The issue may lie with the variant calling methodology rather than the sequencing. If the variant caller reports only genotypes that differ from the reference genome and calls variants independently for each individual, variants where all 400 individuals are homozygous for the reference allele will not produce any genotype calls. Similarly, if only a subset of individuals (e.g., 10) are heterozygous or homozygous for the alternate allele, the calls will only appear for those 10 individuals. This could explain the high number of missing genotypes and the limited distribution of genotyped variants across your dataset.

Call rate cut-offs can be as low as 80% but a 95% cut-off would be more typical and I would expect a call rate of closer to 99%.

Did you check the sequence quality? I would first go for 10 samples and check carefully the BAM file in a genome browser. I am assuming Illumina sequence.

Thank you for your insightful response! Beatrice A Mcgivney Your answer prompted me to revisit my data, and I observed the following trends:

I have 130 variants of a single gene, but only 16 variants show the presence of reference alleles in some individuals. The distribution is as follows:

• 4 variants showed the reference allele in only one individual
• 1 variant showed the reference allele in two individuals
• 1 variant showed the reference allele in five individuals
• 1 variant showed the reference allele in eight individuals
• 1 variant showed the reference allele in eighteen individuals
• 1 variant showed the reference allele in forty-two individuals
• 1 variant showed the reference allele in 134 individuals
• 1 variant showed the reference allele in 157 individuals
• 1 variant showed the reference allele in 162 individuals
• 2 variants showed the reference allele in 165 individuals
• 1 variant showed the reference allele in 167 individuals
• 1 variant showed the reference allele in 206 individuals

In summary, these 16 variants exhibit all three genotypes: homozygous reference (0/0), heterozygous (0/1), and homozygous alternate (1/1). However, the remaining majority of variants are observed only as heterozygous (0/1) or homozygous for the alternate allele (1/1), without any homozygous reference (0/0) calls.

Questions and Concerns:

I am currently uncertain about the best approach to filter significant variants out of my 130 total variants for downstream association analysis with phenotypic data. Since I need to focus on authentic variants to ensure true positive results, any guidance on data handling and filtering best practices would be greatly appreciated! Beatrice A Mcgivney