Skip to content

Output format

Jump to bottom
Jim Shaw edited this page Dec 13, 2024 · 13 revisions

TSV Output

Sylph outputs a TSV (tab-separated values) file. Each row is one genome detected in the metagenome sample.

Sample_file   Genome_file   Taxonomic_abundance   Sequence_abundance   Adjusted_ANI   Eff_cov   ANI_5-95_percentile   Eff_lambda   Lambda_5-95_percentile   Median_cov   Mean_cov_geq1   Containment_ind   Naive_ANI   Contig_name
reads.fq   genome.fa   78.1242   81.8234   97.53   264.000   NA-NA   HIGH   NA-NA   264   264.143   10281/22299   97.53   NC_016901.1 Shewanella baltica OS678, complete genome
  • Sample_file: the filename of the reads/sample.
  • Genome_file: the filename of the detected genome.
  • Taxonomic_abundance: normalized taxonomic abundance as a percentage. Coverage-normalized - same as MetaPhlAn abundance
    • Not present for sylph query
  • Sequence_abundance: normalized sequence abundance as a percentage. The "percentage of reads" assigned to each genome - same as Kraken abundance
    • Not present for sylph query
  • Adjusted_ANI: adjusted containment ANI estimate.
    • If coverage adjustment is possible (cov is < 3x cov): returns coverage-adjusted ANI
    • If coverage is too low/high: returns Naive_ANI (see below)
  • Eff_cov/True_cov: an estimate of the effective, or if -u specified, the true coverage. Always a decimal number.
  • ANI_5-95_percentile: [5%,95%] confidence intervals. Not always a decimal number.
    • If coverage adjustment is possible: float-float e.g. 98.52-99.55
    • If coverage is too low/high: NA-NA is given.
  • Eff_lambda: estimate of the effective coverage parameter. Not always a decimal number.
    • If coverage adjustment is possible: lambda estimate is given
    • If coverage is too low/high: LOW or HIGH is output
  • Lambda_5-95_percentile: [5%, 95%] confidence intervals for lambda. Same format rules as ANI_5-95_percentile.
  • Median_cov: median k-mer multiplicity for k-mers with >= 1 multiplicity.
  • Mean_cov_geq1: mean k-mer multiplicity for k-mers with >= 1 multiplicity.
  • Containment_ind: int/int showing the containment index (number of k-mers found in sample divided by total k-mers), e.g. 959/1053.
  • Naive_ANI: containment ANI without coverage adjustment.
  • kmers_reassigned: the number of k-mers reassigned away from the genome.
    • Not present for sylph query
  • Contig_name: name of the first contig in the genome (or just the contig name for the -i option).

How do I get a taxonomic profile like MetaPhlAn?

See the manual outlined here.

Note on containment ANI

In sylph, ANI implicitly means containment ANI between a genome and a metagenome.

Containment ANI is calculated from the number of k-mers in a reference genome contained in a metagenome.

  1. If the "metagenome" is a single genome, the containment ANI approximates the standard ANI.
  2. If the "metagenome" is a collection of genomes, the containment ANI can be interpreted as a "nearest neighbour ANI".
  3. (profiling): If the metagenome is reads from a collection of genomes, sylph estimates an adjusted (containment) ANI that is the same as case 2 with a statistical model.

Note: containment ANI slightly overestimates the true ANI. See Supplementary Figures in our paper.

Clone this wiki locally