【视频-python一键函数】：热图标注特定基因及不同颜色设置展示marker基因

虽然scanpy有展示marker基因的热图函数（scanpy单细胞转录组python教程（五）：最详尽的基础可视化解析），但是有很明显的缺陷，就是不能label特定基因。为了解决这个问题，我们写了一个通用函数，input是dataframe，不仅可以展示单细胞marker基因平均表达量热图，其他的数据矩阵都可以使用。

视频解说演示链接：
https://www.bilibili.com/video/BV1k23BzREqK/?spm_id_from=333.1387.homepage.video_card.click&vd_source=05b5479545ba945a8f5d7b2e7160ea34

640.png

此外，对于整体cell的热图也可以标注基因，这里是参考了一篇NC文章：

640.png

首先看看函数参数：

640.png

获取marker基因（也可以手动整理）：

import scanpy as sc
import pandas as pd
import numpy as np

import matplotlib.pyplot as plt
import seaborn as sns
import matplotlib.colors as mcolors
#挑选需要plot的基因，读入数据
df_markers = pd.read_csv('./adata_markers.csv')
df_markers

这是scanpy自带函数的plot：

sc.pl.matrixplot(
    adata,
    df_markers['names'],
    "celltype",
    colorbar_title="mean z-score",#legend标题设置
    #layer="log1p_norm",#layer
    #categories_order=['Ciliated', 'UnCiliated','Endo','Stromal', 'Myeloid', 'T cell', 'NKT','B cell','Mast'],#自定义顺序
    vmin=-2,
    vmax=2,
    cmap="RdBu_r",
    swap_axes=False
)

640.png

获取average expressed matrix：

#get gene express data
#scaleed data
#you can set parameter layers
obs_tidy = sc.get.obs_df(adata,keys = df_markers['names'].tolist())
obs_tidy.head()
#keys添加obs，例如celltype或者leiden，就可以在表达矩阵第一列出现注释
genedf = sc.get.obs_df(adata, keys=["celltype", *df_markers['names'].tolist()])
genedf
#average expressed matrix
grouped = genedf.groupby("celltype", observed=True)
mean, var = grouped.mean(), grouped.var()
df = mean.T
df

plot（label的基因及其颜色）：

color_dict = {
  #B cell
  "IGHD" :"#fc8d59",
  "CD79A" : "#fc8d59",
  "CD19":"#fc8d59", 
  "IGHM":"#fc8d59",
  #Ciliated
  "CFAP52":"#9e9ac8", 
  "DNAH5":"#9e9ac8",
  "FOXJ1":"#9e9ac8", 
  #Endo
  "ACKR1":"#96daf7", 
  "HOXD1":"#96daf7",
  "VWF":"#96daf7",
  "EDN1":"#96daf7",
  "PECAM1":"#96daf7", 
  #Mast
  "TPSAB1":"#fed976",
  "TPSB2":"#fed976",
  "CPA3":"#fed976", 
  "MS4A2":"#fed976",
  "KIT":"#fed976",
  #Myeloid
  "FCGR1B":"#b0479a", 
  "TREM2":"#b0479a",
  "CD1E":"#b0479a", 
  #NKT
  "KRT81":"#35978f",
  "KLRC1":"#35978f", 
  "GNLY":"#35978f", 
  "ITGAD":"#35978f", 
  "NKG7":"#35978f", 
  #Stromal
  "COL1A1":"#e31a1c", 
  "COL3A1":"#e31a1c",
  "IGF1":"#e31a1c", 
  "TAS2R1":"#e31a1c",
  #T cell
  "CD3G":"#f6a2a7",
  "CD3D":"#f6a2a7", 
  "FOXP3":"#f6a2a7",
  "CCR4":"#f6a2a7",
  #UnCiliated
  "ASRGL1":"#f9d3d7",
  "KRT8":"#f9d3d7",
  "KRT18":"#f9d3d7", 
  "TMEM210":"#f9d3d7"
}
sc_markerAvgExp_heatmap(df,label=True,label_genes=color_dict,z_score=0)

640.png