The canu progress chatter records statistics such as an input read histogram, corrected read histogram, and overlap types.
The canu progress chatter records statistics such as an input read histogram, corrected read histogram, and overlap types. The layout provides information on where each read ended up in the final assembly. The `GFA <http://lh3.github.io/2014/07/19/a-proposal-of-the-grapical-fragment-assembly-format/>`_ is the assembly graph generated by Canu. The fasta output is split into three types:
- contigs: everything which could be assembled and is part of the primary assembly. This includes both unique and repetitive elements
- bubbles: alternate paths in the graph which could not be merged into the primary assembly.
- unassembled: reads which could not be incorporated into the primary or bubble assemblies.
Correct, Trim and Assemble, Manually
...
...
@@ -178,24 +185,29 @@ After the run completes, we can check the assembly statistics::
lenSuggestRepeat sum 1135952 (genomeSize 12100000)
lenSuggestRepeat num 159
lenSuggestRepeat ave 7144
lenUnassembled ng10 12635 bp lg10 76 sum 1220193 bp
lenUnassembled ng20 9372 bp lg20 188 sum 2424217 bp
lenUnassembled ng30 7287 bp lg30 333 sum 3632625 bp
lenUnassembled ng40 4941 bp lg40 534 sum 4841897 bp
lenUnassembled ng50 2069 bp lg50 883 sum 6050798 bp
lenUnassembled sum 6321159 (genomeSize 12100000)
lenUnassembled num 1061
lenUnassembled ave 5957
lenContig ng10 761684 bp lg10 2 sum 1544923 bp
lenContig ng20 667922 bp lg20 4 sum 2942290 bp
lenContig ng30 567305 bp lg30 6 sum 4156720 bp
lenContig ng40 550087 bp lg40 8 sum 5271140 bp
lenContig ng50 446812 bp lg50 10 sum 6197825 bp
lenContig ng60 251216 bp lg60 14 sum 7356841 bp
lenContig ng70 183999 bp lg70 20 sum 8624087 bp
lenContig ng80 120025 bp lg80 28 sum 9744795 bp
lenContig ng90 84512 bp lg90 40 sum 10907525 bp
lenContig sum 11922888 (genomeSize 12100000)
lenContig num 67
lenContig ave 177953
Consensus Accuracy
-------------------
...
...
@@ -203,9 +215,25 @@ While Canu corrects sequences and has 99% identity or greater with PacBio or Nan
If you have Illumina sequences available, `Pilon <http://www.broadinstitute.org/software/pilon/>`_ can also be used to polish either PacBio or Oxford Nanopore assemblies.
Changes
-------------------
- Support for reads up to 2Mbp in size (up from 130Kbp).
- Incorporate MHAP 2.0 which is 3-5X faster than previous version and has higher specificity
- Improve diploid-aware assembly by categorizing output as primary contigs or unmerged bubbles. Annotate repeat and unique contigs in the output.
- Enable parallel overlap store construction on large genomes
- Enable `minimap <https://github.com/lh3/minimap>`_ as an option for generating overlaps during correction step. Corrected reads are generated as before with falcon_sense.
- Fix bug using shorter rather than longer reads for corrected reads/consensus computation
- Fix bug resuming without providing input sequences which would incorrectly set error rates
- Fix bug in bogart which would demote contained sequences as spurs incorrectly
- Fix bugs in falcon_sense which would hang when input had N bases and limit corrected reads to 65Kbp
Known Issues
-------------------
- Bogart (unitigger) has false positives in repeat breaking. Currently, the workaround is to increase the minimum overlap size to avoid detecting false repeats ca
used by short overlaps. Canu will automatically do this for large (>100MB) genomes.
- LSF support has limited testing
- Large memory usage while unitig consensus calling on unitigs over 100MB in size (140Mb contig uses approximate 75GB).
- Distributed file systems (such as GPFS) causes issues with memory mapped files, slowing down parts of Canu, including meryl (0-mercounts) and falcon-sense (2-correction).
