Sequence variant interpretation is at the heart of delivering clinical genetic test results for hereditary diseases. The first widely adopted guidelines for variant interpretation came in 2015 from the American College of Medical Genetics and Genomics (ACMG) and the Association for Molecular Pathology (AMP). Since then, ClinGen has recommended many modifications to the ACMG/AMP guidelines and is working on a new version (V4), expected to be released in the coming years. A separate set of guidelines for the interpretation of copy number variants (CNV) was introduced by ACMG/ClinGen in 2020. Meanwhile, in the UK, the Association for Clinical Genomic Science (ACGS) developed its own guidelines in 2017, which are deeply rooted in the ACMG guidelines and updated biennially.

With the most recent publication of the ACGS 2024 variant classification guidelines, laboratories may wonder what the differences are between the ACGS and ACMG/AMP guidelines.

Below, we highlight some of the tangible differences that labs should be aware of.

Overall improvements

ACGS combines guidelines for SNVs, CNVs and intronic variants in a single document, while separate guidelines are available from ACMG/AMP/ClinGen. One benefit of this approach is that ACGS addresses scenarios regarding when to apply SNV versus CNV guidelines for intragenic CNVs.

For CNV guidelines, ACGS offers options for genes that either do not or cannot qualify as haploinsufficient, offering a path for interpreting CNVs for recessive disorders. For example, when a gene lacks evidence for haploinsufficiency, ACGS permits the use of evidence of loss of function mechanisms from G2P (absent gene product) or other compelling evidence.

ACGS guidelines explicitly address the mutually exclusive attributes. For example, using the same evidence for PM1, PM5 and PP2 should be avoided as it can result in circular evidence. Also, PS1 and PM4 should not be attributed together.

Attribute level

ACGS guidelines provide a comprehensive breakdown of attributes, offering examples to clarify their use.

PM2/PP3/BP4

One major difference in the ACGS guidelines is the retention of the moderate strength of the allele frequency attribute PM2, aligning with the original ACMG/AMP guidelines, rather than the current ClinGen recommendation to downgrade it to supporting strength. This decision is based on the maintenance of the computational scores—PP3/BP4—at supporting strength. It was thoroughly analyzed and concluded upon a mini-impact assessment by ACGS. Although ACGS applies these strengths to PM2 and PP3, it also indicates that this implementation is subject to the forthcoming ACMG/AMP V4 guidelines.

PS4

ACGS also provides clarity on case control study counts, which are not mentioned in ACMG/AMP guidelines and only available for a few gene-specific guidelines from ClinGen. It recommends attributing PS4 at a supporting level even when one proband is reported for the variant and at a moderate level when two or more probands are reported with the appropriate phenotype. The attribute can be used at a stronger strength when the variant has an odds ratio. Odds ratios can also be calculated with cases available in a paper, and gnomAD data can be used as controls. ACGS also provides the flexibility to attribute PS4 in the absence of PM2 when a considerable number of probands are reported for the variant.

PM4

ACMG/AMP uses the PM4 attribute for all stop-loss variants that disrupt protein function, whereas ACGS aptly suggests attributing stop-loss variants using either PM4 or PVS1. For variants undergoing Non-stop mediated decay (NSD), in the absence of an in-frame termination codon in the 3’UTR, ACGS recommends using PVS1 at a very strong strength. This exhibits strong logical consistency. Conversely, the stop-loss not undergoing NSD can be attributed PM4.

BP7

The ACMG guidelines for BP7 consider silent variants for which splicing prediction algorithms predict no impact on splicing. The ACGS guidelines explicitly mention the application of BP7 to different types of silent variants: synonymous, UTR and intronic variants. In addition to using algorithmic prediction to evaluate potential impact on splicing (using BP7 at a supporting level when there is no evidence of splice impact), ACGS allows the use of laboratory evidence of no splice impact from RNA splicing assays to invoke BP7 at a strong level.

Scoring

ACGS has already shifted to a point-based system for the calculation of final pathogenicity, while current ACMG/AMP guidelines still utilize combinations of supporting/moderate/strong/very strong attributes (Note: V4 of ACMG/AMP guidelines will be based on a point-based system as well).

One advantage of this is that the point-based system provides a straightforward way of dealing with conflicting evidence, which the ACMG/AMP 2015 guidelines do not adequately address.

Because of the different scoring mechanisms, the final pathogenicity might differ when ACGS guidelines are followed versus ACMG/AMP guidelines, but these differences may disappear once ACMG/AMP V4 guidelines come out. For example, it is much easier to reach a classification of Likely Benign with the point-based system. The equivalent of one supporting attribute will get you there with the point-based system, whereas two supporting attributes are needed in the current ACMG/AMP guidelines.

Conclusion

In short, we find that the ACGS variant interpretation guidelines differ in only a few meaningful ways from ACMG/AMP guidelines and they provide some much-needed clarification on the application of different attributes. It is likely that ACGS and ACMG/AMP V4 will be even more closely aligned.

In the meantime, while it is tempting to combine the best of both guidelines when drafting your own SOPs for variant interpretation, caution should be exercised as modifications to some attributes are balanced by modifications to other attributes.

The ACGS guidelines represent a significant step forward in variant interpretation, promoting consistency, accuracy, and uniformity among variant scientists. These guidelines not only help to streamline interpretation processes but also facilitate good clinical practice, ultimately benefiting patient care and outcomes.

If you have any questions or if you would like to know more, please contact us at info@zifornd.com.