Reasons for NOCALL in a gene-level CNV

CNV_IDs are used in the target regions BED file to identify one or more amplicons that represent a single genomic region; most often CNV_IDs are gene symbols.

The CNV algorithm looks for heterogeneity and trends at the CNV_ID level, which may potentially, be indicative of data artifacts, and annotates the calls for CNV_IDs displaying these effects as SUSPECT as it is unclear whether the data for this CNV_ID indicates a true biological focal amplification.

  • The annotation SEVERE_GRADIENT indicates an increasing or decreasing trend in calculated copy number above a specific threshold (range of values > 4*MAPD or > 0.5) across adjacent amplicons. This is calculated as a moving average and is compared to the local MAPD of the CNV_ID.

  • The annotation DIFFERENT_MEAN_SIGNAL relies on a calculation, which looks at differences in groups of amplicons co-located within a CNV_ID. The amplicons are divided into two groups and if the mean difference in the log read ratio is greater than 0.5 this flag is raised.

  • The annotation LOCAL_AVERAGE_SIGNAL_VARIATION is made if any pairwise mean in log2 read ratio of adjacent amplicons within a CNV_ID has a value of less than -4 and another such pairwise mean has a value greater than -2. A CNV_ID with partial total loss usually is flagged by this heuristic.