Biomedical Engineering Reference
In-Depth Information
The third phase of analysis is the process of classifying the variants by utilizing
NGS software and a number of knowledge databases (Fig. 8.1 ). Examples of bio-
logical, genomic, and mutation databases include OMIM ( ) data-
base, and locus-specifi c databases (LSDBs;
html ) and HGMD. For inheritance fi ltering, the method uses simple genotype
comparison utilizing inheritance modeling between parents (plus other additional
family members) and the proband. Multiple inheritance models, including autoso-
mal-dominant, autosomal recessive, X-linked dominant, and X-linked recessive,
are performed on the genotypes of the family members. For the autosomal reces-
sive model, since there needs to be two affected alleles in the proband, variants that
are homozygous or compound heterozygous (bi-allelic) in proband that are inher-
ited from the parents are kept and move down the fi ltering pipeline. An advantage
of having exome data for family members is that variants can be easily focused by
this type of inheritance modeling fi lter. However, this requires data to be available
for each family member adding to the cost and sometimes parental samples may
not be readily available. Examples of commercially available software that per-
forms inheritance modeling are NextGENe from SoftGenetics and SVS Suite from
Golden Helix.
Inheritance Models for Autosomal Recessive Disorders:
A Detailed Example
Autosomal recessive inheritance models can be directly applied to the family
genotypes to fi lter out polymorphisms to obtain a smaller subset of potentially
signifi cant variants. It is possible that in the autosomal recessive model variants are
homozygous (both inherited), compound heterozygous (both inherited), a combina-
tion of an inherited and a de novo variant, or two de novo variants. In the homozy-
gous case, variants are heterozygous in parents and homozygous in the proband. To
see if these variants fi t this model, genotypes based on their genomic coordinates are
compared to the parent in the NextGENe software. In the compound heterozygous
model, the genotypes of parents can be easily sorted to further confi rm the trans
confi guration of alleles. In the de novo and inherited model, the fi ltering of variants
that are homozygous and compound heterozygous where one allele is inherited and
the other de novo is performed. This type of comparison is different from the others
in that genes found by de novo fi ltering must be compared to all the candidate vari-
ants. In the de novo case, parental genotypes are required to rule out transmission of
variants. In this case, parents and unaffected siblings of the affected individual are
negative for a specifi c genotype. Following variant annotation, candidate variants
are further prioritized by functional fi ltering, namely, mutation types, presence of
variants in mutation databases, frequency, predicted pathogenicity of variants, and
case-control comparisons to remove background errors and match phenotypic
description (Fig. 8.1 ) .
Search WWH ::

Custom Search