This week we attended the American Society of Human Genetics meeting in San Diego. This was really a tour de force on the application of novel genomic technologies to the understanding of human biology. Our thoughts on the sessions are below; there was lots of interesting stuff and one of the negative things of the conference was the fact that there were so many interesting concurrent sessions, which meant that one had to choose which talks to see and miss all the other ones.
In addition to our thoughts below, we also live-tweeted the conference, so if you want the play-by-play, check out https://storify.com/thegblab/american-society-of-human-genetics-2014.
Live-tweeting the conference was a particularly satisfying experience; our comments reached an average of 17k people per day with a total of 93.1k people potentially seeing our tweets during the conference. Even if only a tenth of these people were “paying attention”, it is still an extraordinary number.
We actually ended up on the list of the most “influential” accounts live-tweeting the conference – just above Illumina at the time of writing :).
ASHG is known for having a very high twitter volume. Over 20k tweets were made during the conference with its hashtag.
Below are the two posters we presented at the conference: One dealing with our work in Dementia with Lewy Bodies and another with our work studying recessive forms of dementia in the Turkish population.
Below are our thoughts on the conference.
Genomic Alliance for Global Health (GA4GH) Pre-ASHG Plenary Meeting
This was a daylong meeting with several interesting concepts being discussed. We thought the following were particularly relevant:
- We need appropriate standards for data representation, benchmarking tools and data sharing. There are several ideas being developed by the GA4GH members.
- Data integration needs to improve and be better implemented.
- It is inevitable that we will have a cloud-based future. Data storage and analysis from most large-scale projects are being moved to the cloud now.
ASHG14 Distinguished Speaker Symposium
Three talks by IBM, Google and the CDC on “Big Data”.
There were many interesting ideas being presented and discussed. Both IBM and Google are applying computation to the interpretation of genomic data. From a data perspective, genomics is really small in Google’s scale. IBM’s Watson is currently analyzing cancer genomes and making predictions for most applicable individual-level therapies.
A couple of notions were brought up by the CDC that we have become very used to hear in our lab: Garbage In Garbage Out (GIGO) and that the scientific method is fast becoming obsolete.
This was a very innovative session. Crowd mining of genetic data, the fold.it project, and a couple’s quest to cure prion diseases were all great talks showing how useful it can be to involve non-researchers in the process of research.
Complexities of Simple Mendelian Disorders
Great session with Anna Rose from UCL Institute of Ophthalmology delivering a great presentation on non-penetrance in PRPF31-associated retinitis pigmentosa. Natassia Vieira from Boston Children’s gave an interesting talk on using dog models to study muscular dystrophies similar to what we are doing in a dementia-related disorder.
Variant Calling: What makes the difference
Very interesting session with a wide range of topics. Improving the genome reference by using sequences derived from populations closer to the study samples or by using a graph-based pan-genome; new algorithms for sequence alignment and variant calling; detecting CNVs in WGS data. One particularly interesting talk from Deanna Church who made a very compelling case for the update to GRCh38. We will have to deal with this soon.
Genetics and Mechanisms in Neurological disorders
A couple of talks on essential tremor: TENM4 mutations segregating in 3 ET families and the p.G399S HTRA2 in a 6-generation kindred with Essential Tremor and Parkinson’s Disease. Data suggesting a novel disease-causing variation for CMT. GWAS for migraine with 29 new risk loci, pointing to biological pathways of ion homeostasis, nitric oxide signalling and oxidative stress.
The Human Knockout Project
Lots of work done on Loss of Function mutations. Examples of LoF mutations in a spectrum of impact, some are actually good for an organism, while others are deleterious. Several examples of why it can be difficult to identify these from seq data.
Architecture and Impact of human Knockout Alleles
Dan MacArthur releasing ExAC: 63k exome-seq’d individuals (remember, these include 1kG and EVS). Work from deCODE on identifying and ultimately deep phenotyping the 8041 individuals with completely KO genes. Interesting presentation on NUTvar. 1000 samples from the 1958 British Birth Cohort have been exome sequenced and analysed for protein truncating variants.
Cloudy with a Chance of Big Data
Several talks on improved methods for data analysis. Ryan Layer and Chris Chang presented tools we have been testing and applying to our ongoing work. All talks presented exciting developments and ideas.
Genetics of Complex Neuropsychiatric Disorders
Quite a few good talks. Results from autism gene networks and how transcriptome sequencing of human aging brain tissue showed genetic effects on splicing in AD were very interesting, as were the exome chip results from the AD work.
Goncalo Abecassis talked about the completion of the 1kG project. Lots of great work done here. We are all thinking of the 100k genomes, etc, but this is really seminal work.
Neurogenetics: From Gene to Mechanism (II)
Several talks associating mutations with phenotype. The CHCHD10 gene for ALS and FTD was perhaps the most directly relevant to us – this was also a very good talk.
Diagnostic Yield of New Genomic Technologies
Several talks showing how the introduction of genome-wide genotyping arrays and sequencing has revolutionized diagnostics. The amount of diagnosis that are only now possible is staggering. This was amply emphasised by all speakers.
Exome Sequencing as Standard of Care in Clinical Genetics
Another series of great talks showing the power of exome-seq for identifying disease causing mutations and group’s experiences of using this in a clinical setting.
We Have the Technology: Next-Generation Genomic Methods
A series of really interesting novel technological approaches. The targeted locus amplification (TLA) from Cergentis was possibly relevant to our work.
Genomic Variation: Interpreting the Uninterpreted
Talks on how to improve our annotation of variants. Thumbs up from the speaker to in-silico prediction softwares which seem to perform well in predicting novel pathogenic variants in well characterised genes.
Heritability and Risk Prediction for Complex Traits: Regulatory Variants and Polygenic Models
Great talks here as well. We would single out Eleftheria from the Sanger Inst. with data on population isolates from Greece and African cohorts from Uganda. Two very different projects with very interesting results. Also, Frank Dudbridge from UCL with a very thorough talk on polygenic risk analyses.