Home Software Services About Contact     
 
Follow on twitter

Robert C. Edgar on twitter

11-Aug-2018 New paper describes octave plots for visualizing alpha diversity.

12-Jun-2018 New paper shows that one in five taxonomy annotations in SILVA and Greengenes are wrong.

18-Apr-2018 New paper shows that taxonomy prediction accuracy is <50% for V4 sequences.

05-Oct-2017 PeerJ paper shows low accuracy of closed- and open-ref. QIIME OTUs.

22-Sep-2017 New paper shows 97% threshold is wrong, OTUs should be 99% full-length 16S, 100% for V4.

24-Nov-2016
UPARSE tutorial video posted on YouTube. Make OTUs from MiSeq reads.

 

USEARCH v11

Chimera detection benchmarks

See also
 
OTU benchmark home

Algorithm comparisons
  
UCHIME2 benchmark
   UCHIME vs. ChimeraSlayer
   UCHIME vs. Perseus (AmpliconNoise)
   UCHIME vs. DECIPHER
   UPARSE-OTU (cluster_otus) vs. UCHIME

Note
The test described in the ChimeraSlayer paper and used in the first UCHIME paper is misleading because it does not count false negatives as errors, but false negatives are equally important to false positives. The benchmark tests described in the UCHIME2 paper are more realistic.

Conclusions
The UPARSE-OTU algorithm (implemented in the cluster_otus command) is currently (mid-2018) the most effective chimera filter for 97% OTU clustering.

The UCHIME2-denoised-denovo algorithm used by UNOISE3 (as implemented in the unoise3 command) is superior to the DADA2 chimera filterring strategy and is currently (mid-2018) the most effective chimera filter for denoising (see results and discussion in UNOISE2 paper).

Image

Fig 1(a) from the UPARSE paper. Magenta segments in the pie chart show the fraction of chimeric OTUs generated by several algorithms on a mock community test. The QIIME and mothur pipelines used UCHIME for chimera detection. Notice that even after UCHIME filtering, about 50% of the OTUs generated by QIIME and mothur are chimeric, while UPARSE successfully filters all chimeras from Illumina reads and leaves only one chimeric OTU with 454 reads.
 

 


References (please cite)
R.C. Edgar et al. (2014), UCHIME improves sensitivity and speed of chimera detection, Bioinformatics 27(16) 2194-2200
  • Shows UCHIME faster and more accurate than ChimeraSlayer

  • This paper report misleading benchmark tests, see critique in UCHIME2 paper


R.C. Edgar (2016), UCHIME2: improved chimera prediction for amplicon sequencing, https://doi.org/10.1101/074252
  • UCHIME2 algorithm, improved chimera detection

  • "Fake" chimeras are common, valid biological sequences matching two-parent model

  • Perfect chimera filtering impossible even with complete and correct reference

  • Realistic chimera benchmark


R.C. Edgar (2016), UNOISE2: improved error-correction for Illumina 16S and ITS amplicon sequencing, https://doi.org/10.1101/081257
  • UNOISE2 algorithm, improved denoiser

  • Reduces false-positive chimeras compared to UNOISE and DADA2