The pipeline outputs a number of files for further analysis and exploration, as well as to provide an overview of the read preprocessing and distribution.

Report

report.html

This file is generated by Snakemake and outlines a lot of information relating to the Hecatomb run. Under the Results tabs are summary files for things like reads for each sample following different preprocessing steps as well as some summary plots.

SeqTable

hecatomb_out/RESULTS/seqtable.fasta

The SeqTable is the primary output of the read preprocessing and serves as the input for Taxonomic assignment. It is composed of all the representative sequences from the clustered reads for all samples. Samples are clustered individually, and the seq IDs for this fasta file follows the format >sampleID:count:seqNumber. Here, count is the number of reads in that cluster which is important for statistical exploration. Sequences are numbered sequentially (seqNumber) to ensure unique IDs.

BigTable

hecatomb_out/RESULTS/bigtable.tsv

The BigTable is the main output of Taxonomic assignment and can be directly imported into R or Python. The BigTable combines the seqtable IDs with their sampleID, counts, normalised counts, alignment information, taxonomic assignments and Baltimore classification. This file is big, hence the name, but is designed to make merging with sample metadata, plotting, and statistical interrogation as easy as possible.

The header looks like this:

seqID  sampleID  count  percent  alnType  targetID  evalue  pident  fident  nident  mismatches  qcov  tcov  qstart  qend  qlen  tstart  tend  tlen  alnlen  bits  targetName  taxMethod  kingdom  phylum  class  order  family  genus  species  baltimoreType  baltimoreGroup
Column header definitions seqID: Sequence ID (format = sampleID:count:uniqueInt)
sampleID: The sample IDs derived from the read files
count: The number of reads represented by the sequence
percent: Percent of the host-removed reads (normalised count)
alnType: Type of alignment (aa = amino acid, nt = nucleotide)
targetID: The UniProt or NCBI ID of the database target sequence
evalue: expect value of the alignment (less is better)
pident: percent identity (of the alignment)
fident: fraction identity
nident: number of identical bases/residues
mismatches: number of mismatched bases/residues
qcov: coverage of query sequence (query = seqtable sequence)
tcov: coverage of target sequence (target = database sequence)
qstart: query start position
qend: query end position
qlen: query sequence length
tstart: target start position
tend: target end position
tlen: target sequence length
alnlen: alignment length
bits: bit score (more is better)
targetName: target sequence name
taxMethod: Method used to assign taxonomy (either Lowest Common Ancestor, "LCA"; or top hit sequence, "topHit")
kingdom/phylum/class/order/family/genus/species: Taxonomy annotations
baltimoreType: Baltimore classification (double/single strand, DNA/RNA, +/-)
baltimoreGroup: Baltimore classification group


TaxonLevelCounts

hecatomb_report/taxonLevelCounts.tsv

This file is derived from the BigTable and summarises the total sequence counts, for each sample, at all taxonomic levels. The TaxonLevelCounts combines the sampleID with the taxonomic level for which the counts refer, the full taxonomic path, the taxon name, and the total and normalised read counts. The purpose of this file is to expedite statistical interrogation of your data. For instance, if you wanted to compare the numbers of say Flaviviridae reads between two groups of samples, those counts have already been collected, and you can simply run your analysis and plotting on the relevant slice of the table.

The file looks something like this:

sampleID    taxonLevel  taxonPath                                   taxonName       count   CPM
sample1     Kingdom     k_Bacteria                                  Bacteria        3162    3178.818
sample1     phylum      K_Viruses,p_Phixviricota                    Phixviricota    1216    1222.467
sample1     class       K_Viruses,p_Uroviricota,c_Caudoviricetes    Caudoviricetes  1234    1240.564
etc.

Assembly

hecatomb_out/RESULTS/assembly.fasta

These are the contigs generated, unless you run Hecatomb with the --skipAssembly flag. The assembly is used for producing the ContigSeqTable and ContigKrona plots, as well as the direct contig annotations.

hecatomb_out/RESULTS/contig_count_table.tsv

The contig count table contains the coverage information of all contigs for each sample.

Sample  Contig  Length  Reads  RPKM  FPKM  SPM  AverageFold  ReferenceGC  CoveragePercentage  CoverageBases  MedianFold
Column header definitions Sample: The sample IDs derived from the read files
Contig: The contig ID in assembly.fasta
RPKM: Reads Per Kilobase Million - see https://www.rna-seqblog.com/rpkm-fpkm-and-tpm-clearly-explained/
FPKM: Fragments Per Kilobase Million - see https://www.rna-seqblog.com/rpkm-fpkm-and-tpm-clearly-explained/
SPM: Sequences Per Million (counts normalized by library size)
AverageFold: Average read coverage of contig
ReferenceGC: Contig GC content
CoveragePercentage: Percent of contig covered by alignments
CoverageBases: Total bases of contig covered by alignments
MedianFold: Median read coverage of contig


Contig Annotations

hecatomb_out/RESULTS/contigAnnotations.tsv

The contig annotations follows a similar format to the bigtable. Refer to bigtable for column definitions.

contigID  evalue  pident  fident  nident  mismatch  qcov  tcov  qstart  qend  qlen  tstart  tend  tlen  alnlen  bits  target  kingdom  phylum  class  order  family  genus  species

ContigSeqTable

hecatomb_out/RESULTS/contigSeqTable.tsv

The ContigSeqTable combines the read mapping information for the assembly with the read-based taxonomic assignments. This file is intended to assist the user in identifying and binning assembly contigs by applying a consensus approach to contig taxonomic assignment. The file includes the positional mapping information and can also enable investigation of more complex features such as chimeric contigs, recombination or horizontal transfer events.

The header looks like this:

contigID  seqID  start  stop  len  qual  count  percent  alnType  taxMethod  kingdom  phylum  class  order  family  genus  species  baltimoreType  baltimoreGroup
Column header definitions contigID: Contid ID from assembly.fasta
seqID: Sequence ID (format = sampleID:count:uniqueInt)
start: Alignment start position on contig
stop: Alignment end position on contig
len: Alignment length
qual: Alignment quality
count: The number of reads represented by the sequence
percent: Percent of the host-removed reads (normalised count)
alnType: Type of alignment (aa = amino acid, nt = nucleotide)
taxMethod: Method used to assign taxonomy (either Lowest Common Ancestor, "LCA"; or top hit sequence, "topHit")
kingdom/phylum/class/order/family/genus/species: Taxonomy annotations
baltimoreType: Baltimore classification (double/single strand, DNA/RNA, +/-)
baltimoreGroup: Baltimore classification group


krona.html and contigKrona.html

hecatomb_report/krona.html

hecatomb_report/contigKrona.html

The Krona plots are to assist in visual exploration of the read annotations. krona.html is derived from the bigtable and shows the raw distribution of taxon assignments. contigKrona.html is derived from the contigSeqTable and includes the taxon assignment method (either tophit or LCA). The contigKrona plot helps to visualise the distributions of topHit versus LCA assigned reads as well as the distributions over contigs of the identified species, and the distribution of taxonomic assignments for each contig.