Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
• Introduction • ACMG-AMP Classification System • Other Classification Systems • Evaluation of inter-laboratory concordance (Amendola et al. 2016)
Please download ACMG Standards and Guidelines (page 8-10)
https://www.acmg.net/docs/Standards_Guidelines_for_the_Interpretation_of_Sequence_Variants.pdf search: “acmg standards and guidelines”
OVERVIEW
In the not-too-distant future… …we will not perform NIPT anymore …we will do WGS in four seconds
INTRODUCTION
In the not-too-distant distant future… …we will not perform NIPT anymore …we will do WGS in four seconds …INCLUDING interpretation / classification Did you notice the molecularbiologist ?
INTRODUCTION
Starting Point:
Assessing pathogenicity is not an easy task (it never was, but
NGS does not make it better…)
Conflicting systems among labs / reseachers cause confusion
(best case) or wrong clinical recommendations (worst case)
Need to standardize Variant Classification
VARIANT CLASSIFICATION
Qualitative evaluation of different data types (28 defined criteria with assigned code)
Each code is assigned a weight (stand-alone, very strong, strong, moderate, or supporting) and direction (benign or pathogenic)
Variants then can be assigned in one of 5 classes (IARC 5-tier system)
If not enough lines of evidence are invoked to classify a variant as P, LP, LB, or B, or there are valid but contradictory lines of evidence, a variant is interpreted as a VUS
ACMG GUIDELINES
General Considerations: • for variants in all Mendelian genes (single gene, gene panel, exome, genome or transcriptome)
• not for somatic, Px, multigenic/complex disorders
• cave: candiate genes („genes of uncertain significance“; „GUS“; Oct.2016: OMIM 3648 genes)
• the terms „mutation“ and „polymorphism“ should not be used
• instead use „variant“ (pathogenic (5); likely pathogenic (4); uncertain significance (3); likely benign (2); benign (1))
• variants should be reported using the HGVS nomenclature (http://www.hgvs.org/mutnomen)
• to provide flexibility: some criteria listed as one weight can be moved to another weight using professional
judgment, depending on the evidence collected (multiple observations of a variant in trans with path. variant – PM3 to PS)
„Pathogencity should be determined by the entire body of evidence in aggregate, including all cases studied, arriving at a single conclusion“
ACMG GUIDELINES
ACMG GUIDELINES
What are the requirements? • Detailed population frequency data • Thorough literature search (time consuming, often not consistent) • Clinical databases / LSDB´s • Link genes to associated phenotypes and MOI (latter is sometimes not trivial) • Bioinformatic prediction integrated (protein, splice sites)
He who makes no mistakes, makes nothing …. you won´t get the best possible classification for your variant (=VUS)
InSiGHT classification rules (4 MMR genes) • VIC (J.P. Plazzer) • monthly conference calls, ~10 variants classified per session • InSiGHT-LOVD DB
ENIGMA classification rules (BRCA1/BRCA2)
OTHER GUIDELINES
• 9 CSER-labs classifying 99 variants (9 all, 90 by 3 labs, 2 labs contrib. ACMG)
• Using the lab´s own method and the ACMG-AMP Guidelines
• Agreement within labs was high: 79% concordance
• BUT: only 34% concordance for either classification system across labs
• After consensus discussion and detailed review: 71% concordance
EVALUATION OF INTER-LABORATORY CONCORDANCE
EVALUATION OF INTER-LABORATORY CONCORDANCE
SGP7 BTD
No phenotype when homozygous (found homo in ExAC) Pathogenic only when in combination with pathogenic variant
EVALUATION OF INTER-LABORATORY CONCORDANCE
Reasons for Discrepancy • VUS criteria: conflicting pathogenic and benign lines of evidence
Some labs allowed one line of conflicting benign evidence of only a supporting level
(e.g., predictions) to override otherwise strong evidence of pathogenicity. Expert
judgment: disregarding limited conflicting evidence
• 2 discordant variants needed one labs internal data (not accessible to the other)
• Interpreting same variant for different associated conditions (different disease
frequencies…)
• Tabulation errors: codes did not support classification (use a classification tool)
EVALUATION OF INTER-LABORATORY CONCORDANCE
Reasons for Discrepancy • VUS criteria: conflicting pathogenic and benign lines of evidence
Some labs allowed one line of conflicting benign evidence of only a supporting level
(e.g., predictions) to override otherwise strong evidence of pathogenicity. Expert
judgment: disregarding limited conflicting evidence
• 2 discordant variants needed one labs internal data (not accessible to the other)
• Interpreting same variant for different associated conditions (different disease
frequencies…)
• Tabulation errors: codes did not support classification (use a classification tool)