突變相關(guān)分析的時(shí)候我們經(jīng)常會(huì)選擇瀑布圖進(jìn)行展示，瀑布圖看起來(lái)十分的復(fù)雜高端奕剃，但是實(shí)際上只需要用一個(gè)GenVisR的R包即可解決。今天就讓我們來(lái)看一下如何做瀑布圖吧捐腿。另外纵朋，因?yàn)楸救瞬皇亲鲈摲较虻模绻忻枋霾粶?zhǔn)確的地方茄袖，請(qǐng)大家指正操软。

什么是瀑布圖 Waterfall Plot

Wiki上介紹的瀑布圖分為兩種，一類是2D形式宪祥，另一類是3D形式聂薪。我們簡(jiǎn)單介紹一下2D形式的瀑布圖。該類瀑布圖用于描述一系列中間正值或負(fù)值如何影響初始值品山。通常胆建，初始值和最終值（端點(diǎn)）由整列表示，而中間值則顯示為基于上一列的值開始的浮動(dòng)列肘交。這些列可以用不同的顏色標(biāo)記笆载，以區(qū)分正值和負(fù)值。

傳統(tǒng)瀑布圖的例子

可以看到該例子展示了獲利能力的分析涯呻。但是用于展示突變的瀑布圖和傳統(tǒng)的瀑布圖并不完全一樣凉驻，不過(guò)他們的展現(xiàn)形式很像。

SNP的瀑布圖

在SNP的瀑布圖中复罐，橫軸表示的是不同的樣本涝登，縱軸是基因，填充則代表該基因發(fā)生突變效诅，不同的顏色代表不同的突變情況胀滚。上面的柱狀圖是對(duì)于每個(gè)樣本突變情況的統(tǒng)計(jì)。

所以從該圖我們既能夠獲得每個(gè)樣本的具體信息乱投，也能夠從全局了解這一組樣本的整體情況咽笼，很好地展示了突變的情況。

怎么做瀑布圖

本次作圖我們使用一個(gè)叫做GenVisR的R包,該包除了提供瀑布圖還提供了多種其他形式較為復(fù)雜的戚炫、用于展現(xiàn)多個(gè)樣本突變情況的數(shù)據(jù)圖（見(jiàn)下圖）剑刑，具體的作圖方法大家可以參考GenVisR使用手冊(cè)。

GenVisR提供的可視化方式

今天我們要使用該包提供的一個(gè)叫做brcaMAF的數(shù)據(jù)表，通過(guò)名字也可以看出這是乳腺癌的數(shù)據(jù)施掏，該數(shù)據(jù)包含50個(gè)樣本钮惠，來(lái)源于TCGA，格式為MAF格式七芭。

MAF格式是Mutation Annotation Format的縮寫素挽，是一個(gè)以制表符分隔的文本文件，其中聚合了來(lái)自VCF文件的突變信息抖苦，該文件格式標(biāo)準(zhǔn)由TCGA制定毁菱，包含了一些關(guān)于突變的常見(jiàn)信息，進(jìn)一步的具體信息詳見(jiàn)MAF格式介紹

1）需要什么格式的數(shù)據(jù)
我們首先來(lái)看一下brcaMAF數(shù)據(jù)的情況锌历，可以看到該數(shù)據(jù)包括了55列信息税灌，如Hugo_Symbol柴钻、Chromosome等等,一共觀察到了2773個(gè)突變隐轩。

colnames(brcaMAF)
 [1] "Hugo_Symbol"                   "Entrez_Gene_Id"                "Center"                        "NCBI_Build"                    "Chromosome"                    "Start_Position"               
 [7] "End_Position"                  "Strand"                        "Variant_Classification"        "Variant_Type"                  "Reference_Allele"              "Tumor_Seq_Allele1"            
[13] "Tumor_Seq_Allele2"             "dbSNP_RS"                      "dbSNP_Val_Status"              "Tumor_Sample_Barcode"          "Matched_Norm_Sample_Barcode"   "Match_Norm_Seq_Allele1"       
[19] "Match_Norm_Seq_Allele2"        "Tumor_Validation_Allele1"      "Tumor_Validation_Allele2"      "Match_Norm_Validation_Allele1" "Match_Norm_Validation_Allele2" "Verification_Status"          
[25] "Validation_Status"             "Mutation_Status"               "Sequencing_Phase"              "Sequence_Source"               "Validation_Method"             "Score"                        
[31] "BAM_File"                      "Sequencer"                     "Tumor_Sample_UUID"             "Matched_Norm_Sample_UUID"      "chromosome_name_WU"            "start_WU"                     
[37] "stop_WU"                       "reference_WU"                  "variant_WU"                    "type_WU"                       "gene_name_WU"                  "transcript_name_WU"           
[43] "transcript_species_WU"         "transcript_source_WU"          "transcript_version_WU"         "strand_WU"                     "transcript_status_WU"          "trv_type_WU"                  
[49] "c_position_WU"                 "amino_acid_change_WU"          "ucsc_cons_WU"                  "domain_WU"                     "all_domains_WU"                "deletion_substructures_WU"    
[55] "transcript_error"      
head(brcaMAF)
  Hugo_Symbol Entrez_Gene_Id           Center NCBI_Build Chromosome Start_Position End_Position Strand Variant_Classification Variant_Type Reference_Allele Tumor_Seq_Allele1 Tumor_Seq_Allele2 dbSNP_RS dbSNP_Val_Status
1       A2ML1         144568 genome.wustl.edu         37         12        8994108      8994108      +      Missense_Mutation          SNP                G                 G                 C    novel                 
2       AADAC             13 genome.wustl.edu         37          3      151545656    151545656      +      Missense_Mutation          SNP                A                 A                 G    novel                 
3       AADAT          51166 genome.wustl.edu         37          4      170991750    170991750      +                 Silent          SNP                G                 G                 A    novel                 
4        AASS          10157 genome.wustl.edu         37          7      121756793    121756793      +      Missense_Mutation          SNP                G                 G                 A    novel                 
5        ABAT              0 genome.wustl.edu         37         16        8857982      8857982      +                 Silent          SNP                G                 G                 A    novel                 
6       ABCA3             21 genome.wustl.edu         37         16        2335631      2335631      +      Missense_Mutation          SNP                C                 T                 T    novel                 
          Tumor_Sample_Barcode  Matched_Norm_Sample_Barcode Match_Norm_Seq_Allele1 Match_Norm_Seq_Allele2 Tumor_Validation_Allele1 Tumor_Validation_Allele2 Match_Norm_Validation_Allele1 Match_Norm_Validation_Allele2
1 TCGA-A1-A0SO-01A-22D-A099-09 TCGA-A1-A0SO-10A-03D-A099-09                      G                      G                        G                        C                             G                             G
2 TCGA-A2-A0EU-01A-22W-A071-09 TCGA-A2-A0EU-10A-01W-A071-09                      A                      A                                                                                                              
3 TCGA-A2-A0ER-01A-21W-A050-09 TCGA-A2-A0ER-10A-01W-A055-09                      G                      G                                                                                                              
4 TCGA-A2-A0EN-01A-13D-A099-09 TCGA-A2-A0EN-10A-01D-A099-09                      G                      G                        G                        A                             G                             G
5 TCGA-A1-A0SI-01A-11D-A142-09 TCGA-A1-A0SI-10B-01D-A142-09                      G                      G                                                                                                              
6 TCGA-A2-A0D0-01A-11W-A019-09 TCGA-A2-A0D0-10A-01W-A021-09                      C                      C                                                                                                              
  Verification_Status Validation_Status Mutation_Status Sequencing_Phase Sequence_Source Validation_Method Score BAM_File      Sequencer                    Tumor_Sample_UUID             Matched_Norm_Sample_UUID
1             Unknown             Valid         Somatic         Phase_IV             WXS Illumina_WXS_gDNA     1    dbGAP Illumina GAIIx b3568259-c63c-4eb1-bbc7-af711ddd33db 17ba8cdb-e35b-4496-a787-d1a7ee7d4a1e
2             Unknown          Untested         Somatic         Phase_IV             WXS              none     1    dbGAP Illumina GAIIx de30da8f-903f-428e-a63d-59625fc858a9 1583a7c5-c835-44fa-918a-1448abf6533d
3             Unknown          Untested         Somatic         Phase_IV             WXS              none     1    dbGAP Illumina GAIIx 31ed187e-9bfe-4ca3-8cbb-10c1e0184331 2bc2fdaf-fb2f-4bfd-9e20-e20edff6633a
4             Unknown             Valid         Somatic         Phase_IV             WXS Illumina_WXS_gDNA     1    dbGAP Illumina GAIIx 12362ad7-6866-4e7a-9ec6-8a0a68df8896 ad478c68-a18b-4529-ad7a-86039e6da6b1
5             Unknown          Untested         Somatic         Phase_IV             WXS              none     1    dbGAP Illumina GAIIx e218c272-a7e1-4bc9-b8c5-d2d1c903550f fbcab9dc-4a6b-4928-9459-699c9932e3e1
6             Unknown          Untested         Somatic         Phase_IV             WXS              none     1    dbGAP Illumina GAIIx 3f20d0fe-aaa1-40f1-b2c1-7f070f93aef5 bbf1c43d-d7b3-4574-a074-d22ad537829c
  chromosome_name_WU  start_WU   stop_WU reference_WU variant_WU type_WU gene_name_WU transcript_name_WU transcript_species_WU transcript_source_WU transcript_version_WU strand_WU transcript_status_WU trv_type_WU
1                 12   8994108   8994108            G          C     SNP        A2ML1        NM_144670.3                 human              genbank                58_37c         1            validated    missense
2                  3 151545656 151545656            A          G     SNP        AADAC        NM_001086.2                 human              genbank                58_37c         1             reviewed    missense
3                  4 170991750 170991750            G          A     SNP        AADAT        NM_016228.3                 human              genbank                58_37c        -1             reviewed      silent
4                  7 121756793 121756793            G          A     SNP         AASS        NM_005763.3                 human              genbank                58_37c        -1             reviewed    missense
5                 16   8857982   8857982            G          A     SNP         ABAT        NM_000663.4                 human              genbank                58_37c         1             reviewed      silent
6                 16   2335631   2335631            C          T     SNP        ABCA3        NM_001089.2                 human              genbank                58_37c        -1             reviewed    missense
  c_position_WU amino_acid_change_WU ucsc_cons_WU                                                    domain_WU
1        c.1224              p.W408C        0.995                                                         NULL
2         c.896              p.N299S        0.000      HMMPfam_Abhydrolase_3,superfamily_alpha/beta-Hydrolases
3         c.708               p.L236        1.000 HMMPfam_Aminotran_1_2,superfamily_PLP-dependent transferases
4         c.788              p.T263M        1.000                                          HMMPfam_AlaDh_PNT_C
5         c.423               p.E141        0.987     HMMPfam_Aminotran_3,superfamily_PyrdxlP-dep_Trfase_major
6        c.3295             p.D1099N        0.980                                                         NULL
                                                                                                                                                                                                                                               all_domains_WU
1             HMMPfam_A2M,HMMPfam_A2M_N,superfamily_Terpenoid cyclases/Protein prenyltransferases,HMMPfam_A2M_recep,superfamily_Alpha-macroglobulin receptor domain,HMMPfam_A2M_N_2,HMMPfam_A2M_comp,HMMPfam_Thiol-ester_cl,PatternScan_ALPHA_2_MACROGLOBULIN
2                                                                                                                                                                         PatternScan_LIPASE_GDXG_SER,HMMPfam_Abhydrolase_3,superfamily_alpha/beta-Hydrolases
3                                                                                                                                                                                                HMMPfam_Aminotran_1_2,superfamily_PLP-dependent transferases
4 HMMPfam_Saccharop_dh,HMMPfam_AlaDh_PNT_C,HMMPfam_AlaDh_PNT_N,superfamily_NAD(P)-binding Rossmann-fold domains,superfamily_Formate/glycerate dehydrogenase catalytic domain-like,superfamily_Glyceraldehyde-3-phosphate dehydrogenase-like C-terminal domain
5                                                                                                                                                                    HMMPfam_Aminotran_3,PatternScan_AA_TRANSFER_CLASS_3,superfamily_PyrdxlP-dep_Trfase_major
6                                                                                                                            HMMPfam_ABC_tran,HMMSmart_SM00382,PatternScan_ABC_TRANSPORTER_1,superfamily_P-loop containing nucleoside triphosphate hydrolases
  deletion_substructures_WU transcript_error
1                         -        no_errors
2                         -        no_errors
3                         -        no_errors
4                         -        no_errors
5                         -        no_errors
6                         -        no_errors

那么我們的MAF文件也需要那么多信息嗎图云？并非如此，和很多其他作圖所需數(shù)據(jù)一樣卤材，其中有一些信息是必須提供的遮斥，另外一些是非必須的。
具體地分為三類情況扇丛，瀑布圖地功能提供了三種數(shù)據(jù)格式的選擇：
1.MAF
必須包括列名為"Tumor_Sample_Barcode","Hugo_Symbol","Variant_Classification"的信息
2.MGI
必須包括列名為"sample","gene_name","trv_type"的信息
3.Custom
必須包括列名為"sample", "gene", "variant_class"的信息

MGI也是一種以制表符分割的文本文件术吗，具體的可以見(jiàn)鏈接MGI格式介紹

2）如何作圖
waterfall函數(shù)有很多參數(shù)，可以根據(jù)需求展示突變信息帆精，那么下面就來(lái)一步作圖较屿，我們展示幾種常用的參數(shù)用途，其他更多具體參數(shù)的意義可以查看幫助?waterfall（本文的代碼來(lái)源GenVisR官方手冊(cè)）

# 最基本的作圖
waterfall(brcaMAF, fileType="MAF")
# 展示至少在6%的樣本中存在的突變
waterfall(brcaMAF, mainRecurCutoff = 0.06)
#特定基因的突變圖譜
waterfall(brcaMAF, plotGenes = c("PIK3CA", "TP53", "USH2A", "MLL3", "BRCA1"))

waterfall

我們常常需要結(jié)合樣本的臨床信息分析突變情況卓练，那么要怎樣同時(shí)展示樣本的臨床信息呢隘蝎？

#建立臨床信息
#分組情況為5組
subtype <- c("lumA", "lumB", "her2", "basal", "normal")
#隨機(jī)放回的分配組別
subtype <- sample(subtype, 50, replace = TRUE)
#年齡分為4組
age <- c("20-30", "31-50", "51-60", "61+")
#隨機(jī)分配年齡
age <- sample(age, 50, replace = TRUE)
#獲取樣本號(hào)
sample <- as.character(unique(brcaMAF$Tumor_Sample_Barcode))
#合并獲得臨床數(shù)據(jù)
clinical <- as.data.frame(cbind(sample, subtype, age))
head(clinical)
                        sample subtype   age
1 TCGA-A1-A0SO-01A-22D-A099-09   basal 51-60
2 TCGA-A2-A0EU-01A-22W-A071-09   basal 31-50
3 TCGA-A2-A0ER-01A-21W-A050-09  normal   61+
4 TCGA-A2-A0EN-01A-13D-A099-09  normal 31-50
5 TCGA-A1-A0SI-01A-11D-A142-09    lumA 31-50
6 TCGA-A2-A0D0-01A-11W-A019-09  normal   61+

# Melt the clinical data into 'long' format.
library(reshape2)
#整理數(shù)據(jù)表
clinical <- melt(clinical, id.vars = c("sample"))
head(clinical)
                        sample variable  value
1 TCGA-A1-A0SO-01A-22D-A099-09  subtype  basal
2 TCGA-A2-A0EU-01A-22W-A071-09  subtype  basal
3 TCGA-A2-A0ER-01A-21W-A050-09  subtype normal
4 TCGA-A2-A0EN-01A-13D-A099-09  subtype normal
5 TCGA-A1-A0SI-01A-11D-A142-09  subtype   lumA
6 TCGA-A2-A0D0-01A-11W-A019-09  subtype normal

# Run waterfall
waterfall(brcaMAF, clinDat = clinical, 
clinVarCol = c(lumA = "blue4", lumB = "deepskyblue", 
    her2 = "hotpink2", basal = "firebrick2", normal = "green4", `20-30` = "#ddd1e7", `31-50` = "#bba3d0", `51-60` = "#9975b9", `61+` = "#7647a2"), #設(shè)定顏色
plotGenes = c("PIK3CA", "TP53", "USH2A", "MLL3", "BRCA1"), clinLegCol = 2, 
clinVarOrder = c("lumA",  "lumB", "her2", "basal", "normal", "20-30", "31-50", "51-60", "61+"))#設(shè)定順序

添加臨床信息的waterfall

自己的數(shù)據(jù)可以整理成MAF格式，也可以選擇Custom格式襟企，要注意的是MAF格式和MGI格式對(duì)mutation type的類別名字有固定要求嘱么，如果你的mutation命名方式或者有不在下列類型中的突變類型，請(qǐng)選擇Custom類別顽悼，該作圖方式對(duì)mutation type的類別名字沒(méi)有限制曼振。

那么關(guān)于瀑布圖的分享就到這里啦～

往期R數(shù)據(jù)可視化分享
R數(shù)據(jù)可視化12: 曼哈頓圖
R數(shù)據(jù)可視化11: 相關(guān)性圖
R數(shù)據(jù)可視化10: 蜜蜂圖 Beeswarm
R數(shù)據(jù)可視化9: 棒棒糖圖 Lollipop Chart
R數(shù)據(jù)可視化8: 金字塔圖和偏差圖
R數(shù)據(jù)可視化7: 氣泡圖 Bubble Plot
R數(shù)據(jù)可視化6: 面積圖 Area Chart
R數(shù)據(jù)可視化5: 熱圖 Heatmap
R數(shù)據(jù)可視化4: PCA和PCoA圖
R數(shù)據(jù)可視化3: 直方/條形圖
R數(shù)據(jù)可視化2: 箱形圖 Boxplot
R數(shù)據(jù)可視化1: 火山圖