Clinics and Mobile Elements in WES data

Question

[German.M.Demidov]

Dear all,

I am performing an analysis of MEIs in clinical exomes (analogously to https://www.nature.com/articles/s41436-020-0749-x ) and there is something around 0.1% of cases affected by MEIs (as shown in a paper), however, I wander if I miss anything. The questions:

1) so far I detect around 3 different types of MEI using the tool described in the paper ( https://github.com/GeneDx/scramble/blob/master/cluster_analysis/resources/MEI_consensus_seqs.fa ) - are there more? where can I get a catalogue of sequences for different mobile elements?

2) we also detected large retroduplications ( https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3847774/ ) with third-party tools - is there any paper on clinical value of such insertions of whole mRNAs?

3) discussion: which are other potentially genetic disruptions except SNVs/indels/SVs[CNVs]/MEI/retroduplications which may play at least some role in some super rare cases AND are potentially detectable with WES samples? (for cancer we also detect microsatellites - but this is a different story, I don't expect them to play a role in germline)

Thanks a lot!

Answer 1

[rciskow]

Hi,

Thanks for your interest in our paper. We found a diagnostic rate of MEIs in about 0.03% of cases. The 0.1% figure comes from this paper: https://pubmed.ncbi.nlm.nih.gov/15643617/.

To answer your questions as your enumerated them:

1) We provided fasta sequences for the 3 most active families of MEIs in humans. SCRAMble is sensitive enough to detect other subfamilies based on these sequences (for example, the Alu sequence provided is a consensus AluYa5, but SCRAMble will still be able to detect an AluYb8). If you're interested in using different sequences, I recommend going to RepBase (https://www.girinst.org/repbase/) to find consensus MEI sequences.

2) The DDD cohort has recently published their analysis of MEIs and retroduplications in a clinical cohort (https://www.nature.com/articles/s41467-019-12520-y). They used MELT (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5668948/) to detect both types of insertions. That paper is a good place to start. In our own experience, we've found that rare processed pseudogene insertions affect sensitivity in detecting variants due to dilution of signal from the "real" gene. Similarly, variants that may appear to be of clinical significance might not confirm if they are coming from the processed pseudogene instead of from the original gene.

3) I would add potentially balanced events such as translocations and inversions to this list.

If you have any more questions, you can email me directly by going to the Genetics in Medicine paper and finding the corresponding author info.

Best, Rebecca