Was walking around Harvard and saw this. Things have not gotten better in 62 years

www.mobilize.us/nokings/map/... October 18 just about everywhere

Turning off the fire alarms won’t prevent the fire.

#AGBTPH25 JY: Can do all by all testing of compounds x cancer models. 50 cancer models, 1200 drug treatments, 60K experiments, 100M cells

#AGBTPH25 Johnny Yu from Tahoe Therapeutics to talk about virtual cell models. Single cell work has advanced dramatically in 10 yrs

#AGBTPH25 JK: Could apply functional data to reclassify 682 VUS. Look at All of Us data, rare LOF missense have higher risk of cancer. AlphaMissense prediction calls too many variants pathogenic

#AGBTPH25 JK: Lynch syndrome is first inherited cancer risk syndrome, LOF in DNA mismatch repair genes. Most are missense with most being VUS. Apply 6-TG to cells to select MMR- proteins. Mapped 73,568 missense variants. Most have no effect but there are hot spots for impact

#AGBTPH25 Jacob Kitzman from UMich will talk about variants in hereditary cancer syndromes

#AGBTPH25 TY: N=1 collaborative working on sharing knowledge and guidelines for use, need innovative regulatory models

#AGBTPH25 TY: Another dozen genes being treated, more than 30 patients so far. ASOs great version 1 but gene editing could have broader applications

#AGBTPH25 TY: de novo KCNT1 R474H mutation leads to increased channel firing. Used ASO to shut down mutant protein. Daily seizures dropped from 10-15 to 0-2 and stayed down for 50 days but patients developed hydrocephalus. Now being tracked for other ASOs, try to keep dose lower.

#AGBTPH25 TY: Proposal for N=1 medicine to help patients with no other options. Led to FDA guidance for individualized medicine. New trials in ATM kinase, KCNT1 and FLVCR1. ATM trial is longest running N=1. Have seen no clinical progression on primary outcomes.

#AGBTPH25 TY: Milasen is ASO that introduced N=1 medicine. SIgnificantly lowered seizure frequency and improved quality of life but did not save life. Post-mortem studies showed it was working.

#AGBTPH25 TY: Going straight from genetics to therapeutics will increase efficiency. Antisense oligonucleotides (ASOs) are easy to manufacture, easy to predict effects, multiple uses

#AGBTPH25 TY: Mind the gap. Genome sequencing has shortened diagnostic odyssey but not the path to therapy. Many rare diseases are too rare to be commercially investable. Not obvious they weren't being addressed because causes unknown. Now we know what we're not doing.

#AGBTPH25 Tim Yu from BCH/Harvard Med talking about interventional genetics.

#AGBTPH25 HM: ongoing trials of targeted therapies including SCN1A with ASO targeting poison exon and a cell selective gene therapy. For SCN2A, ASO knockdown. How to scale? 13 genes with ASO in preclinical development, 5 in clinic or trial

#AGBTPH25 HM: UBA5 Dad was Hypo/+, mom was LOF/+. Child was Hypo/LOF. Hypo/Hypo can be healthy so try to boost expression of Hypo allele. Tried SINEUPs using lncRNA and CRISPR dCAS9 with activator domain to boost protein expression

#AGBTPH25 HM: identify the target, model the disease, engage patients. Can try gene replacement, ASOs, gene editing, gene regulation

#AGBTPH25 HM: developmental and epileptic encephalopathies - most drugs don't work; early onset, intractable seizures; hundreds of genes identified, 50% diagnostic rate, de novo variants are important, overlap with ID, autism

#AGBTPH25 Heather Mefford from St Jude Childrens Hosp to discuss precision diagnosis to targeted therapies in epilepsies.

#AGBTPH25 SM: Used CRISPR to change isoform abundance (DESeq2). Depletion strongly enriches isoforms expressed in nerve, muscle, adipose, and skin. Depletes blood, spleen, small intestine, lung. Can find intron retentions not normally visible in blood

#AGBTPH25 SM: Looked at outliers of outliers of splicing. See people with hundreds of intron retentions. Even more dramatic if look at only products of minor splicesome. 5 had mutations in RNU4ATAG, one RNU12

#AGBTPH25 SM: If look at extreme outliers, can narrow down to a handful of individuals. Looked at teenage girl with apparent overgrowth syndrome, not Prader-Willi. Saw region on chr11 with reduced methylation near 2.7Mb. If look at all genomes, z score goes from -7.5 to -6.8

#AGBTPH25 SM: outlier molecular phenotypes can inform gene function. Looking at methylation outliers. Find deviant regions relative to the whole population. How unusual is a particular pattern?

#AGBTPH25 SM: Gregor - Genomics Research to Elucidate the Genetics fo Rare Diseases. U11 release >1000 RNAseq, lots more DNA seq

#AGBTPH25 SM: multi omics being used increasingly but usually applied to only a few cases and controls

#AGBTPH25 SM: Whole genome sequencing at scale increases likelihood of finding variants but increases interpretation burden. What is the strategy most likely to be effective for a patient?

#AGBTPH25 Stephen Montgomery from Stanford to talk about multi-omics strategies for diagnosis of rare diseases.

#AGBTPH25 JS: 3 factors and dosage give each disease a distinct signature

#AGBTPH25 JS: Diagnostic categories differ by cell type specific pathways - interaction terms most important

#AGBTPH25 JS: CNVs have dose-dependent effects on traits with SCZ but not ASD. Gene set burden analysis: captures enrichment of variants in genes. Depression and bipolar disorder have different gene set expression patterns across brain

#AGBTPH25 JS: Genome-wide CNV burden only important in a few disorders. Positive and negative associations cancel out burden effect.

#AGBTPH25 JS: Not surprising to see same genes involved in multiple diseases. Large copy number variants provides direction of functional effect right away. All major psychiatric disorders have contributions from rare variants, some plus some minus

#AGBTPH25 JS: Often have rare variants emerge in common diseases in adults. Functional convergence of genes implicated in autism. CNV/WES/WGS point to genes in early brain development. Also schizophrenia and congenital heart disease

#AGBTPH25 Jonathan Sebat from UCSD to discuss linking genotype to phenotype in rare disease

#AGBTPH25 IO: Large degree of data missingness in aggregated studies

#AGBTPH25 Switching to Inez Oh. Leveraging clinical data for approvals. Use AI to aggregate and analyze existing data.

#AGBTPH25 PD: Some companies dropped Sanfillippo drugs. Patient families argued for more use of biomarkers for approval. Use heparin sulfate as biomarker. Placebo controlled trials for such diseases unethical

#AGBTPH25 PD: Sanfilippo syndromes - MPS disorders leading to lysosomal storage disease. Can correct by enzyme replacement therapy

#AGBTPH25 Patricia Dickson and Inez Oh from Wash Univ to talk about computational approaches to clinical trial data harmonization

#AGBTPH25 HR: Have draft standards for sequence variant classification v4.0. New variants start at likely benign. Rates of reclassification vary based on where in VUS spectrum they occur.

#AGBTPH25 HR: Rate of VUS is ~32% in inconclusive genetic tests. Can we eliminate VUS by 2030? Requires data sharing, look at global populations and expand database diversity. Trying to get federated, aggregated summary data for gnomAD.

#AGBTPH25 HR: 1607 variants nominated for high throughput splicing assay to assess function, many hit

#AGBTPH25 HR: Use automated interpretation pipeline to reanalyze genomic data to include more recent information. For VUS, use matchmaker exchange to search for same rare variants across sites. Problem with system is that lists are private. Trying to get more sharing with the gene curation coalition

#AGBTPH25 HR: Long read does not add much to yield. RNAseq has been more useful. Provides information on splicing. Increased yield by 14%. Helps more if have access to disease tissue like muscle.

#AGBTPH25 HR: 45% cases are negative, variant may be missing from sequencing or may be overlooked in analysis. Genome adds ~8% to diagnostic yield over exomes originally. RNU genes now found and add to yield.

#AGBTPH25 Heidi Rehm from MGH talking about strategies to solve rare diseases. 50% of cases not currently resolved.

#AGBTPH25 RR: enrollment planned to reach 10,000 by end of 2026