Mutation Analysis (MutSig 2CV v3.1)
Sarcoma (Primary solid tumor)
28 January 2016  |  analyses__2016_01_28
Maintainer Information
Citation Information
Maintained by David Heiman (Broad Institute)
Cite as Broad Institute TCGA Genome Data Analysis Center (2016): Mutation Analysis (MutSig 2CV v3.1). Broad Institute of MIT and Harvard. doi:10.7908/C137785M
Overview
Introduction

This report serves to describe the mutational landscape and properties of a given individual set, as well as rank genes and genesets according to mutational significance. MutSig 2CV v3.1 was used to generate the results found in this report.

  • Working with individual set: SARC-TP

  • Number of patients in set: 247

Input

The input for this pipeline is a set of individuals with the following files associated for each:

  1. An annotated .maf file describing the mutations called for the respective individual, and their properties.

  2. A .wig file that contains information about the coverage of the sample.

Summary
  • MAF used for this analysis:SARC-TP.final_analysis_set.maf

  • Blacklist used for this analysis: pancan_mutation_blacklist.v14.hg19.txt

  • Significantly mutated genes (q ≤ 0.1): 13

Results
Lego Plots

The mutation spectrum is depicted in the lego plots below in which the 96 possible mutation types are subdivided into six large blocks, color-coded to reflect the base substitution type. Each large block is further subdivided into the 16 possible pairs of 5' and 3' neighbors, as listed in the 4x4 trinucleotide context legend. The height of each block corresponds to the mutation frequency for that kind of mutation (counts of mutations normalized by the base coverage in a given bin). The shape of the spectrum is a signature for dominant mutational mechanisms in different tumor types.

Figure 1.  Get High-res Image SNV Mutation rate lego plot for entire set. Each bin is normalized by base coverage for that bin. Colors represent the six SNV types on the upper right. The three-base context for each mutation is labeled in the 4x4 legend on the lower right. The fractional breakdown of SNV counts is shown in the pie chart on the upper left. If this figure is blank, not enough information was provided in the MAF to generate it.

Figure 2.  Get High-res Image SNV Mutation rate lego plots for 4 slices of mutation allele fraction (0<=AF<0.1, 0.1<=AF<0.25, 0.25<=AF<0.5, & 0.5<=AF) . The color code and three-base context legends are the same as the previous figure. If this figure is blank, not enough information was provided in the MAF to generate it.

CoMut Plot

Figure 3.  Get High-res Image The matrix in the center of the figure represents individual mutations in patient samples, color-coded by type of mutation, for the significantly mutated genes. The rate of synonymous and non-synonymous mutations is displayed at the top of the matrix. The barplot on the left of the matrix shows the number of mutations in each gene. The percentages represent the fraction of tumors with at least one mutation in the specified gene. The barplot to the right of the matrix displays the q-values for the most significantly mutated genes. The purple boxplots below the matrix (only displayed if required columns are present in the provided MAF) represent the distributions of allelic fractions observed in each sample. The plot at the bottom represents the base substitution distribution of individual samples, using the same categories that were used to calculate significance.

Significantly Mutated Genes

Column Descriptions:

  • nnon = number of (nonsilent) mutations in this gene across the individual set

  • npat = number of patients (individuals) with at least one nonsilent mutation

  • nsite = number of unique sites having a non-silent mutation

  • nsil = number of silent mutations in this gene across the individual set

  • p = p-value (overall)

  • q = q-value, False Discovery Rate (Benjamini-Hochberg procedure)

Table 1.  Get Full Table A Ranked List of Significantly Mutated Genes. Number of significant genes found: 13. Number of genes displayed: 35. Click on a gene name to display its stick figure depicting the distribution of mutations and mutation types across the chosen gene (this feature may not be available for all significant genes).

rank gene longname codelen nnei nncd nsil nmis nstp nspl nind nnon npat nsite pCV pCL pFN p q
1 TP53 tumor protein p53 1314 5 0 3 51 8 11 20 90 85 75 1e-16 0.00064 1e-05 1e-16 1.8e-12
2 RB1 retinoblastoma 1 (including osteosarcoma) 2891 12 0 0 1 8 6 11 26 24 26 2.6e-16 1 0.37 9.4e-15 5.8e-11
3 ATRX alpha thalassemia/mental retardation syndrome X-linked (RAD54 homolog, S. cerevisiae) 7615 2 0 0 7 14 3 14 38 36 38 6.3e-16 1 0.17 9.5e-15 5.8e-11
4 NUMBL numb homolog (Drosophila)-like 1866 425 0 0 0 0 0 8 8 8 1 4e-09 1e-05 0.97 1.3e-12 5.8e-09
5 MSH3 mutS homolog 3 (E. coli) 3508 8 0 2 0 0 0 9 9 7 5 3e-07 1e-05 1 8.1e-11 3e-07
6 LTBP3 latent transforming growth factor beta binding protein 3 4020 2 0 0 0 0 0 5 5 5 2 0.000016 1e-05 0.84 3.9e-09 0.000012
7 EOMES eomesodermin homolog (Xenopus laevis) 2081 117 0 0 0 0 0 5 5 5 1 0.00026 1e-05 1 5.4e-08 0.00014
8 PTEN phosphatase and tensin homolog (mutated in multiple advanced cancers 1) 1244 284 0 0 6 0 0 2 8 7 8 9e-08 1 0.04 2e-07 0.00046
9 KRTAP5-5 keratin associated protein 5-5 716 24 0 1 3 0 0 6 9 7 8 3.5e-06 0.14 0.55 8.8e-06 0.018
10 CABLES1 Cdk5 and Abl enzyme substrate 1 1992 206 0 0 0 0 0 3 3 3 1 0.0011 0.001 0.84 0.000016 0.029
11 TRAF7 TNF receptor-associated factor 7 2093 166 0 0 2 0 0 2 4 4 3 0.00086 0.019 0.039 2e-05 0.033
12 LOR loricrin 943 0 0 2 0 0 0 6 6 6 2 0.0084 0.00032 1 0.000044 0.067
13 R3HDM1 R3H domain containing 1 3396 13 0 0 3 1 0 1 5 4 5 0.00061 0.0059 0.89 6e-05 0.085
14 NF1 neurofibromin 1 (neurofibromatosis, von Recklinghausen disease, Watson disease) 8807 1 0 1 3 1 1 5 10 9 10 0.00037 0.017 0.93 9e-05 0.11
15 COL4A3 collagen, type IV, alpha 3 (Goodpasture antigen) 5291 12 0 0 5 0 1 0 6 4 5 0.13 6e-05 0.86 0.000098 0.11
16 MEGF9 multiple EGF-like-domains 9 1948 278 0 0 0 0 0 3 3 3 1 0.0077 0.001 0.98 0.000099 0.11
17 SCAP SREBF chaperone 3928 15 0 0 3 0 1 0 4 3 4 0.0027 0.0076 0.16 0.00025 0.27
18 PTPRU protein tyrosine phosphatase, receptor type, U 4483 137 0 2 7 0 0 0 7 7 7 0.00045 1 0.037 0.0003 0.3
19 CPNE4 copine IV 1733 21 0 0 4 0 0 0 4 2 4 0.28 0.0001 0.71 0.00032 0.31
20 CHD5 chromodomain helicase DNA binding protein 5 6031 9 0 0 4 0 2 0 6 5 6 0.014 0.03 0.051 0.00035 0.31
21 TMPRSS5 transmembrane protease, serine 5 (spinesin) 1422 93 0 0 2 0 0 0 2 2 2 0.0033 1 0.01 0.00037 0.31
22 KRT4 keratin 4 1817 29 0 0 3 0 0 0 3 2 3 0.033 0.02 0.013 0.00038 0.31
23 PKD1 polycystic kidney disease 1 (autosomal dominant) 13092 1 0 0 0 0 2 1 3 3 3 0.00036 1 0.17 0.00065 0.52
24 APOB apolipoprotein B (including Ag(x) antigen) 13804 53 0 4 8 1 0 1 10 9 10 0.014 0.015 0.29 0.00096 0.73
25 PKD2 polycystic kidney disease 2 (autosomal dominant) 2965 214 0 0 1 0 0 2 3 3 2 0.025 0.0043 0.69 0.0012 0.85
26 MUC5B mucin 5B, oligomeric mucus/gel-forming 17492 26 0 4 12 0 0 0 12 8 12 0.23 0.043 0.0031 0.0012 0.85
27 DOCK3 dedicator of cytokinesis 3 6303 0 0 0 4 1 1 0 6 6 6 0.041 0.013 0.17 0.0013 0.86
28 FRAS1 Fraser syndrome 1 12406 2 0 1 4 0 1 0 5 4 5 0.2 0.0036 0.024 0.0014 0.89
29 KDM3A lysine (K)-specific demethylase 3A 4066 18 0 0 3 0 1 0 4 3 4 0.0073 0.02 0.89 0.0014 0.9
30 RABEP1 rabaptin, RAB GTPase binding effector protein 1 2657 7 0 0 1 1 0 1 3 3 3 0.0018 1 0.083 0.0015 0.91
31 RSAD1 radical S-adenosyl methionine domain containing 1 1363 10 0 2 2 1 0 0 3 3 2 0.0043 0.026 0.46 0.0015 0.91
32 IGF2BP1 insulin-like growth factor 2 mRNA binding protein 1 1790 173 0 0 3 0 1 0 4 4 4 0.0028 1 0.037 0.0018 0.97
33 NUP210L nucleoporin 210kDa-like 5825 4 0 0 2 1 0 0 3 1 3 0.17 0.001 0.77 0.0019 0.97
34 DAPK1 death-associated protein kinase 1 4389 35 0 0 2 1 1 0 4 4 4 0.00021 1 0.52 0.002 0.97
35 PLEKHG4B pleckstrin homology domain containing, family G (with RhoGef domain) member 4B 3886 41 0 0 4 0 0 0 4 3 4 0.11 0.0078 0.25 0.002 0.97
TP53

Figure S1.  This figure depicts the distribution of mutations and mutation types across the TP53 significant gene.

RB1

Figure S2.  This figure depicts the distribution of mutations and mutation types across the RB1 significant gene.

ATRX

Figure S3.  This figure depicts the distribution of mutations and mutation types across the ATRX significant gene.

NUMBL

Figure S4.  This figure depicts the distribution of mutations and mutation types across the NUMBL significant gene.

MSH3

Figure S5.  This figure depicts the distribution of mutations and mutation types across the MSH3 significant gene.

LTBP3

Figure S6.  This figure depicts the distribution of mutations and mutation types across the LTBP3 significant gene.

EOMES

Figure S7.  This figure depicts the distribution of mutations and mutation types across the EOMES significant gene.

PTEN

Figure S8.  This figure depicts the distribution of mutations and mutation types across the PTEN significant gene.

KRTAP5-5

Figure S9.  This figure depicts the distribution of mutations and mutation types across the KRTAP5-5 significant gene.

CABLES1

Figure S10.  This figure depicts the distribution of mutations and mutation types across the CABLES1 significant gene.

TRAF7

Figure S11.  This figure depicts the distribution of mutations and mutation types across the TRAF7 significant gene.

LOR

Figure S12.  This figure depicts the distribution of mutations and mutation types across the LOR significant gene.

R3HDM1

Figure S13.  This figure depicts the distribution of mutations and mutation types across the R3HDM1 significant gene.

Methods & Data
Methods

MutSig and its evolving algorithms have existed since the youth of clinical sequencing, with early versions used in multiple publications. [1][2][3]

"Three significance metrics [are] calculated for each gene, using the […] methods MutSigCV [4], MutSigCL, and MutSigFN [5]. These measure the significance of mutation burden, clustering, and functional impact, respectively […]. MutSigCV determines the P value for observing the given quantity of non-silent mutations in the gene, given the background model determined by silent (and noncoding) mutations in the same gene and the neighbouring genes of covariate space that form its 'bagel'. […] MutSigCL and MutSigFN measure the significance of the positional clustering of the mutations observed, as well as the significance of the tendency for mutations to occur at positions that are highly evolutionarily conserved (using conservation as a proxy for probably functional impact). MutSigCL and MutSigFN are permutation-based methods and their P values are calculated as follows: The observed nonsilent coding mutations in the gene are permuted T times (to simulate the null hypothesis, T = 108 for the most significant genes), randomly reassigning their positions, but preserving their mutational 'category', as determined by local sequence context. We [use] the following context categories: transitions at CpG dinucleotides, transitions at other C-G base pairs, transversions at C-G base pairs, mutations at A-T base pairs, and indels. Indels are unconstrained in terms of where they can move to in the permutations. For each of the random permutations, two scores are calculated: SCL and SFN, measuring the amount of clustering and function impact (measured by conservation) respectively. SCL is defined to be the fraction of mutations occurring in hotspots. A hotspot is defined as a 3-base-pair region of the gene containing many mutations: at least 2, and at least 2% of the total mutations. SFN is defined to be the mean of the base-pair-level conservation values for the position of each non-silent mutation […]. To determine a PCL, the P value for the observed degree of positional clustering, the observed value of SCL (computed for the mutations actually observed), [is] compared to the distribution of SCL obtained from the random permutations, and the P value [is] defined to be the fraction of random permutations in which SCL [is] at least as large as the observed SCL. The P value for the conservation of the mutated positions, PFN, [is] computed analogously." [6]

Download Results

In addition to the links below, the full results of the analysis summarized in this report can also be downloaded programmatically using firehose_get, or interactively from either the Broad GDAC website or TCGA Data Coordination Center Portal.

References
[1] Getz G, Höfling H, Mesirov JP, Golub TR, Meyerson M, Tibshirani R, Lander ES, Comment on "The Consensus Coding Sequences of Human Breast and Colorectal Cancers", Science 317(5844):1500b (2007)
[3] TCGA, Integrated genomic analyses of ovarian carcinoma, Nature 474(7353):609-615 (2011)
[4] Lawrence MS, et al., Mutational heterogeneity in cancer and the search for new cancer-associated genes, Nature 499(7457):214-218 (2013)
[6] Lawrence MS, et al., Discovery and saturation analysis of cancer genes across 21 tumour types, Nature 505(7484):495-501 (2014)