TAXONOMY PROFILING

Microbiome research main questions:

• Who?
  • Corelations with environmental variables and time.
  • Experimental models can be used to confirm corelations.
• What is it doing?
  • Explore the functional potential can propose environment/host-microbe interactions and improve the correlational studies

Taxonomy profiling

• Comparison of the microbiome sequences properties against microbial reference genomes.
• Amplicon sequencing
  • Uses oligonucleotide probes to amplify specific genome regions. The conserved and hypervariable regions in the ribosome are used as primers.
    • 16S – prokaryotes
    • 18S – eukaryotic microbes
    • ITS – fungi
  • Advantages
    • Cheaper
    • Less susceptibility to contamination
    • Precise
    • Functional predictions based in reference genomes annotations
    • Better databases
  • Disadvantages
    • It reports a small set of organisms.
    • Functional potential of your samples will be ignored. Gene counts inaccessible.
    • Low confidence for taxonomic abundance estimation
    • Resolution: Genus
  • Shotgun sequencing
    • All the metagenomic DNA is sequenced.
    • Advantages
      • It detects bacteria, archaea, viruses, etc.
      • Accesses to the functional POTENTIAL in the microbiome
      • Resolution: Strains
      • Recovers single nucleotide variants
      • Better predictions of the taxonomic abundance
    • Disadvantages
      • Expensive
      • More susceptible to contamination
      • Less extensive databases: It is more likely to deplete.
    • Types of taxonomy profilers
      • Reads profiling: More confident representations of the DNA content. Better for fragments shorter than 1 kb
        • Homology-based
          • DNA: More restrictive than the protein programs.
            • Kraken: k-mer based classifier; reads input.
            • CLARK: k-mer based; DNA/RNA reads/scaffolds/contigs/etc input; *A single index per taxonomic rank. Lower memory/time consumption.
          • Protein: Mutational rate, and taxonomy bias in the incomplete databases makes tricky to ask very strict hits as the DNA programs. Useful in cross reference the reads with a predicted protein with the taxonomy classification of this read.
            • MEGAN: reads input. More precision values at higher time/classification rates. Requires previous alignment of your reads. Easy visualization.
            • Kaiju: reads input.
            • CARMA: reads input. Pfam based.
          • Marker-genes: Genes shared across domains, specific clade genes. Reduces classified seqs/false positives; faster.
            • MetaPhlAn
            • MetaPhyler
          • Composition-based: Wider perspective. Smaller db. Useful to classify complex environments. Lower memory requirements and time consumed.
            • PhyloPythia (SVM): Assembly input.
            • Phymm (IMMs): Reads input.
            • PhymmBL: accuracy improvement.
          • Assembled microbiome profiling
          • Heuristic solutions: use of k-mers. Used in different types of taxonomy profilers.
            • Short k-mer length: the precision will be reduced. It is more suitable if your microbiota is composed of distant taxa
            • Longer k-mer sizes: more restrictive
• What should I consider?
  • Resources
  • Expertise
  • How many samples do I have?
  • Do I have a complex environment?
  • Priority: Taxonomy or functional potential?
  • Expected resolution
  • Do I suspect taxonomically distant taxa in my samples?