转换OTU表和序列文件为PICRUST2需要的格式

Listenlii-生物信息知识分享

发布于 2022-07-30 14:02:47

9860

发布于 2022-07-30 14:02:47

文章被收录于专栏：Listenlii的生物信息笔记

作者：需要鼓励的小昱

审核：Listenlii

Load data

library(phyloseq)
packageVersion("phyloseq")
## [1] '1.32.0'
data("GlobalPatterns")# phyloseq 自带文件
GlobalPatterns

1.准备biom格式的OTU表

Export otu_table.txt

write.table(otu_table(GlobalPatterns), "GlobalPatterns_otu_table.txt", sep="\t", row.names=TRUE, col.names=NA, quote=FALSE)

Change .txt file to .biom file

在服务器运行这部分代码

# 把文件放到工作文件夹中
echo -n "#OTU Table" | cat - GlobalPatterns_otu_table.txt > GlobalPatterns_otu_biom_table.txt

biom convert -i GlobalPatterns_otu_biom_table.txt -o GlobalPatterns_otu.biom --table-type="OTU table" --to-hdf5

2. 准备.fna格式的代表性序列

Phyloseq 提供的数据提取不出来序列文件。只能用自己的了

library

library("phyloseq") 
library("dplyr") 
library("biomformat")
library("Biostrings")
load(file = "ps_rs_rsingle_CK_AT.rda")#载入一个phyloseq文件

whole rep seq

whole_seq_fasta <- readDNAStringSet("../raw_data/RDP-Bacteria-sequences.fasta")# 代表性序列文件
whole_sequences_fasta_ASV <- names(whole_seq_fasta)
whole_sequences_fasta_sequences <- paste(whole_seq_fasta)
whole_sequences_fasta_df <- data.frame(whole_sequences_fasta_ASV,
                                       whole_sequences_fasta_sequences)

Abundant taxa data

filter subset repseq：

从整个数据的fastq文件，提取出来我们想要的丰富种序列。

AT_tax_picrust <- taxa_names(ps_rs_CK_AT)
AT_tax_seq_fasta_df <- whole_sequences_fasta_df[with
                                                (whole_sequences_fasta_df,whole_sequences_fasta_ASV %in% AT_tax_picrust),]

rename the dataframe columns

AT_tax_seq_fasta_df_names <- rename(AT_tax_seq_fasta_df, 
                                   c( "whole_sequences_fasta_ASV" = "abundance",
                                      "whole_sequences_fasta_sequences" = "sequence")
                                   )

Export using dada2


uniquesToFasta(AT_tax_seq_fasta_df_names , 
                fout = "../rsingle_data/R_output/PICRUST2_data/AT_tax_rep_seqs.fna", #路径
                ids = AT_tax_seq_fasta_df_names$abundance)

最后就得到了PICRUST2需要的OTU表和序列文件了！

PS:

一个问题，如果想分别分析稀有种和优势种得功能，是整个跑Picrust之后再挑出不同类型的物种，还是先挑出不同类型物种再做分析？

github上有人回答建议整个跑出来之后再挑不同类型物种的结果查看。

It would be best to generate the stratified output (mentioned in the tutorial), which you can then analyze based on the different sets of ASVs you're interested in. You might also just want to look at the output of hsp.py specifically, as it sounds like you just want to compare the predicted genome annotations for rare and abundant ASVs. https://github.com/picrust/picrust2/issues/258

本文参与腾讯云自媒体同步曝光计划，分享自微信公众号。

原始发表：2022-06-24，如有侵权请联系 cloudcommunity@tencent.com 删除

github