My short review on best practices for variant calling in clinical sequencing is online (and open access) @GenomeMedicine, just in time for #ASHG20: genomemedicine.biomedcentral…

Great sessions on the last day, too. AI in genetics, beyond the exome for NDDs, and more.

PY: T1 MRI of study participants provides 1,324 brain imaging phenotypes. GWAS yielded 1044 associations (56 novel), gene burden test yielded 148 associations, *(145 novel) #ASHG23

Now: Peng Yang from Takeda (Japan) who will share data from 10,000 Japanese genomes. He led off by pointing out the representation gap in genetic research, which tends to focus on European ancestry individuals #ASHG23

Just bumped into my old friend @dgmacarthur. It reminds me why coming to these meetings in person is so worthwhile. #ASHG23

TIL that the developers of STR detection tools really make the most of naming possibilities... I knew of LobSTR and HipSTR, and now learning about PrancSTR #ASHG23

Tony Mander from Invitae discussing the opportunities of their massive dataset, ~4.5m patients sequenced, of which ~2.5m had accompanying clinical information. A great opportunity for machine learning #ASHG23

PB: International collaboration identified 14 patients with LOF variants, not only survivable but all patients had milder disease than classic CdL. There are some LOF variants in gnomAD in SMC3 (despite pLI=1), but these have low-skewed VAF suggesting mosaicism #ASHG23

PB: Properties Medium-hanging fruit discoveries in Mendelian disorders: Rare conditions Generic or hard to recognize phenotypes Reduced penetrance Variable expressivity Not Mendelian Unique allelic architecture #ASHG23

Now up in Rare Disease Diagnostics: Philip Boone (MGH), Heterozygous LOF variants in SMC3: Lessons from medium-hanging fruit era of Mendelian disease gene discovery. #ASHG23

DK tidbit: variants previously known as P/LP involved in 40% of diagnoses. In 4.3%, they "rescue" variants not otherwise prioritized (e.g. noncoding, higher MAF) #ASHG23

DK now talking about CVA: Virtual single lab across england stores full case info, HPO terms, automated algorithms, variant classifications, evidence logs, and outcomes across all the labs in UK. #ASHG23

DK (great initials): Their pipeline uses some DRAGEN tools, some in-house tools, standard prioritization, and Exomiser. 1 in 3 patients get a diagnosis (confirms this overall Dx rate with huge #'s) #ASHG23

First up: Dalia Kasperaviciute: Interpretaion of whole genome sequencing for clinical care in rare disease on a national scale: the UK experience... they have done 60k WGS for 30k RD referrals through NHS. Wow. #ASHG23

Good morning #ASHG23! I am starting out in "Advancing diagnostics for rare diseases" which promises some great talks.

A nice blog post on putting intentional thought into new bioinformatics methods before developing them. Creative new approaches to study biology are probably more useful than ~2% performance gains over existing methods

5:07 pm yesterday: submitted my abstract to #ASHG23 5:08 pm: ASHG extends deadline to Monday at 9am.

What a week at #ASHG22! Enjoyed all the outstanding science, the joy of seeing old friends & colleagues... and of course, Maroon 5. Feeling inspired & reconnected.

Breaking: 2023 variant classification guidelines from the expanded group of ACMG/AMP/CAP/ClinGen (per H. Rehm) will use a point system allowing for easier tiering, draft terms for VUS tiers are VUS-low, VUS-mid, VUS-high #ASHG22