The problem of generating chain files using cactus

[jihui-boop]

Dear Professor, when I was running the script, I encountered an issue where the chromosome length exceeded the calculated range, causing the script to fail in generating the chain file. Could you please help me with this problem? I sincerely hope you can take the time to offer guidance. Thank you, Professor.

03_cactus_55549295.txt

toga-cactus.txt

[DustinSokolowski]

Hello!

I'm sorry for the slow reply, for some reason I didn't see this come up on my e-mail. Just to get oriented here, did the cactus alignment run successfully? In other words do you see a hal file that is populated with alignments?

If so, I noticed that you have this line echo "Running Cactus on primary chromosomes only..."

You then have an if statement in your hal2fasta line where it looks like you'll only extract the fasta file from the hal file if the original fasta's don't yet exist:

echo "Extracting FASTA from HAL..."

if [ ! -f "${CACTUS_DIR}/${REF}.fa" ]; then
    singularity exec \
      --bind "${BASE},${EXTERNAL_DIR}" \
      "${CACTUS_SIF}" \
      hal2fasta "${HAL_FILE}" "${REF}" > "${CACTUS_DIR}/${REF}.fa"
fi

if [ ! -f "${CACTUS_DIR}/${TARGET}.fa" ]; then
    singularity exec \
      --bind "${BASE},${EXTERNAL_DIR}" \
      "${CACTUS_SIF}" \
      hal2fasta "${HAL_FILE}" "${TARGET}" > "${CACTUS_DIR}/${TARGET}.fa"
fi

I'm not totally certain what the primary chromosomes are, but you need an exact match in chromosomes and chromosome lengths between ${TARGET}.fa//${REF}.fa and whatever you inputted into cactus. If these files include extra information other than "primary chromosomes" (e.g. unplaced contigs), I could see your error being thrown.

One way to test this would be to get the chromsome sizes pulled out from the fasta files generated by your fasta file and compare them to the chromosome sizes of the ref/target fasta's that you are trying to annotate to make sure you don't have any inconsistencies.

Other less likely but possible options off the top of my head is that I don't know if cactus would be good with a nucleotide alphabet that is not just ATCGN, so if you have any of the IUPAC extended code or something like a "." or "*" instead of an N to denote a gap or "any base" then cactus may throw weird errors. While I haven't experienced this with cactus, TOGA also doesn't like having unusual characters in the fasta headings (e.g., "|"). These headings can cause some unusual behaviour.

Best!
Dustin