I'm running variant effect predictor on VCF with 100 individuals. Is there a build in way to annotate each sample separately and make stats out of it, e.g. distribution of number of pathogenic variants in the population of individuals?
VEP annotates variants; not individuals, so you'll be getting the same variant-level annotations. It's more efficient to annotate the vcf; then split it out per sample for statistics -- or you could use bcftools query in combination with bcftools stats -s <sample_id> to compute such counts without needing to split out individual VCFs. (Or: Use a vcf library in your language of choice for fine control of filters)