2019-07-29 【Pacbio三代转录组测序】拆分read

CCS: Generate Highly Accurate Single-Molecule Consensus Reads

Latest ccs can be installed via bioconda package pbccs.
官网 https://github.com/PacificBiosciences/ccs

#安装
conda install -c bioconda pbccs

image.png

Input: Subreads from a single movie in PacBio BAM format (.subreads.bam).

Output: Consensus reads in a format inferred from the file extension: unaligned BAM (.bam); FASTQ (.fastq); or SMRT Link XML (.consensusreadset.xml) which also generates a corresponding BAM file.

用以下代码跑
Run on a full movie:

ccs movie.subreads.bam movie.ccs.bam --noPolish --minPasses 1

Usage: ccs [options] INPUT OUTPUT
Generate circular consensus sequences (ccs) from subreads.

Basic Options:
  -h,--help                 Output this help.
  --version                 Output version information.
  --logFile                 Log to a file, instead of stderr.
  --log-level,--logLevel    Set log level: "TRACE", "DEBUG", "INFO", "WARN",
                            "FATAL". ["WARN"]
  -j,--numThreads           Number of threads to use, 0 means autodetection.
                            [0]

Input Filter Options:
  --minLength               Minimum length of subreads to use for generating
                            CCS. [10]  默认
  --maxLength               Maximum length of subreads to use for generating
                            CCS. [21000]  也许可以设置长一些，得到一些意想不到的转录本
  --minPasses               Minimum number of subreads required to generate
                            CCS. [3]   最低至少有3个subreads，就是转3圈之后，认为可以生成可信的ccs
  --minIdentity             Minimum identity of a subread aligned to the draft
                            consensus to use it for polishing. 0 disables this
                            filter. [0.82]    默认了，不知道做什么
  --minSnr                  Minimum SNR of input subreads. [2.5]   
  --zmws                    Generate CCS for the provided comma-separated
                            holenumber ranges only. Default = all

Model Override Options:  模型？不知道什么意思
  --modelPath               Path to a model file or directory containing model files.
  --modelSpec               Name of chemistry or model to use, overriding default selection.

Processing Options:  
  --byStrand                Generate a consensus for each strand.
  --noPolish                Only output the initial template derived from the
                            POA (**faster, less accurate**).  不修正序列？
  --richQVs                 Emit dq, iq, and sq "rich" quality tracks.

Output Filter Options:
  --minPredictedAccuracy    Minimum predicted accuracy in [0, 1]. [0.9]
  --minReadScore            Minimum read score of input subreads. [0.75]
  --maxDropFraction         Maximum fraction of subreads dropped by polishing
                            (not input filters) before skipping ZMW. [0.34]

Output Files Options:
  --force                   Overwrite OUTPUT file if present.   强制覆盖并重写文件
  --reportFile              Where to write the results report.   report文件路径
                            ["ccs_report.txt"]

Options:
  --emit-tool-contract      Emit tool contract.
  --resolved-tool-contract  Use args from resolved tool contract.

Arguments:
  input                     Input file.
  output                    Output file.

影响CCS reads数目和质量的因素：
The longer the polymerase read gets, more readouts (passes) of the SMRTbell are produced and consequently more evidence is accumulated per molecule. This increase in evidence translates into higher consensus accuracy, as depicted in the following sketch:

image.png

运行时间：
20190728: 0点开始 - 结束 100M转录组数据