Why do I have * in my FASTA? g2gtools diploid

Question

[jon.klonowski]

Hi, I am using a bespoke program (g2gtools extract) that extracts sequences from a FASTA based on a GTF and the features you want (in my case I am extracting all transcripts). My FASTA is an entire genome of a diploid organism customized to include all SNPs and INDELS the organism carries.

When I run this program, it returns an error: "Sequence contains non-DNA character '*' at position 393"

At first, I was thinking that maybe these denote stop codons but when I investigate my FASTA, there are only 35. But upon second thought, that doesn't make much sense because its a genomic FASTA

Thanks in advance!

---

UPDATE TO THE ISSUE

Ok, so I found out the issue. I found out that the patient VCF has overlapping INDELs and, although they are not denoted like this within the VCF, when I extracted a set of them in a patient with bcftools query, the * appears. I looked at GATK and this is how overlapping INDELs are annotated: https://gatk.broadinstitute.org/hc/en-us/articles/360035531912-Spanning-or-overlapping-deletions-allele

My issues now what to do to prevent the reporting of overlapping INDELs in this way... I wonder if there is a way to prevent this annotation...

Original trio VCF - so 2 parents and the patient:

#CHROM  POS      REF     ALT     
chr1    154590147  CCG     C     
chr1    154590148  CG      C    
chr1    154590149  G       *      
chr1    154590149  G       C      

and then, if I just extract the patient:

#CHROM  POS         REF   ALT     GT  
chr1     154590148   CG  C      0|1 
chr1     154590149   G   *      1|0 
chr1     154590149   G   C      0|1 

after extracting the proband genotypes using bcftools query.

[GenoMax]

customized to include all SNPs and INDELS the organism carries.

Must be something related to this then.

[jon.klonowski]

I am using the same software for the entire pipeline - g2gtools -, so it would be weird if the output from a software is incompatible with that software....

[jon.klonowski]

Yea, youre right. g2gtools does introduce the into the fasta. I am just not sure what they mean. I am going to try to see what SNPs or INDELs align with the location of the in the FASTA

[dsull]

Can you show me what g2gtools commands you are running? I've been using g2gtools a bit in the past few weeks.

[jon.klonowski]

hey dsull ! Check out my update to the question please!

Answer 1

[Pierre Lindenbaum]

search for non ATGC in you reference.

grep -v '^>' in.fasta | tr -d 'ATGCatgcNn\n\r'

[jon.klonowski]

The output is: *YKY********R**RRRYWMWWYMKYYWYKKMYRRRWR***RYYBRRRRKWRS*RYY*RYWB***YYW*YKYRSKYWYYRYRYKYYMYYMRRRYWMWWYMKYYWYKKMYRRRWRRYYBRRRRKWRSRYSY*RYSYY*RYWYYKWM******MRRSKYWYYRYRYKYRYWYYKWMR****RYWBYY*WMRYRYYYMMRMMR*RWYRYRWYRY**YYRSWYYRSW

but the letters are all IUPAC standard and acceptable in a FASTA: https://www.bioinformatics.org/sms/iupac.html

would it not be the 35 '*', which is not IUPAC standard?

[Pierre Lindenbaum]

does g2tools support non ATGC chars ?

[jon.klonowski]

yes, but not asterisks. which is weird because it was their tool that introduced them in the first place :/ I am going to try to see what SNPs or INDELs align with the location of the *

[jon.klonowski]

hey Pierre Lindenbaum! Check out my update to the question please! Let me know if you have experience/thoughts.

[jon.klonowski]

I found your response to a similar question that seems to suggest that I can remove these variants. I need to think through whether I can still do this for a diploid (phased) organism... I am not sure if these spanning / overlapping indels MUST be on the same allele, or they are annotated as such even if they are not on the allele. Wouldnt removing the INDEL with the '*' be meaningful if the spanning/overlapping indels are on seperate alleles? bcftools norm resulting in '*' in alternate allele