Population, samples and DNA extraction.
We included blood samples of 50 phenotypically normal individuals (MXM, 30 females and 20 males) and 97 aneuploid patients (MXM_A, 41 with Trisomy 21, 28 with Monosomy X, 11 with Trisomy 13 and 17 with Trisomy18), all with Mexican Mestizo origin.
Genomic DNA was extracted from the blood samples obtained from the participants with the saline precipitation method (Gentra Puregene Kit, QIAGEN, Venlo, Limburg, NL).
Genome-Wide Human SNP arrays.
We analyzed the gDNA with the Affymetrix GeneChip® Genome-Wide Human SNP Array 6.0 (Santa Clara, CA, USA). This array contains 906,600 SNP probes and 946,000 non-polymorphic oligonucleotides; the median inter-marker distance over all 1.8 million SNP and copy number markers combined is less than 700 bases.
The procedures for DNA digestion, ligation, PCR amplification, fragmentation, labeling, denaturing and hybridization into the array were performed in 147 DNA samples (two DNA samples were not included, because they did not pass quality controls), according to the protocols provided by the supplier. Arrays were stained and washed in the Affymetrix GeneChip Fluidic Station 450 and scanned using an Affymetrix GeneChip Scanner 3000 7G (Affymetrix, Santa Clara, CA, USA). We analyzed the files obtained with the appropriate bioinformatics tools.
Long-contiguous stretches of homozygosity
We visualized the long-contiguous stretches of homozygosity (LCSH) in the software Chromosome Analysis Suite (ChAS) software version 4.1, provided by Affymetrix (Affymetrix, Santa Clara, CA, USA). For the analysis, we used the NetAffx 33 hg19 annotation files (http://www.affymetrix.com). For LCSH>3Mb, the analysis configuration was set at LOH with marker count = 50 and size = 3000 Kbp, and for LCSH>5Mb was set at marker count = 50 and size = 5000 Kbp.
Estimation of coefficient of inbreeding (F)
Individual inbreeding coefficients (F) were estimated using LCSH >3Mb data; F was the total length of autosomal LCSH in Kbp divided by the total autosomal size covered by the Genome-Wide Human SNP Array 6.0 (2,881,033,286 Kbp for hg19). We report the average of the F value for each group. The F value of 0.25 could reflect a first-degree parental relationship, 0.125 a second-degree, 0.0625 a third-degree and 0.03125 a fourth-degree [22].
Additional files
To compare our studied population, we used the following CEL files obtained from Affymetrix GeneChip® Genome-Wide Human SNP Array 6.0 (Santa Clara, CA, USA)
a) 99 files from healthy tissues obtained from women with Breast cancer (MXM_WBC) [29] of GEO open database (GSE87048).
b) 30 Spanish (IBS), files downloaded from GEO data set GSE67047 [30]
c) 30 Caucasian (CEU) and 10 Yoruba (YRI), part of the population studied in the HapMap project4, CEL files kindly provided by Affymetrix/ Thermofisher.
Ancestry
The MAP and PED files obtained by the Genotyping Console software (Affymetrix) were used for the ancestry analysis; we eliminated two files that did not pass quality control.
The global ancestry of 49/50 healthy Mexican Mestizo files (MXM) and 96/97 individuals with aneuploidy (MXM_A), was deducted through a supervised maximum likelihood ADMIXTURE approach, from K = 2 to K = 3 ancestral components, and compared with the global ancestry of the group MXL (Mexican ancestry from Los Angeles California US) (n=64). For this analysis, genotypic frequencies of Northern European (CEU) (n=99), Yoruba (YRI) (n=108), and Native American (NAT) (n=43), reported in the 1KGP, were considered as reference parental populations for demographic and historical reasons.
Copy number variants (CNV) calling. The CNVs were visualized in the software provided by Affymetrix (Affymetrix, Santa Clara, CA, USA). GeneChip Command Console (AGCC) software was used to generate the .CEL files and .ARR files from each microarray scanned. The .CEL and .ARR files were analyzed with Genotyping Console software (Affymetrix, Santa Clara, CA, USA) in order to obtain the CN data files (.CNCHP) and .MAP and .PED files for do ancestry. CNCHP data were analyzed using Chromosome Analysis Suite (ChAS) software version 3.1 (Affymetrix, Santa Clara, CA, USA).
The annotation file used in our analysis can be found on the Affymetrix website, listed as NetAffx 33.1 (hg19). The reporting threshold of the copy number was set at "High Resolution" settings, which give us a total of gains or losses in 100Kb with marker count ≥ 50. The copy number variants were compared among all samples. The CN analysis in ChAS software (http://www.affymetrix.com) has two ways of showing the results: one as a table that includes both coordinates and CN for each region; or a graphical interphase that displays the results of CN in an image of each of the 23 chromosomes, highlighting the regions with CN>2 in blue bars, regions with CN=1 in red bars, or CN regions with normal copy number (CN=2), which are only marked in lines (Figure 5).
To corroborate the changes in the CN obtained by GW SNP 6.0, we analyzed the absolute changes of copy number with qPCR in a representative sample. Oligonucleotides were designed to be used as primers for each locus with a change in CN and for a control region within the same chromosome with CN=2. The qPCR was made with LightCycler ® Master (ROCHE diagnostics) using hydrolysis TaqMan® probes (Universal probes, ROCHE), with 100ng of DNA, 45 amplification cycles and a single quantification. For those chromosomal regions in which it is possible to have a CN=2 control DNA, the method proposed by Livak of 2-DDCt was used [31]; for the loci where all subjects showed the change in the CN, the 2-DCt method was used [31].
For an additional CNV calling method, we used Copy-number estimation using Robust Multichip Analysis, and this method provides full-resolution raw total copy-number estimates by a pre-processing and probe summarization. CRMA v2, is available in Bioconductor [32] implemented in R (http://www.aroma-project.org/vignettes/CRMAv2/).
Briefly, the .CEL files were analyzed for quality after the program realizes the calibration for crosstalk between allele probe pairs, followed by the normalization for 25-mer nucleotide-position probe sequence effects. Next, analysis of performance of robust probe-summarization, the normalization for PCR fragment-length effects on summary signals, and finally, the calculation of full-resolution total copy numbers. When a standard reference is used, it is often the average of a pool of samples CnChipEffectFile.
The reference control samples for CRMA v2 analysis were: a) The same 50 healthy individuals of MXM; b) 30 IBS, 30 CEU) and 10 Yoruba (YRI). All groups were analyzed with the same pipeline.
We made the following comparisons: a) MXM vs. MXM reference; b) MXM vs. IBS reference; c) MXM vs. CEU reference; d) MXM vs. YRI reference, and; e) MXM vs. HapMap reference (CEU+YRI).