To determine the genetic relationship between Hainan aborigines and other populations, we performed two methods of clustering analyses (Fig. 2) using the Y-SNP haplogroup frequencies of Hainan aborigines and other East Asian populations [2], [4], [8] [10], [12], [13], [28].
Furthermore, the set of genes used for clustering analyses are in part different among validation datasets, depending on the availability of gene-expression data in a specific study.
Here we map results of cluster analyses using both Sorenson's (similar to Jaccard's) and Simpson's coefficients with Group Average clustering, and Infomaps on both 2009 and a 2015 OBIS data set (used in10 (Fig. 2).
We performed three types of cluster analyses using sequence similarities derived from: I) The 60-amino acid homeodomain identified by McGinnis et al. [7], II) Full-length Hox-protein sequences and III) Extended homeodomain sequences, as described in the methods section.
After repeated trials of cluster analyses using genes of various minimum sizes (or probe numbers per gene), we observed that most CNAs from the same tumour were clustered more closely together than any samples from another UGC case when we analysed only genes with 3 or more probes per gene: a total of 5019 genes.
It is also worth noting that there only ∼31 60% of genes from this 187 gene signature that were used in clustering analyses in each of validation datasets due to the specificity of platforms, sample availability and missing values in microarray experiments.
The epitypes were identified using hierarchical clustering analyses (HCA) of 2000 bootstrapped data sets as previously described by Lindgren et al. (Supplemental Material).
We used hierarchical clustering analyses (HCA) of bootstrapped data sets based on the CpGs that showed highest variation in DNA methylation levels (Supplemental Material) to identify four major subtypes of bladder cancer: Epitypes A, B, C, and D (Fig. 4; Supplemental Material).
The resulting gene signatures were used in hierarchical clustering analyses of the primary lung adenocarcinomas with complete linkage agglomeration, based on the Euclidean distance measure.
To avoid including axes describing variance unrelated to population structure [ 51] only the 3 to 5 largest eigenvectors weighted by eigenvalues, each accounting for at least 2% of the variance, were used for subsequent clustering analyses.
