【绘图进阶】之绘制PCA biplot图(二)

    常规的biplot绘图，虽然可以把样品和物种标在PCA图上，但是效果非常不理想。ggbiplot是一款PCA分析结果可视化的R包工具，可以直接采用ggplot2来可视化R中基础函数prcomp()结果，按分组绘图，添加椭圆、箭头和物种名称。

2.PCA分析

In [2]:

data_pca = prcomp(df,scale=TRUE)

In [3]:

data_pca$rotation

Out[3]:
PC1   PC2   PC3   PC4   PC5   PC6   PC7   PC8   PC9   PC10   PC11   PC12   PC13   PC14   PC15   PC16
X68 -0.2430500 0.38916585 0.023282518 0.05234324 0.36699759 0.05506225 -0.050560004 -0.10593654 -0.11444462 0.36953735 -0.30537404 0.338903555 0.41402670 -0.284624230 -0.090105023 -0.12800184
X69 -0.2402164 0.36218483 -0.013061765 0.51194696 -0.07174752 0.45384168 0.176893288 -0.18258457 0.11230165 0.10908458 0.25096766 -0.373622075 -0.11261695 0.091127424 0.058957993 -0.14682460
X106 -0.2557063 -0.12447526 0.173438926 -0.13426542 -0.44009941 -0.03620825 -0.171205288 -0.27565035 -0.07107470 0.06744872 0.24642356 -0.132560788 0.22757381 -0.559334120 0.272614351 0.20942325
X6 -0.2536670 0.05623670 -0.468005262 0.15240255 -0.21849821 0.21291012 -0.042266340 0.08882465 -0.05819861 -0.63923630 -0.04586830 0.368626963 0.18738234 -0.041111689 0.002259739 0.02606331
X74 -0.2566932 0.04852627 0.119204747 0.06838263 0.54204129 -0.26228748 -0.168889220 -0.15750400 0.37401913 -0.24653397 0.43918652 0.132795250 -0.06540488 0.027949070 0.035667605 0.28180399
X112 -0.2484910 0.32559769 0.059756361 -0.02354247 -0.01974062 -0.15863691 -0.236439490 0.13963796 -0.52135144 -0.02553914 -0.12506557 -0.257332430 0.01738885 0.385639179 0.082978501 0.46575660
X75 -0.2450634 -0.34701579 -0.056762940 -0.29495653 0.02895928 0.34410897 0.015339684 0.05938054 0.25976962 0.18702376 0.06575778 -0.115475427 0.53309938 0.434943704 -0.057135102 0.09767001
X77 -0.2513740 -0.22241477 -0.275059754 -0.16383243 0.24778626 0.37816523 -0.304629744 0.31352885 -0.15620294 0.21428547 0.08568085 -0.041715898 -0.46765947 -0.285282621 -0.088286154 0.04530152
X76 -0.2184556 -0.53765837 0.359586523 0.60243062 0.04023701 -0.03870977 -0.169555479 0.06393069 -0.03699587 -0.05759508 -0.35819968 -0.007589647 0.01859631 0.003219111 -0.037389048 -0.01473870
X7 -0.2449620 -0.21491066 -0.616404879 0.17990580 0.01817630 -0.53453347 0.281215337 -0.09008429 -0.05890281 0.27059539 0.03529958 -0.167312985 0.02918828 -0.001724188 0.002416606 -0.03795324
X104 -0.2587459 -0.03355179 0.008686834 -0.16896009 -0.08510343 0.03558645 -0.005192017 -0.29396775 0.09086539 0.13010662 -0.24205025 0.337891028 -0.35686557 0.297527487 0.611915023 -0.14174029
X105 -0.2577352 -0.06682350 0.184230342 -0.10012624 -0.21442142 0.04157705 0.311069102 -0.32783475 -0.16500884 0.07210566 0.08267707 0.313885377 -0.25793701 0.103284263 -0.617888919 0.19085013
X107 -0.2568870 0.06880793 0.019220849 -0.32495584 0.14486672 -0.01338294 0.153501873 -0.19078478 0.22179854 -0.35954343 -0.50832904 -0.485621957 -0.10601452 -0.191133770 -0.135823579 -0.05717753
X80 -0.2564216 -0.09961860 0.240161326 -0.18274753 0.16787236 -0.09335768 0.143164978 0.09703855 -0.45169269 -0.22824042 0.32685081 -0.057313696 0.09276323 0.064916502 0.064533540 -0.61904805
X110 -0.2565415 0.11833528 0.226514035 -0.01270806 -0.10483179 -0.07664522 0.524717549 0.65239299 0.21146831 0.06159903 -0.01427284 0.115320466 -0.04797013 -0.149323315 0.172812582 0.18554028
X111  -0.2528704  0.22570474  0.029792623  -0.04277772  -0.38337351  -0.28980173  -0.483304622  0.21887859  0.35215399  0.11053195  0.01820724  
0.002325398  -0.06549265  0.117608596  -0.291131534  -0.36124728

3.biplot图绘制

In [4]:

biplot(data_pca)

Out[4]:

    install_github 需要安装软件rtools（不是R包），最新的是rtools40，针对的是R 4.0 版本，一般安装的R version 3.6.*，所以rtools要选择版本下载，默认安装到C盘，避免麻烦。

1.加载R包

In[5]:
library(devtools)
library(ggbiplot)

2.读取数据

In[6]:

df = read.table('ggbiplot_data.txt',header = T,sep = ' ',row.names = 1)
head(df)

Out[6]:
Veillonella Neisseria Prevotella Porphyromonas Streptococcus Lachnoanaerobaculum Treponema Leptotrichia
UC01 0.01339965 0.16395553 0.08344061 0.002750146 0.1247513 0.000000000 0.053481568 0.04324166
UC02 0.05143359 0.13288473 0.07963721 0.065652428 0.1256290 0.006260971 0.003861908 0.03282621
UC03 0.05418373 0.04967817 0.14101814 0.017846694 0.1716208 0.006085430 0.015740199 0.02387361
UC04 0.05149210 0.01141018 0.10854301 0.004505559 0.2196021 0.007021650 0.032358104 0.02527794
UC05 0.04031597 0.09520187 0.08759509 0.075775307 0.2674664 0.001696899 0.001930954 0.02375658
UC06  0.05839672  0.02293739  0.11854886  0.003393798  0.1412522  0.005032183  0.018607373  0.03089526

3.指定绘图中的标签排列

In [7]:
Group = read.table('ggboxplot_group.txt',header = T,sep = ' ')
group = factor(Group$group,levels = c("CD","UC","HC"))
group

Out[7]:

UC UC UC UC UC UC UC UC UC UC CD CD CD CD CD CD CD CD CD CD CD CD HC HC HC HC HC HC HC HC

4. PCA分析
In [8]:

df_pca = prcomp(df,scale. = T)
df_pca

Out[8]:

Standard deviations (1, .., p=8):
[1] 1.7821933 1.2878327 1.0918639 0.7871798 0.7236448 0.6625366 0.4725125
[8] 0.4093539

Rotation (n x k) = (8 x 8):

PC1 PC2 PC3 PC4 PC5
Veillonella 0.4573712 -0.24397676 0.18983692 -0.3538455 0.23524282
Neisseria -0.2236327 -0.39722188 -0.52758737 0.1988236 0.61197685
Prevotella 0.3992054 0.34811232 0.03861493 -0.2718197 0.61713984
Porphyromonas -0.3377305 -0.42471400 0.21734617 -0.1666548 0.12510516
Streptococcus -0.1964093 0.10167920 0.71839929 0.5098384 0.36876531
Lachnoanaerobaculum 0.4728501 -0.21213688 0.11285891 0.2122485 -0.16521656
Treponema -0.1520557 0.64532333 -0.24043181 0.1116204 0.09942415
Leptotrichia 0.4267181 -0.09665939 -0.22414483 0.6451039 -0.02151154

PC6 PC7 PC8
Veillonella -0.03135291 0.12042148 0.70655408
Neisseria 0.15586753 -0.25749323 0.09597301
Prevotella -0.07492942 0.06084414 -0.50388271
Porphyromonas -0.74952775 0.15434102 -0.17110997
Streptococcus 0.14946847 -0.05381796 0.11758812
Lachnoanaerobaculum -0.25907424 -0.74950230 -0.13211541
Treponema -0.51264597 -0.20198211 0.42033615
Leptotrichia -0.23462707 0.53500084 -0.02074083

5.绘图

In[9]:

ggbiplot(df_pca, obs.scale = 1, var.scale = 1, groups = group, ellipse = TRUE, circle = TRUE) + scale_color_discrete(name = '') + theme(legend.direction = 'horizontal', legend.position = 'top')

Out[9]:

var.axes= T/F 是否添加物种及箭头

In[10]:

ggbiplot(df_pca, obs.scale = 1, var.scale = 1, groups = group, ellipse = TRUE,var.axes = F)

Out[10]:

In [11]:

ggbiplot(df_pca, obs.scale = 1, var.scale = 1, groups = group, ellipse = TRUE,var.axes = T)

Out[11]: