Hello BioStars Community,

I've been working with VCF files generated by the nf-core RNAvar pipeline, which employs the GATK best practices for RNA-seq variant calling. Upon reviewing the genotype information in my VCF files, I encountered a variety of genotype formats beyond the common ones I'm familiar with ("./.", "1/1", "0/1", "1/2").

Here are the unique genotype formats I extracted from my data:

# > unique_values
# [1] "./."   "1/1"   "0/1"   "1/2"   "2/2"   "0/2"   "2/1"   "0/3"   "1/3"   "2/3"   "3/2"   "3/1"   "3/3"   "3/4"   "0/4"   "4/3"   "1/4"   "2/4"   "4/2"   "4/4"  
# [21] "5/2"   "3/5"   "5/5"   "4/5"   "5/4"   "0/5"   "4/1"   "5/6"   "5/1"   "1/5"   "6/7"   "2/5"   "0/6"   "1/7"   "5/3"   "6/5"   "4/8"   "3/6"   "4/6"   "6/2"  
# [41] "0/7"   "4/7"   "6/1"   "6/6"   "6/3"   "0/8"   "7/7"   "2/7"   "1/6"   "7/2"   "9/7"   "2/6"   "5/8"   "7/5"   "8/2"   "6/4"   "3/7"   "0/9"   "9/3"   "7/4"  
# [61] "7/8"   "9/9"   "8/9"   "5/10"  "7/1"   "10/4"  "0/10"  "9/1"   "10/10" "2/8"   "9/2"   "8/5"   "7/3"   "8/8"   "5/7"   "8/3"   "10/3"  "3/8"   "9/5"   "1/8"  
# [81] "3/9"   "1/10"  "3/10"  "2/11"  "8/1"   "8/10"  "1/9"   "0/11"  "8/6"   "2/9"   "10/6"  "4/11"  "8/4"   "7/6"   "6/8"   "12/13" "9/10"  "11/10" "11/7"  "4/9"  
# [101] "7/9"   "0/14"  "6/9"   "11/12" "5/9"   "6/10"  "10/11" "12/1"  "12/12" "11/11" "2/10"  "9/11"  "0/13"  "8/7"   "5/11"  "10/1"  "0/12"  "2/12"  "9/8"   "10/2" 
# [121] "4/10"  "10/5"  "8/11"  "10/7"  "1/11"  "1/13"  "9/6"   "0/15"  "11/2"  "15/1"  "0/16"  "7/10"  "11/8"  "3/11"  "17/17" "12/2"  "4/12"  "8/16"  "11/9"  "13/13"
# [141] "0/18"  "0/17"  "11/3"  "13/8"  "6/11"  "11/5"  "10/9"  "11/14" "9/14"  "11/1"  "13/12" "1/12"  "13/14" "9/4"

The common genotypes ("./.", "1/1", "0/1", "1/2") make sense to me, representing missing data, homozygous reference, heterozygous, and homozygous variant calls, respectively. However, the presence of other formats such as "1/3", "2/4", or "5/5" is confusing, and I am unsure how to interpret these, especially considering the RNA-seq context and the GATK pipeline used.

I suspect these unusual genotypes might have originated from specific nuances in the RNA-seq variant calling process or might represent multi-allelic sites, but I'm not entirely sure. Before I proceed with filtering these genotypes or interpreting the variant calls, I'd appreciate any insights or recommendations on how to handle these genotypes.

Should I be concerned about the validity of these unusual genotypes, or are they expected outcomes given the RNA-seq variant calling methodology? Any advice on best practices for filtering or interpreting these would be greatly appreciated. Also this is human data so its a diploid. Thank you in advance for your help!