Searching for Capsids in Venom Proteomes

This repo documents searching for viral capsid proteins in venom transcriptomes. This repository is associated with the pub Identification of capsid-like proteins in venomous and parasitic animals. To use the scripts in this repo, create an environment with:

conda env create -n capsid_searches -f environment.yml

The most updated set of results are in 2023-06-01-updated-vog-results where VOG capsid HMMs were run against the entire venom/tick database and additional tick proteins from the TSA. The specific capsid VOGs from release 217 are listed in metadata/Updated_capsid_VOGs_v217.txt.

Searching for viral capsid proteins

Run capsid HMMs against sets of venomous species proteins with python3 scripts/run_parse_hmms.py dbs/capsid_hmm.hmm proteins/all_proteins/all_venom_proteins.fasta --evalue 1e0 with no evalue cutoff. The results are in results/2023-04-11_capsid_v1/ where I experimented with some different evalue cutoffs before settling with no cutoffs.

To combine multiple FASTA files into one FASTA file where the filename is carried through to the header (important if you want to retain the accession or species number for example), use the script reorganize_fastas.py.

Searching for a specific viral capsid protein from Bracovirus

We were specifically interested in a Bracovirus capsid protein that wasn't being captured from the previous HMM search, and found that for whatever reason this particular capsid wasn't included in the HMM above. We found the protein sequence for this capsid, BLASTed it in NCBI, and downloaded the hits and turned it into an HMM.

Then we created an alignment of the multi-FASTA file of BLAST results to the capsid protein with muscle -align vp39_psiBLAST_hits.fasta -output vp39_psiBLAST_hits.aln then build an HMM profile with hmmbuild vp39_capsid.hmm vp39_psiBLAST_hits.aln. This was used to run hmmsearch against the venom proteins using the run_parse_hmms.py script with: python3 ../scripts/run_parse_hmms.py vp39_capsid.hmm ../proteins/all_proteins/all_venom_proteins.fasta --evalue 1e0 --output_file ../results/vp39_capsid_summary.txt. The files for creating this vp39 HMM and the HMM are in hmms.

BLAST searches for Amblyomma americanum

BLAST should already be installed in the environment you created above. First download the tick genome:

wget -O Arcadia_Amblyomma_americanum_asm001_purged_cleanedup1.fasta https://zenodo.org/record/7783368/files/Arcadia_Amblyomma_americanum_asm001_purged_cleanedup1.fasta\?download\=1

Then extract protein sequences of interest from the pangenome that was created by the Arcadia-Science/rehgt pipeline:

for acc in EEC17452.1 KAG0427517.1 KAG0443635.1 EEC12039.1 KAG0444350.1 KAG0420414.1 XP_042148722.1
do
grep -A 1 ${acc} ../../outputs/genus_pangenome_clustered/Ixodes_cds_rep_seq.fasta >> query_cds.fasta
done

Make a BLAST db of the genome and run blast:

makeblastdb -in Arcadia_Amblyomma_americanum_asm001_purged_cleanedup1.fasta -dbtype nucl -out Arcadia_Amblyomma_americanum_asm001_purged_cleanedup1

 blastn -task megablast -db Arcadia_Amblyomma_americanum_asm001_purged_cleanedup1 -query query_cds.fasta -dust no -max_target_seqs 100 -outfmt 6 -out outputfile.txt

And do the same for proteins:

for acc in EEC17452.1 KAG0427517.1 KAG0443635.1 EEC12039.1 KAG0444350.1 KAG0420414.1 XP_042148722.1
do
grep -A 1 ${acc} ../../outputs/genus_pangenome_clustered/Ixodes_aa_rep_seq.fasta >> query_aa.fasta
done

tblastn -db Arcadia_Amblyomma_americanum_asm001_purged_cleanedup1 -query query_aa.fasta -max_target_seqs 100 -outfmt 6 -out outputfile_tblastn.txt

The queries and results files are in results/2023-04-20-blast-results.

Structural clustering to identify capsids in Ornithodoros turicata

In order to compare structures of tick proteins to viral capsid proteins, we ran ProteinCartography using v0.4.2.

First, we cloned the ProteinCartography repository and followed the README.md to install the conda environments. We used the notebook in this repository, notebooks/1_prep_metadata.ipynb to fetch UniProt metadata for the Ornithodoros proteins in metadata/ornithodoros.txt, which we created by searching UniProt for all Ornithodoros proteins. This notebook should be ran using the envs/web_apis.yml environment from the ProteinCartography repository.

We then used the notebook in this repository, notebooks/2_get_alphafold_structures.ipynb to fetch predicted structures from the AlphaFold database using the list of proteins in metadata/ornithodoros.txt. This notebook should also be ran using the envs/web_apis.yml environment from the ProteinCartography repository. We folded all of the proteins from the Virus Orthologous Groups Database (VOG database, version 217) that contained "capsid" in the name of the group using ESMFold.

Finally, we ran ProteinCartography using the config_ff.yml that can be found in the notebooks folder of this repository. The pipeline was initiated from inside the ProteinCartography repository with the envs/cartography_tidy environment activated using the following command:

snakemake --snakefile Snakefile_ff --configfile ../capsids/metadata/config_ff.yml --use-conda --cores 4

We then used the notebook in this repository, notebooks/3_plotting_overlays.ipynb to create custom plotting overlays for the UMAP. This notebook should be ran using the envs/plotting.yml environment from the ProteinCartography repository.

The results from this analysis can be found in the Zenodo repo.

Contributing

If you have suggestions or encounter any problems with the workflow or associated scripts please file an issue. You are also welcome to submit a pull request if you wish to dive in to the code. See this guide to see how we recognize feedback and contributions on our code.