HGT_analysis

Documentation for reproducing analysis of the Quercus virginiana and Belonocnema kinseyi genomes for potential HGT to each other and from bacteria and fungi

Software Requirements

Conda Environment

The analysis requires several bioinformatics tools that are managed through a conda environment. To recreate the analysis environment: conda env create -f environment.yml conda activate genomics

Data sources:

Quercus virginiana (HAP1; Q_virginiana_protein_prefixed.faa; provided privately)

Belonocnema kinseyi (Build: GCF_010883055.1/B_treatae_v1)

1. Insect References:

   • Apis mellifera (Build: GCF_003254395.2/Amel_HAv3.1)
   • Nasonia vitripennis (Build: GCF_009193385.2/Nvit_psr_1.1)

2. Plant References:

   • Arabidopsis thaliana (Build: GCF_000001735.4/TAIR10.1)
   • Populus trichocarpa (Build: GCF_000002775.5/P.trichocarpa_v4.1)

3. Bacterial data:

• NCBI RefSeq bacterial protein sequences
• Source: ftp://ftp.ncbi.nlm.nih.gov/refseq/release/bacteria/ (sourced on September 12, 2024)
• Database created using makeblastdb with protein sequence type (-dbtype prot)

4. Fungi:

• NCBI RefSeq fungal protein sequences
• Source: ftp.ncbi.nlm.nih.gov/refseq/release/fungi/fungi.*.protein.faa.gz (sourced on September 12, 2024)
• Database created using makeblastdb with protein sequence type (-dbtype prot)

Scripts:

Prep data:

Prep data file contains code to be run in the console to acquire, unzip, concatenate, and make BLAST dbs for each dataset

Run BLASTP:

wasp_to_reference.sh: blasts the B kinseyi proteins against Q virginiana, N vitripennis, and A mellifera sequences

oak_to_reference.sh: blasts the Q virginiana proteins against B kinseyi, P trichocarpa, and A thaliana sequences

blastp_wasp_vs_bact.sh: blasts the B kinseyi proteins against N vitripennis, A mellifera, and the combined bacterial sequences

blastp_oak_bact.sh: blasts the Q virginiana proteins against P trichocarpa, A thaliana, and the combined bacterial sequences

blastp_wasp_fungal.sh: blasts the B kinseyi proteins against N vitripennis, A mellifera, and the combined fungal sequences

blastp_fungal.sh: blasts the Q virginiana proteins against P trichocarpa, A thaliana, and the combined fungal sequences

Analysis:

HGT Candidate Filtering Criteria:

• Minimum identity: 75% (eliminate low-similarity matches likely to be spurious)
• Maximum own-lineage identity: 50% (remove matches that are highly similar to phylogenetic neighbors)
• Minimum difference: 20% (eliminate proteins that are very similar to both potential donor and to neighbors)
• Maximum identity: 90% (to exclude universally conserved proteins)

Analyze Results

filter_bkins_v_oak.py

Description: Finds B. kinseyi proteins for which identity to an oak protein is significantly higher than to Nasonia or Apis.

• Result: No candidates.

---

filter_qvirg_v_bkins.py

Description: Finds Q. virginiana proteins for which identity to a B. kinseyi protein is significantly higher than to Populus or Arabidopsis.

• Result: No candidates.

---

filter_fungal_hgt.py

Description: Finds Q. virginiana proteins for which identity to a fungal protein is significantly higher than to Populus or Arabidopsis.

• Result: 1 potential HGT candidate:

Oak_protein	Best_fungal_hit	Fungal_identity	Best_plant_hit	Plant_identity	Difference
Qv_Qvirginiana.HAP1.v1.g2518.t1	XP_024712885.1	76.5	Pt_XP_052308549.1	42.2	34.2

This is protoheme IX farnesyltransferase, mitochondrial [Candidozyma pseudohaemuli].

• Assessment: Not a likely candidate for genuine HGT.

---

filter_wasp_fungal_hgt.py

Description: Finds B. kinseyi proteins for which identity to a fungal protein is significantly higher than to Nasonia or Apis.

• Result: 1 potential HGT candidate:

Query_protein	Best_fungal_hit	Fungal_identity	Best_insect_hit	Insect_identity	Difference
Bk_XP_033211786.1	XP_003671371.2	77.8	Nv_XP_031779649.1	48.1	29.7

This is hypothetical protein NDAI_0G03510 [Naumovozyma dairenensis CBS 421].

• Assessment: Not a likely candidate for genuine HGT.

---

filter_oak_bact_hgt.py

Description: Finds Q. virginiana proteins for which identity to a bacterial protein is significantly higher than to Populus or Arabidopsis.

• Result: (BLAST outputs were deleted but there were no candidates.)

---

filter_wasp_hgt.py

Description: Finds B. kinseyi proteins for which identity to a bacterial protein is significantly higher than to Nasonia or Apis.

• Result: (BLAST outputs were deleted but there were no candidates.)

kmer analysis

check_fasta_format.py - look for issues in the formatting of the genomes that will be an obstacle to kmerizing them conda install -n genomics biopython - install biopython clean_genome.py - write cleanup script clean_genomes.slurm - run cleanup script conda install -c bioconda kmc bwa samtools - install kmc and bwa and samtools

Create directories for k-mer analysis

mkdir -p /export/martinsons/adam/kmer_analysis/{kmers,shared,mapped} mkdir -p /export/martinsons/adam/kmer_analysis/temp # KMC needs a temp directory

using a kmer length of 31 run_kmer_analysis.slurm - reciprocally checks for kmer matches

This found: Q. virginiana coverage: Average coverage: 0.000951613 Regions with coverage: 597155

B. kinseyi coverage: Average coverage: 0.000481438 Regions with coverage: 470286

conda install -c bioconda pysam - install pysam

find_clusters.py - write script to find overlapping matches find_clusters.slurm - run script to find overlapping matches

conda install -c bioconda bedtools - install bedtools

extract_and_blast.py - script to pull out the nucleotide sequences associated with the top matches

Results: Lots of highly repetitive sequences in both directions, which can't be BLASTed and are not likely to be HGT. One high match density region in the B kinseyi genome turns out to be an insect LSU rRNA sequence. Also not likely to be HGT.