Pathogenicity/Causality of Variants

Damaging effect (impairment) of a variant

First, we should differentiate the damaging effect of a variant from the pathogenicity/causality of the variant.

Variants that cause gross protein malformations are here defined as destructive variants, including truncating variants (nonsense and frameshift), canonical splice-site variants, and variants at initiation codon or with single/multi exon deletion. Destructive variants mainly lead to complete loss of function and haploinsufficiency, i.e., null variants. In contrast, missense variants and those with indel/insert result in variable damaging effect, which are generally evaluated or predicted by functional studies or in silico tools. Damaging effects are associated with pathogenicity of variants. However, due to the differences in GDQ, variants with the same damaging effect present different pathogenicity in different genes.

Pathogenicity of a variant

Standards and guidelines for the interpretation of sequence variants by the American College of Medical Genetics and Genomics (ACMG) provided a general strategy for evaluating the pathogenicity of variant, based on evaluation of the functional impairment and indirect evidence from other sources such as population data, inheritance information, and previous reports.

Theoretically, in genes with pathogenic mechanism of LOF, the pathogenicity of a variant is determined by the impairment caused by the variant and the GDQ of the gene. The pathogenicity of a variant could be defined as the damaged GDQ that potentially produces abnormal phenotype (Figure a). However, only in a limited number of genes GDQs have been quantitatively determined (109 genes in this database); and the functional impairment could not been quantitatively defined in many cases.

Based on the information of KO and clinical-genetics, GDQ of genes have been grossly divided into high, middle-high, middle (~50%), middle-low, and low dependence (Fig a-e), which would provide a general criterion for evaluating the pathogenicity of new identified variant (s).


In genes featured by high dependence (Fig a), monoallelic variants of haploinsufficiency are pathogenic and potentially associated with relatively severe phenotypes and sporadic cases (de novo variants); while variants of less functional impairments are potentially associated with milder phenotypes that may be co-segregated in families. In genes featured by middle-high dependence (Fig b), monoallelic variants of haploinsufficiency are potentially pathogenic but may be associated with relatively mild phenotypes; whereas missense variants with less damaging effect may not be pathogenic. Biallelic variants are potentially associated with severe phenotypes. In genes featured by middle-dependence (Fig c), monoallelic variants of haploinsufficiency cause susceptibility to diseases, whereas biallelic variants are pathogenic. In genes featured by middle-low dependence (Fig d), biallelic variants are pathogenic, and one of the variants may be less destructive, such as variants in deep intronic regions. In genes featured by low dependence (Fig e), biallelic variants with severe function deficiency are required to cause the diseases.

In genes with specific pathogenic mechanisms, such as GOF and toxic effect, the pathogenicity of variants should be evaluated with considerations on their pathogenic mechanism.

Causality of a variant

In practice, the question is whether the identified variant is the cause of disease (abnormal phenotype) of the individual (patient), i.e., estimating the causality of the variants, which is defined as the possibility of a variant as the cause of abnormal phenotype in the patient. Apparently, the clinical manifestation of the patient should be an important consideration.

We propose a protocol to evaluate the causality of variants by combing the variants information with clinical concordance, i.e., comparing the clinic-genetic features of the patient with the pathogenic feature of variants of the gene. The pathogenic features of the genes were summarized in this database, including phenotype specificity, pathogenic geno/funo-type or mechanism, correlation between genetic impairment and phenotype severity, and inheritance pattern, aimed to provide individualized criterion for evaluating the causality of variants in each gene.

Variants sorting and filtration are required before clinical concordance evaluation. Similar to ACMG, allelic data (inheritance) and minor allele frequency in populations are considered in sorting and filtration.

Flow chart of evaluation of causality of variants