Last Updated: 07/26/2019
Cupcake是一個三代測序后續(xù)分析軟件的集合，可以
cDNA_Cupcake is a miscellaneous collection of Python and R scripts used for analyzing sequencing data. Most of the scripts only require Biopython. For scripts that require additional libraries, it will be specified in documentation.
https://github.com/Magdoll/cDNA_Cupcake
Current version: 8.2

發(fā)現(xiàn)一個不錯的介紹

https://github.com/Magdoll/cDNA_Cupcake/wiki#refgmap

1. 首先通過git拉包

git clone https://github.com/Magdoll/cDNA_Cupcake.git

出現(xiàn)錯誤

(base) [jing@localhost ~]$ git clone https://github.com/Magdoll/cDNA_Cupcake.git
Cloning into 'cDNA_Cupcake'...
error: RPC failed; curl 56 OpenSSL SSL_read: SSL_ERROR_SYSCALL, errno 104
fatal: the remote end hung up unexpectedly

查詢得解答：
使用git clone error: RPC failed

#Solution:
#修改Git的傳輸字節(jié)限制即可。
 git config --global http.postBuffer  524288000

運行以上代碼后，正常下載了
這步驟比較慢，，14:20-15:05，斷線，重新上， 大約60分鐘

(base) [jing@localhost ~]$ git clone https://github.com/Magdoll/cDNA_Cupcake.git
Cloning into 'cDNA_Cupcake'...
remote: Enumerating objects: 164, done.
remote: Counting objects: 100% (164/164), done.
remote: Compressing objects: 100% (115/115), done.
Receiving objects:  18% (301/1615), 9.45 MiB | 49.00 KiB/s

運行以下：

export PATH=$PATH:/home/jing/cDNA_Cupcake/sequence/
export PATH=$PATH:/home/jing/cDNA_Cupcake/rarefaction/
改為自己的路徑

2. 裝Cupcake ToFU
   因為： The only exception is Cupcake ToFU, which does require compiling and installation.
   https://github.com/Magdoll/cDNA_Cupcake/wiki/Cupcake-ToFU%3A-supporting-scripts-for-Iso-Seq-after-clustering-step

下載下來之后，
cd cDNA_Cupcake
python setup.py build
python setup.py install

報錯

image.png

缺啥安啥
conda install numpy
yum search zlib
install之后重新運行安裝

image.png

繼續(xù)yum search gcc
install
還是不行

image.png

<meta charset="utf-8">

What to do after Iso Seq Cluster?https://github.com/PacificBiosciences/IsoSeq_SA3nUP/wiki/What-to-do-after-Iso-Seq-Cluster%3F

Cupcake ToFU 能做什么？

在經(jīng)過cluster步驟之后，我們應(yīng)該已經(jīng)獲得了高質(zhì)量isoforms（HQ isoform sequences.），滿足以下條件：

1. 所得序列為全長（包含5‘UTR，序列中包含polyA）
2. 高質(zhì)量(predicted accuracy by default is >= 99%)
3. 有至少2個全長序列支持（subreads？）

獨白：可能用不著那么高質(zhì)量的reads，也可以挖掘很多有用的信息

這寫高質(zhì)量isoforms中，依舊存在冗余序列（isoforms），因此前步驟產(chǎn)出的序列，并不能真正代表樣品中的所有unique isoforms。有兩個原因：

1. Clustering algorithm tradeoff between sensitivity and specificity.

2. Natural 5' degradation in RNA.
   所以，下面需要做的步驟有

3. Best practice for aligning Iso Seq to reference genome: minimap2, GMAP, STAR, BLAT

4. Collapse identical isoforms to obtain final set of unique, full-length, high-quality isoforms

5. Obtain associated count information for each unique isoform

6. Robust ORF prediction using ANGEL

7. Fusion finding -- tutorial to come soon

Cupcake TOFU 可以做第 (2), (3), and (5)步