Wednesday, 9 September 2015

#GenSci15 day3 clinical genomics session

Andrew Beggs is chairing the clinical genomics session – this is why I am here, cool stuff!

Interesting that the Tolstoy quote from Pippa Thomson's talk (see below) "Happy families are all alike; every unhappy family is unhappy in its own way" came up, as it could be argued that the Anna Karenina principle probably applies to the clinical use of genomics right now: a deficiency in any one of a number of factors dooms it to failure. Clinical genomics can only have a successful impact if every possible deficiency in the study has been avoided i.e. clinical and family history, quality of exome coverage, issues with alignment etc, etc, etc,

Sarah Ennis, Professor of Genomics, University of Southampton "Paediatric exomes: diagnostic and data dilemmas"
Talking about clinical diagnostics work in Southampton. Case report describing metabolic SCID "boy in the bubble" syndrome, genetic basis. Sequenced RAG1, RAG2, JAK2 at about £800 each using Sanger. Hoped that sibling would be a good match but had same diagnosis of SCID, now two kids with profound SCID. Exomes done externally but nothing found in SCID genes.

Asked to analyse data. QC checks on family structure, turned out that family was not Caucasian as described but had Sri Lankan ancestry, important to be clear about family history. Some clues in folate responsive immunodeficiency, analysis of folate genes identified a mutation in MTHFD1 L51P that causes kink in enzyme, a key enzyme in the folate pathway. Very clear in Sanger validation, however also in Mother. In depth analysis of the father showed that exon 13 appeared to be much lower coverage than seen in controls, they found a deletion in Alu repeat in MTHFD1 exon 13 and the mRNA was degraded before translation via nonsense mediated decay. Likelihood is that normal copy of protein not processed due to exon 13 deletion so only mutant version present. Boys now on folate supplement and pretty much back to normal. Interesting case of compound heterozygosity.

Second case: aHUS case referred from Nephrology for exome analysis. Had previously Sanger sequenced multiple genes and found nothing. Targeted analysis of the exome to 220 genes involved in aHUS. Found 2 variants of unknown significance. No real resolution for case. Masters student carefully reviewed notes found Moyamoya (stroke), developmental delay, chryptorchidism, persistent thrombocytopaenia, splenommegaly, dysmorphic – repeated analysis more thoroughly. Identified 51 variants in 44 candidate genes filtered down to a CBL gene and determined a de novo mutation in the child that affects splicing and a Noonan-like syndrome. CBL mutation carries risk of leukaemia due to LOH, checked b-allele freq and found clear LOH on 11q and is showing progression to leukaemia. Horrible diagnostic dilemma as the only two other cases died within 12 months, but bone marrow transplant not an easy therapy to push through the different clinical groups without a confirmed leukaemia diagnosis.

Sped through a 3rd case history, again family QC showed information not presented to team that would signifjcantly impact clinical care.

Nick asked about the continued use of Sanger validation, response from Sarah was that until the clinic can make a decision based on the NGS the Sanger results from a clinically accredited lab need to be done to make decisions.

Pippa Thomson, Centre for Genomic and Experimental Medicine, University of Edinburgh "Using high-throughput sequencing to understand the genetic complexity of psychiatric illness"
Happy families are all alike; every unhappy family is unhappy in its own way Leo Tolstoy Anna Karenina, see Anna_Karenina_principle. Schizophrenia, Bipolar and major depressive disorder: common, serious and debilitating, poorly understood, major unmet clinical need. Strong evidence for shared genetic basis to these complex disease but very complex genetics and strong environmental effect. 2014 Nature 80,000 sample meta analysis shows lots of hits in the genome – lots of rare variants, balance between common and rare variants.

Blackwood 2001 paper describes a large family with balanced 1:11 translocation affecting DISC1 that showed a significant risk of psychiatric disease, although some family members are not affected – incomplete pep entrance. Updated analysis now 6 generations from 56 individuals to 105, now a small number of individuals without the translocation have major depression. WGS on 49 family members at CSHL. Fund two regions with strong association which are the translocation breakpoints and other variation in Chr 2 & 5. Extended family shows LOD score of almost 8 for t(1:11) if any psychiatric disorder considered.

Performed NGS on 528kb in 1500 cases to 80% coverage 30x depth used lrPCR (ouch - could be made easier using lrPCR on AccessArray?) found an abundance of rare variants, 60% novel, almost a variant in each affected individual! Nothing showed up for Schizophrenia, Bipolar but found association for recurrent major depressive disorder. See Thomson et al Mol Psyhc 2104

Anthony Rogers, Congenica "Why we care about rare diseases"
Described DDD project, trio sequencing of 14000 children, 24 UK regional Gnetics units, 180 clinicans, recently published in The Lancet. Rare disease not really that rare 5:10,000 for any individual "rare" disease, but works out to 1:17 people affected by something rare. Fragile X 1:5000 compared to oculootodental syndrome 1:1,000,000. Impact of rare disease is high lots of time and emotion spent in diagnostic odysseys averaging 4.8 years and 7.3 doctors!

UK genetic test centres spread across UK, lots of different capabilities and scale, limited bioinformatics support, generally using targeted panels 10-100s genes. NHS IT systems inadequate. Clinical scientist interpretation bottleneck.

Ant described the Congenica Sapienta platform that can take multiple filtypes in, analyse multiple gene panels and take relatives genomes. This is one of the platforms being developed for use in the 100,000 Genomics England project. Goes from data to report in a single workflow, and allows users to share data more effectively whch is very important for rare diseases. Individual data is not shared, Congenica provides ananymised sample data. Consisten analysis makes interpreation much easier.

Examples: Described Sclerosteosis with less than 100 cases worldwide, caused by mutations in SOST gene. Can cause hearing and sight loss due to abnormal bone growth. Rare disease informs on comon disease and comparison of Sclerosteosis, normal and osteoporosisby Amgen allowed development of a therapeutic agent in phase III clincal trials. Also Viagra being used in pulmonary areterial hypertensions (1:100,000) and remarketed as Revatio.

Congenica can provide tools to help in analysis of data.

Saskia Sanderson, Icahn School of Medicine and UCL "Should everyone have their genomes sequenced?"
Sarah graduated from Psychology in Birmingham! She's talking about insights from interviews with personal genomics research participants in the HealthSeq project. As costs for WGS have dropped should a healthy person get their genome sequenced? Problems can be psychosocial/psychological especially around false +ve results and variants of unknown significance.

Saskia is involed in the research to understand the impact of WGS results on individuals see ESHG publication and the video below...

35 individuals WGS and comprehensively interviewed, slightly higher end of income and education levels. Emotional response ranged from positive to negative but most people were generally positive. Negative responses were based on interpretation and technology, could not get something that actively helped to improve health; or confusion around results and interpretation. One case stood out, young man late 20s in HealthCare highly educated got pathogenic result SCN5A gene (sudden cardiac death syndrome). Spent time reading literature and was concerned with his results, very worried about insurance implications, saw cardiologist and got reassurance on best ways to proceed. Also had APOE e4/e4 with 80% lifetime risk! Across the study there was no significant behavioural response, look at smoking - you don't need to sequence a genome to know you should not smoke.

Interesting to understand the impact of negative results across age spectrum, I'd imagine that this will vary quite significantly. Although this was a small study it certainly  underpins the need to do more research like this. PeopleSeq brings together HealthSeq, UYG and PGP participants into a single study.

Awesome talk!

1 comment: