Skip to content

Additional information

Freya Arthen edited this page Mar 3, 2022 · 17 revisions

Gene descriptors and PCA variables

The following variables are computed for each gene and the results are written to gene_table_taxon_assignment.csv. Any combination of these can be used in the PCA.

Contig related

c_name name of the contig the given gene is located on

  • Positional
    • c_num_of_genes number of genes annotated on the given contig
    • c_len contig length
    • c_pct_assembly_len percentage of assembly length: proportion of total assembly length contained in the given contig
    • c_genelenm mean gene length on the given contig
    • c_genelensd standard deviation of the gene lengths in the set of genes annotated on the given contig
  • Coverage
    • c_cov mean read coverage of the given contig
    • c_covsd standard deviation of read coverage along the given contig
    • c_covdev extent to which 'c_cov' deviates from the mean assembly coverage, in units of total assembly coverage SD
    • c_genecovm mean read coverage in the set of genes annotated on the given contig
    • c_genecovsd standard deviation of the read coverage in the set of genes annotated on the given contig
  • Sequence composition
    • c_pearson_r Pearson’s correlation coefficient for the tetranucleotide-derived z-score vectors of the given contig and the overall assembly
    • c_pearson_p probability (p-value) for two random sequences to show a correlation that is at least as high as 'c_pearson_r'
    • c_gc_cont percentage of GC content in the given contig
    • c_gcdev extent to which 'c_gc_cont' deviates from the mean assembly GC content, in units of total assembly GC SD

Gene related

g_name gene name (as given in the GFF file, 9th column, ID attribute)

  • Positional
    • g_len length of the given gene
    • g_lendev_c extent to which 'g_len' deviates from the corresponding c_genelenm, in units of 'c_genelensd'
    • g_lendev_o extent to which 'g_len' deviates from the average gene length in the assembly, in units of the length SD over the complete set of genes in this assembly
    • g_abspos absolute position of the gene (the closer to contig centre, the closer 'g_abpspos' to 0; the closer to either contig end, the closer g_abpspos to 1)
    • g_terminal classifier of terminal genes (1 if gene is terminal, 0 if not)
    • g_single classifier of single genes (1 if the given gene is the only gene annotated on its contig, 0 if not)
  • Coverage
    • g_cov mean read coverage of the given gene
    • g_covsd standard deviation of the read coverage along the given gene
    • g_covdev_c extent to which 'g_cov' deviates from the corresponding 'c_genecovm', in units of 'c_genecovsd'
    • g_covdev_o extent to which 'g_cov' deviates from the average gene coverage in the assembly, in units of the coverage SD over the complete set of genes in this assembly
  • Sequence composition
    • g_pearson_r_o Pearson’s correlation coefficient for the tetranucleotide-derived z-score vectors of the given gene and the overall assembly
    • g_pearson_p_o probability (p-value) of a random data set to show a correlation with the overall assembly composition that is at least as high as 'g_pearson_r_o'
    • g_pearson_r_c Pearson’s correlation coefficient for the tetranucleotide-derived z-score vectors of the given gene and its respective contig
    • g_pearson_p_c probability (p-value) for two random sequences to show a correlation that is at least as high as 'g_pearson_r_c'
    • g_gc_cont percentage of GC content in the given gene
    • g_gcdev_c extent to which 'g_gc_cont' deviates from the mean gene GC content, in units of GC SD over the set of genes on the corresponding contig
    • g_gcdev_o extent to which 'g_gc_cont' deviates from the mean GC content in overall gene set, in units of GC SD over the complete set of genes in this assembly

Taxonomic assignment information

  • protID (protein ID) header in FASTA file that was matched to gene ID and corresponding sequence is used for taxonomic assignment [header is truncated at first whitespace]
  • lcaID NCBI taxon ID of LCA of hits in 10% score range of best hit
  • lca name of taxon corresponding to 'lcaID'
  • best_hitID NCBI taxon ID of best hit
  • best_hit name of taxon corresponding to 'best_hitID'
  • bh_evalue e-value of best hit
  • bh_pident percentage of identical matches between best hit and query protein
  • corrected_lca corrected LCA: LCA of best hit and query species
  • taxon_assignment final taxonomic assignment: LCA if LCA is not in query's lineage, else the corrected LCA
  • plot_label final label for gene displayed in plot after merging of taxonomic assignments

Adding taxonomic assignment to GFF

To compare the taxonomic assignment of neighbouring genes using for examle a genome viewer, the taxonomic assignment can be added to the list of attributes of the genes in the GFF. For this, either call:

python add_ta2gff.py -c 'path/to/config.yml'

or

python add_ta2gff.py -g 'path/to/input.gff' -o 'path/to/modified_output.gff' -m 'path/to/taXaminer_report/'

The path stated with the parameter '-m' can either be

  • path to the directory of the taXaminer report (as in the example)
  • path to the 'taXaminer_report/taxonomic_assignment/gene_table_taxon_assignment.csv' file or the taXaminer report directory

GFF source

In taXaminer, you can select the source where gene and protein information will be picked from in the GFF. The following information is needed in the file containing the rule on how to parse gene and protein information from GFF file to retrieve the correct sequences and also match gene to protein features:

  • source: 'source' tag in GFF to consider (2nd field)
  • gene_tag: tag with which to identify genes (3rd field)
  • fasta_header_type: type of feature where to find the header of the proteins in the FASTA in (3rd field)
  • fasta_header_attr: attribute which corresponds to FASTA header of proteins
  • gene_connection: how connection between gene and FASTA header can be drawn; available options are:
    • inline: gene ID can be found in same feature as the FASTA header -> fill in 'gene_attr'
    • parent/child: connection can be drawn via Parent attribute or alike -> fill in 'parent_child_types' and 'parent_child_attr'
  • gene_attr: in which attribute in the line of 'fasta_header_type' can the gene ID be found (gene ID = attribute 'ID' in lines with 'gene_tag')
  • parent_child_types: which features to parse for the parent/child connection
  • parent_child_attr: which attribute is the 'child attribute' and which is the 'parent attribute'
    • state as dictionary like this: {parent: Parent, child: ID}

File setup: information each line by line, identifier and information separated by ':'

Clone this wiki locally