Algorithm Information

Variable Non-Overlapping Window

This was implemented using Nathan Lazar's scripts to create constant expected reads per window and to generate observed reads per window.

Circular Binary Segmentation (CBS)

CBS was implemented using the DNAcopy R package.

Hidden Markov Model (HMM)

HMM was implemented using the HMMcopy R package.

CBS/HMM Intersect (CHI)

This was implemented using Kristof Torkenczy's scripts.

Classification Rules

1. Identify samples that have no whole genome amplification (no.WGA) and exclude these samples from the analysis.

From VNOWCHI_Summary.txt, if UniqueMappingPositions < 50,000, then "no.WGA" 

2. In each sample, determine if the impact on chromosomal regions are whole or segmental.

For each region with the agreeing CNV call of gain or loss,
  if the region is > 90% of the chromosome's length, then it's "Whole Chromosome CNV"
  if the region is >= 15Mb, then it's "Large Segmental CNV", else "small Segmental CNV"

3. In each sample, identify ploidy and sex status of each individual sample.*

Ploidy status:
  If no CNVs were called, then "euploid"
  If 1-4 chromosomes have CNVs, then "aneuploid"   
  If 5 or more chromosomes have CNVs, then "chaotic aneuploid"
Sex status:
  If no CNVs were called on the X chromosome (CN=2), then "female" (XX)
  If CN = 1 for the X chromosome, then "male" (XY)
  If any other CNV calls exist, then "unknown" and flag the sample for inspection

*regardless of the number of CNVs a single chromosome has, ploidy status only counts each chromosome with a CNV once (Ex. If chr1 has 1 gain and 1 loss, then ploidy status will only count chr1 once not twice.)

4. Identify embryo's ploidy and sex status based on samples in an embryo.

For the samples from each embryo, 
  if all samples are euploid, then the embryo is "euploid"
  if all samples are aneuploid or chaotic aneuploid, then the embryo is "aneuploid"
  if there is a mix of euploid samples with aneuploid and/or chaotic aneuploid samples, then the embryo is "mosaic”