GATK3 error message

Question

[shuksi1984]

I ran following command:

java -Xmx4g -jar /path/to/gatk3.8.jar -T RealignerTargetCreator -R /path/to/Homo_sapiens_assembly38.fasta -I SP1_markduplicate.bam -o SP1_realignedtargets.list

I got following error message:

ERROR MESSAGE: Bad input: We encountered a non-standard non-IUPAC base in the provided reference: '92'

What does this mean? Did Homo_sapiens_assembly38.fasta got corrupted?

Answer 1

[Pierre Lindenbaum]

what is the version of gatk ?

what is the output of:

grep -v '^>' /path/to/Homo_sapiens_assembly38.fasta | tr -d '[ATGCNatgcn\n]' | grep -o '.' | sort | uniq -c

[shuksi1984]

what is the version of gatk ?

GATK3.8

Output of

grep -v '^>' /path/to/Homo_sapiens_assembly38.fasta | tr -d '[ATGCNatgcn\n]' | grep -o '.' | sort | uniq -c

is:

 2 \
  2 B
  1 F
  1 I
 15 K
  8 M
  2 O
 26 R
  4 S
 13 W
 33 Y

[Pierre Lindenbaum]

what devon said:

your fasta file is likely corrupted, since it has a \ in it as a base.

futhermore there is a 'O'.

I also suggest you have a look at this fasta file:

grep -n -A 5 -B 5 '/' ref.fasta

[shuksi1984]

Can you share the link to download hg38gatkbundle reference genome? I am trying with the given link:

ftp://[e-mail]/bundle/hg38/Homo_sapiens_assembly38.fasta.gz

[Pierre Lindenbaum]

there is no problem with that resource. your data is corrupted.

$ curl   "ftp://[e-mail]/bundle/hg38/Homo_sapiens_assembly38.fasta.gz"  | gunzip -c | grep -v '^>' | tr -d '[ATGCNatgcn\n]' | grep -o '.' | sort | uniq -c

      2 B
      8 K
      8 M
     26 R
      4 S
     13 W
     33 Y

[shuksi1984]

your data is corrupted

bam file is corrupted?

[Pierre Lindenbaum]

your data is corrupted

try to understand what I've done in my previous comment please C: GATK3 error message

[shuksi1984]

I ran the above command and got same output. I could not find any file in my working directory, except the output as shown above. I am unable to understand. Kindly, explain

Answer 2

[Devon Ryan]

Yes, your fasta file is likely corrupted, since it has a \ in it as a base.