGS/WES were 8.3x higher than that for CMA

Kingsmore found only 2 small studies which compared diagnostic sensitivity of WES versus WGS; both reported no significant difference. He said “the literature needs to catch up,” which I take to mean that he thinks WGS has a clear diagnostic advantage.

Automated reanalysis of clinical exome data in a fee-for-service clinical Dx lab.

One of my favorite talks was from Sam Baker from CHOP on their pipeline for automated re-analysis of clinical exomes which were reported as negative (80% of their first 300 probands in the lab). They have an approach for mining weekly PubMed abstracts (from new papers) to see if they inform past exome cases. He described “Clinical correlation” (manual assessment of gene-disease phenotype overlap) as the most important and most time-consuming step. They have a

Phenotype terms retrieved from chart review were used with genotype data. Automated correlation comes from 3 filters:

1.) Patient phenotype to gene-disease compares gene symbols and phenotype terms in PubMed Abstracts/OMIM entries

2.) Filters based on variant characteristics, quality, MAF, effect

3.) Filter based on segregation / inheritance

Their reanalysis of 240 negative exomes revealed 27 novel diagnoses

• 17 new disease gene
• 6 classification update of previous VUSes
• 1 new phenotype described for a known disease gene
• 3 candidate genes that don’t meet lab’s diagnostic criteria but they think they have it
• The diagnostic yield went from 20% in the initial analysis to to 29% with automated reanalysis. Their pipelines have really whittled down the monthly re-analysis workload to make it feasible (it’s usually 1-5 variants per case per month).

 

 

Next-generation sequencing technologies have transformed the way we study rare genetic disorders. Exome sequencing for Mendelian disease is the poster child for success in this arena. However, despite the frenzied pace of disease gene discovery and the growth of public sequence databases, the diagnostic/success rates for exome sequencing have remained somewhat constant over the past several years. Depending on the type of disorder, available samples, mode of inheritance, and other factors, one can expect that ~35-50% of clinical cases may be solved by exome sequencing.

On the research side, the success rates for rare disease studies have often gone down, for the simple reason that as we improve the catalogue of disease-causing genes/mutations, genetic tests become increasingly effective. In other words, easy-to-solve cases never make it to a research study. The cases that do make it through, while valuable for gene discovery, will almost certainly have a low success rate.

Knowing this intellectually does not make it any less frustrating when you can’t solve a rare disease patient or family where there’s clearly something to find. When an exome is negative, what recourse do we have? Whole-genome sequencing is often what researchers consider pursuing next — in theory, it enables the detection of variants that are not detectable by exome sequencing (e.g. in the 5-10% of coding bases that exome sequencing does not sufficiently cover, or structural variants with noncoding breakpoints). Yet the “rescue rate” of WGS for negative exome cases has been disappointing.

Transcriptome sequencing offers a possible alternative to further investigate negative exome cases. For certain categories of variants (e.g. nonsense, frameshift, and splice site mutations), RNA-seq can provide a molecular confirmation of the predicted pathogenic effect. This is especially helpful in the gray area of splice-region variants, the effect of which is difficult to assess from genomic information alone.

A recent paper from the MacArthur lab demonstrates the utility of RNA-seq in patients with genetically undiagnosed rare muscular disorders. They applied RNA-seq to muscle tissue from 50 unsolved cases. Four of these had a splice region variant of uncertain significance. Another 12 had a strong candidate gene based on clinical presentation and/or the presence of a single pathogenic variant in a known recessive gene. The majority (34) had undergone exome/genome sequencing but yielded no candidates.

Pathogenic Extended Splice Variants (2 cases)

First, the straightforward analysis. In the four cases with candidate variants in the extended splice region (i.e. not the canonical splice site, but close to it), RNA-seq confirmed pathogenicity in 2 of them. This is a small sample size, so all it really does is highlight the difficulty of predicting variant effects outside of the canonical splice site.

Unrecognized/Undetected Splice Variants (4 cases)

Interestingly, RNA-seq enabled the diagnosis of a three cases in which a splicing variant was not suspected:

• 1 with a splice-region variant was pathogenic but had been missed by exome sequencing
• 1 with a missense variant that alters splicing
• 2 synonymous variants that alter splicing

The missense variant is interesting: it creates a stronger splicing motif than the wild-type sequence, leading to an exonic splice gain:

Deep Intronic Variants that Alter Splicing

The real power of RNA-seq in cases like these is to identify novel splicing events that are not due to obvious splice site/region variants. In seven patients, the authors of this study identified aberrant splicing caused by deep intronic variants in known disease genes (DMD and COL6A1).

RNA-seq is really the only approach that can identify an event like this. Exome sequencing doesn’t cover this region. Whole-genome sequencing does, but undoubtedly will return dozens of intronic variants, most of which have no effect on splicing.

Rescue Rate for RNA-Seq

Overall, transcriptome sequencing enabled the diagnosis of 17 patients, or 34%. This rate is skewed a bit high by cases with candidate splice region variants (2/4 solved) and cases in which there was already a strong candidate gene (66%). Among cases without such strong leads, the rescue rate was 21%. This is a not-inconsiderable improvement to the overall diagnostic rate, and a strong argument that RNA-seq should be considered as a possible next step when exome sequencing alone does not achieve a diagnosis.

Importance of the Right Tissue

It should be noted that muscular disorders were well-suited to a transcriptome approach, because the relevant tissue is both accessible and often already banked due to the fact that muscle biopsy is a routine (if painful) diagnostic procedure. For other types of rare disorders, one may not have this luxury. In retinitis pigmentosa, for example, the relevant tissue is the retina. Good luck getting any RP patients to volunteer any of their diminishing retinal cells. Developmental timing is also a consideration: for some disorders, like birth defects, the genetic aberration affects a certain phase of development that is no longer ongoing when the patient is in the hospital.

You’ll note that the figures shared in this post also included, in blue, the splice pattern observed in normal individuals. These were gleaned from a subset of the skeletal muscle samples sequenced by the GTex consortium that were carefully matched to the patient specimens. A tissue-matched normal is incredibly powerful for analyses such as this one, and reinforces the need for more large-scale efforts to put transcriptome data in the public domain.

 

 

One significant challenges accompanied by advances in high-throughput genomic technologies (i.e. high-density arrays and next-gen sequencing) is that the speed at which we can detect candidate variants outstrips our ability to interpret their effect on phenotypes. In other words, genomic discovery is fast, but functional validation is slow. Also, the relationship between a sequence variant and an observable human trait tends to be complex, making it difficult to elucidate the mechanisms underlying an association.

Gene expression offers an appealing intermediate phenotype. It’s influenced by sequence variation, quantifiable, and can be assayed in high-throughput fashion. Early systematic efforts to identify expression quantitative trait loci (eQTLs) revealed some fascinating insights into the genetic regulation of gene expression, such as the fact that much of that genetic control occurs in noncoding regions. Also, most eQTLs were identified in proximity to the affected gene (cis-eQTLs), but some were millions of base pairs away or on other chromosomes entirely (trans-eQTLs).

Most eQTL surveys conducted thus far are limited to a certain cell type — most often, lymphoblastoid cell lines — and utilize genetic information from SNP arrays or low-pass whole genome sequencing (WGS). That’s why I’ve been excited to see the results of the Genotype-Tissue Expression (GTEx) project, which has applied RNA-seq to numerous different human tissues from donors that also underwent genome sequencing. In addition to a wide set of tissues profiled, their WGS data allows for a fairly comprehensive interrogation of genomic variation including large structural variants (SVs).

In the latest issue of Nature Genetics, members of the GTEx consortium led by several of my former WashU colleagues report the most extensive study to date of the impact of structural variation on gene expression. Their dataset comprises RNA-seq from 13 different tissues along with deep WGS (blood) from a total of 147 individuals.

Genome-wide Structural Variation

A strength of this study is that the authors have considerable expertise in identifying SVs from genome sequencing data. Across the 147 individuals, they identified 23,602 high-confidence SVs, including deletions, duplications, inversions, and other SVs.

Common SVs Affect Gene Expression

Of the 9,000 common SVs (MAF>0.05) in their dataset, 1,634 were associated with gene expression variation in at least one tissue. Because an SV can influence gene expression in multiple tissues and vice-versa, their complete analysis uncovered 5,128 individual eQTLs affecting 2,064 different genes.

Some 11% of the eQTLs involved an SV that altered one or more exons of the affected gene. More than 90% of the time, the effect on gene expression is consistent with the type of SV involved, i.e. deletions tend to decrease expression and duplications tend to increase it:

That’s a nice result if not an altogether surprising one, since you’d hope that changing the genomic copy number of a gene would alter its expression in somewhat predictable ways.

Contribution of SVs to eQTLs

The authors also identified eQTLs involving SNVs and small indels, which are more numerous and generally easier to detect than SVs with current next-gen sequencing technologies. This offered a more complete picture of genomic variants affecting gene expression, and allowed them to characterize the relative contribution of each type of variant to eQTLs. Using fine-mapping approaches, Chiang et al found that for 3.5 to 6.8% of cases, a structural variant was the probable causal variant.

This might seem like a trivial contribution, but remember that the typical human genome harbors ~4 million SNVs and indels, but just 5-10,000 SVs. On a per-variant basis, an SV was 28 to 54 times more likely to modulate gene expression than an SNV or indel. Moreover, this is probably an under-estimate of the contribution of SVs, since our ability to detect/genotype them with current WGS technologies is considerably less than our ability to detect smaller variants.

Noncoding SVs Affect Gene Expression

Interestingly, 88.3% of SVs predicted to cause an eQTL did not affect gene dosage or structure. Instead, they likely exert their effects via noncoding regulatory mechanisms. Cross-referencing them with several classes of predicted regulatory elements revealed that SV-eQTLs were significantly enriched in regions that were:

• 10kbp upstream or downstream of the gene transcript
• Within 1 kbp of an enhancer
• Predicted to bind transcription factors (by FunSeq)

Overall, this study beautifully illustrates the important role that structural variants have in modulating gene expression. It also lends additional evidence to what I’ve been saying for a long time: that noncoding regulatory variants probably underlie the majority of inter-individual differences in humans.