CNV Informatics Baseline calculations

Copy number estimates are made using a proprietary algorithm.

The algorithm for CNV Informatics Baselines normalize some known sources of variability include pool imbalance, median mean base coverage of all tiles, and per tile attributes of GC proportion.

In Ion AmpliSeq™ assays, germline copy number estimates are performed in small units called tiles. Tiles are short segments of sequences that are about 100bp long. Contigs that are formed by overlapping amplicons are divided into disjointed tiles of equivalent size closest to 100bp.

When Ion Reporter™ Software calculates the mean base coverage for each tile, the same normalization that was done by the CNV Informatics Baseline creation step is performed on the test sample. The software then compares the mean coverages to expected mean coverage for the tiles in the baseline samples. The software further derives for every tile, posterior likelihood of its being every possible ploidy value (0 to 10), given the coverage observation.

The algorithm uses a model for the whole chromosome to determine the ploidy value of every tile, instead of the most likely ploidy value for a CNV segment tile by tile. A transition probability between 2 adjacent tiles to model the likelihood that two adjacent tiles are having the same ploidy value is used. The likelihood of an assignment of ploidy values to a list of tiles is the multiplication of the likelihood of the assigned ploidy of every tile times all the transitional probabilities between adjacent tiles based on the assigned ploidy values. The software finds the most likely assignment, and that is reported as the CNV result.

For details about CNV call results shown in analysis results, see the user guide for the assay that you use.