Quantile normalization is frequently used in microarray data analysis. It was introduced as quantile standardization and then renamed as quantile normalization. Quantile, quartile, percentile ??? Quantiles are just the lines that divide data into equally sized groups. percentiles are just quantiles that divide the data into 100 equally sized groups

What’s Multilevel models, and how to deal with it What is multilevel model Multilevel model AKA: multilevel Models random-effects models hierarchical models variance-components models random-coefficient models mixed models Many kinds of data, including observational data collected in the human and biological sciences, have a hierarchical or clustered structure, or non-hierarchical

蒙特卡罗方法，又称统计模拟方法(statistical simulation method), 通过概率模型的随机抽样进行进行近似数值计算

奇异值分解(SVD)是一种矩阵因子分解方法，在线性代数中，被广泛应用。 奇异值分解也是一种矩阵近似的方

提升（Boosting）方法： 通过改变训练样本的权重（概率分布），学习n个分类器，并将这些分类器线性

隐马可夫模型（HMM）描述隐藏的马可夫链随机生成观测序列的过程，属于生成模型。 HMM在语音识别、自然

潜在语义分析（LSA）是一种非监督学习方法，用于文本话题分析。其特点是通过矩阵分解发现文本于单词之间

CRF条件随机场，可应用于标注问题 概率无向图模型Probabilistic undirected graphical model(Markov random field) 是一个可以由无向

Probability, P-value, Likelihood Probability: the level of possibility of something happening or being true. We determine the possibility of an event. We know the parameters associated with the event and assume them to be trustworthy. Likelihood: the chance that something will happen. We have some observations. We have an explanation

GATK is design for human genetics, but it also work well for inbred mice. However, one of my colleague who studies mouse genetics, said, I tried the haplotype caller from GATK. But it seems that the haplotype caller is designed for heterogeneous genome like human than for mice. Therefore, the result coming out of HC is worse than samtools, as I manually inspected a few regions that HC calls didn’t make sense.