方案1: 不纠错直接组装。输入数据分为两种情况，一种是用所有的原始数据（包括长度过滤2K/5K以下），一种则是选择最长的30X, 40X作为输入。

• miniasm: https://github.com/lh3/miniasm

• WTDBG2: https://github.com/ruanjue/wtdbg2

• SMARTdenovo: https://github.com/ruanjue/smartdenovo

• RA: https://github.com/lbcb-sci/ra，个人对这个工具极其失望，运行速度慢，内存管理差，中断之后会删除所有中间文件。

• Flye: - https://github.com/fenderglass/Flye

  注: 不纠错直接组装得到的contig，需要先进行三代自纠错，才能用二代纠错，否则比对率会比较低。

方案2: 纠错后组装，然后挑选所有数据或者最长的30X, 40X作为输入。可供选择的纠错工具如下

• NextDenovo: https://github.com/Nextomics/NextDenovo

• Canu: https://github.com/marbl/canu

• MECAT: https://github.com/xiaochuanle/MECAT

• NECAT: https://github.com/xiaochuanle/NECAT

注: NECAT, Canu可以从头跑到尾，其中NECAT是MECAT的继任者，专门处理Nanopore组装。

Examples

### 对于PacBio data：

## wget https://zenodo.org/record/1172816/files/E.coli_PacBio_40x.fasta

## flye --pacbio-raw E.coli_PacBio_40x.fasta --out-dir out_pacbio --genome-size 5m --threads 4

Parameter descriptions

usage: flye (--pacbio-raw | --pacbio-corr | --nano-raw |

             --nano-corr | --subassemblies) file1 [file_2 ...]

             --genome-size SIZE --out-dir PATH

             [--threads int] [--iterations int] [--min-overlap int]

             [--meta] [--plasmids] [--no-trestle] [--debug]

             [--version] [--help] [--resume]


Assembly of long and error-prone reads

optional arguments:

  -h, --help            show this help message and exit

  --pacbio-raw path [path ...]
PacBio raw reads

  --pacbio-corr path [path ...]
PacBio corrected reads

  --nano-raw path [path ...]
ONT raw reads

  --nano-corr path [path ...]
ONT corrected reads

  --subassemblies path [path ...]
 high-quality contigs input

  -g size, --genome-size size
    estimated genome size (for example, 5m or 2.6g)

  -o path, --out-dir path
Output directory

  -t int, --threads int
number of parallel threads [1]

  -i int, --iterations int
number of polishing iterations [1]

  -m int, --min-overlap int
minimum overlap between reads [auto]

  --asm-coverage int    reduced coverage for initial contig assembly [not set]

  --plasmids            rescue short unassmebled plasmids

  --meta                metagenome / uneven coverage mode

  --no-trestle          skip Trestle stage

  --polish-target path  run polisher on the target sequence

  --resume              resume from the last completed stage

  --resume-from stage_name
resume from a custom stage

  --debug               enable debug output

  -v, --version         show program's version number and exit

1. Estimated genome size (required)

根据kmer值估计你的基因组大小，例如5M，2.3g

2. Minimum overlap length

这设置了一个最小的重叠长度，两个读操作被认为是重叠的。在最新的Flye版本中，这个参数是根据读取长度分布(读取N90)自动选择的，不需要手动设置

3. Metagenome mode

微生物用的

4. Reduced contig assembly coverage

对于一个普遍的装配，30X就足够了。

5. Number of polishing iterations

默认情况下，Flye运行一个迭代。额外的迭代可能会纠正少量的额外错误。可以设置为1-3即可

6. Re-starting from a particular assembly stage

使用——恢复，以恢复先前可能提前终止的汇编程序的运行( 使用相同的输出目录)，会继续从之前完成的哪一步开始，就是支持断点续传。

当然你也可以从值重新某一步： --resume-from stage_name, where stage_name is a choice of assembly, consensus, repeat, trestle, polishing

Flye output

改程序主要包括assembly, consensus, repeat, trestle, polishing五个阶段。所以最终会生成这几个文件夹。

生成文件夹

主要文件包括：

• assembly.fasta - Final assembly. Contains contigs and possibly scaffolds (see below).

• assembly_graph.{gfa|gv} - Final repeat graph. Note that the edge sequences might be different (shorter) than contig sequences, because contigs might include multiple graph edges (see below).

• assembly_info.txt - Extra information about contigs (such as length or coverage).