• 数据标准化的原因？
• 样本内标准化
  • RPM/CPM
  • RPKM/FPKM
  • TPM
• 样本间的标准化
  • 直接计算比例
  • Quantile
  • REL(Relative Log Expression)
  • TMM(Trimmed Mean of M-value)
• 基本统计学概念
  • 标准差（Standard Deviation）
  • MetaboAnalyst标准化的方法包括
• 中位数相同
• 参考

数据标准化的原因？

不同基因的长度不同
不同批次数据的测序量不同

ref:Computational methods for transcriptome annotation and quantification using RNA-seq

样本内标准化

RPM/CPM

• RPM: Reads per million mapped reads
• CPM: Count per million mapped reads
  
  

  图片alt

RPKM/FPKM

• RPKM: Reads per kilo base per million mapped reads
• FPKM: Fragment per kilo base million mapped reads
  
  

  图片alt

在双端测序中：FPKM=RPKM/2

TPM

• TPM: Transcript per million
  
  

  图片alt

TPM强行要求基因最终加和相同，高表达基因会将整体的表达量拉高

样本间的标准化

直接计算比例

• Cj 表示样本j的系数或矫正因子
• Dj表示样本j的总测序量

Quantile

• Cj 表示样本j的系数或矫正因子
• Dj表示样本j的总测序量
• Q(p)j表示样本j的p分位数

REL(Relative Log Expression)

使用中位数进行最终标准化

TMM(Trimmed Mean of M-value)

基本统计学概念

标准差（Standard Deviation）

• 总体标准差：
  

  图片alt

• 样本标准差：
  

  图片alt

• 标准误差：
  

  图片alt

MetaboAnalyst标准化的方法包括

• Sample Normalization
  • Sample-specific normalization
  • Normalization by sum
  • Normalization by median
  • Normalization by reference sample (PQN)
  • Normalization by a pooled sample from group
  • Normalization by reference feature
  • Quantile normalization
• Data transformation
  • Log transformation
  • Cube root transformation
• Data scaling
  • Mean centering (mean-centered only)
  • Auto scaling (mean-centered and divided by the standard deviation of each variable)
  • Pareto scaling (mean-centered and divided by the square root of the standard deviation of each variable)
  • Range scaling (mean-centered and divided by the range of each variable)

中位数相同

DEqMS蛋白数据差异分析
Use boxplot to check if the samples have medians centered. if not, do median centering.

# Here the data is already median centered, we skip the following step. 
# dat.log = equalMedianNormalization(dat.log)

参考

• R语言——标准化与归一化
• What the FPKM? A review of RNA-Seq expression units
• countToFPKM
• How to perform FPKM on raw reads in R