Clustering of miRseq precursor expression: consensus NMF
Lung Squamous Cell Carcinoma (Primary solid tumor)
28 January 2016  |  analyses__2016_01_28
Maintainer Information
Citation Information
Maintained by Hailei Zhang (Broad Institute)
Cite as Broad Institute TCGA Genome Data Analysis Center (2016): Clustering of miRseq precursor expression: consensus NMF. Broad Institute of MIT and Harvard. doi:10.7908/C1KP81MR
Overview
Introduction

This pipeline calculates clusters based on a consensus non-negative matrix factorization (NMF) clustering method , . This pipeline has the following features:

1. Convert input data set to a non-negitive matrix by column rank normalization.

2. Classify samples into consensus clusters.

3. Determine differentially expressed marker miRs for each subtype.

Summary

The most robust consensus NMF clustering of 478 samples using the 150 most variable miRs was identified for k = 4 clusters. We computed the clustering for k = 2 to k = 10 and uused the cophenetic correlation coefficient and the average silhouette width calculation to determine the robust clusters.

Results
miR expression patterns of molecular subtypes
Silhouette widths, Cophenetic Correlation Coefficients and Consensus matrix
Samples assignment with silhouette width

Table 1.  Get Full Table List of samples with 4 subtypes and silhouette width.

SampleName cluster silhouetteValue
TCGA-18-5592-01 1 0.079
TCGA-18-5595-01 1 0.11
TCGA-21-5782-01 1 0.033
TCGA-21-5786-01 1 0.11
TCGA-21-5787-01 1 0.2
TCGA-22-5471-01 1 0.026
TCGA-22-5474-01 1 0.05
TCGA-22-5480-01 1 0.12
TCGA-22-5481-01 1 -0.15
TCGA-22-5489-01 1 0.097

Table 2.  Get Full Table List of samples belonging to each cluster in different k clusters.

SampleName K=2 K=3 K=4 K=5 K=6 K=7 K=8
TCGA-18-5592-01 1 1 1 1 1 1 1
TCGA-18-5595-01 1 1 1 2 2 1 1
TCGA-21-5782-01 1 1 1 1 1 1 2
TCGA-21-5783-01 1 1 2 2 3 2 3
TCGA-21-5786-01 1 1 1 1 1 1 5
TCGA-21-5787-01 1 1 1 1 1 1 1
TCGA-21-A5DI-01 1 2 3 3 4 3 4
TCGA-22-4593-01 1 2 3 3 4 3 4
TCGA-22-5474-01 1 1 1 1 1 1 1
TCGA-22-5478-01 1 1 2 5 3 2 3
Marker miRs of each subtype

Samples most representative of the clusters, hereby called core samples were identified based on positive silhouette width, indicating higher similarity to their own class than to any other class member. Core samples were used to select differentially expressed marker miRs for each subtype by comparing the subclass versus the other subclasses, using Student's t-test.

Table 3.  Get Full Table List of marker miRs with p <= 0.05 (The positive value of column difference means miR is upregulated in this subtype and vice versa).

Composite.Element.REF p difference q subclass
HSA-LET-7A-1 2.3e-10 -0.38 9.1e-09 1
HSA-LET-7A-2 1.5e-10 -0.38 7.3e-09 1
HSA-LET-7A-3 2.1e-10 -0.38 9e-09 1
HSA-LET-7C 0.015 -0.31 0.033 1
HSA-LET-7D 0.0017 0.21 0.005 1
HSA-LET-7E 1.4e-09 -0.52 3.2e-08 1
HSA-LET-7F-2 4.2e-08 -0.51 6.4e-07 1
HSA-MIR-1-2 0.00046 -0.57 0.0017 1
HSA-MIR-100 7.7e-07 -0.56 8.1e-06 1
HSA-MIR-101-1 9.2e-07 -0.38 9.5e-06 1
Methods & Data
Input

miRseq (at precursor level) of RPM value (reads per million reads aligned to miRBase precursor) with log2 transformed was as the input data for the clustering.

• Input file for selecting top 150 genes = *.miRseq_RPKM_log2.txt from miRseq_Preprocess

• Input file for the clustering module = /xchip/cga/gdac-prod/tcga-gdac/jobResults/GDAC_TopgenesforCluster/LUSC-TP/22507489/LUSC-TP.expclu.gct

CNMF Method

Non-negative matrix factorization (NMF) is an unsupervised learning algorithm that has been shown to identify molecular patterns when applied to miR expression data , . Rather than separating miR clusters based on distance computation, NMF detects contextdependent patterns of miR expression in complex biological systems.

Cophenetic Correlation Coefficient and How to select the best cluster

We use the cophenetic correlation coefficients to determine the cluster that yields the most robust clustering. The cophenetic correlation coefficient is computed based on the consensus matrix of the CNMF clustering, and measures how reliably the same samples are assigned to the same cluster across many iterations of the clustering lgorithm with random initializations. The cophenetic correlation coefficients and average silhouette values are used to determine the k with the most robust clusterings. From the plot of cophenetic correlation versus k, we select modes and the the point preceding the greatest decrease in cophenetic correlation coefficient, and from these choose the k with the highest average silhouette value.

Silhouette Width

Silhouette width is defined as the ratio of average distance of each sample to samples in the same cluster to the smallest distance to samples not in the same cluster. If silhouette width is close to 1, it means that sample is well clustered. If silhouette width is close to -1, it means that sample is misclassified .