MAUDE Adverse Event Report: NEBULA EXPLORE (0.5-30X WHOLE GENOME SEQUENCING); AUTOSOMAL RECESSIVE CARRIER SCREENING GENE MUTATION DETECTION SYSTEM

Device Problem Incorrect, Inadequate or Imprecise Result or Readings (1535)

Patient Problem No Known Impact Or Consequence To Patient (2692)

Event Date 07/11/2019

Event Type  malfunction  

Event Description

I have 2 concerns, but i think only 1 of them may still be true for new customers because i believe nebula currently offers higher coverage sequencing results (30x, rather than 0.5x).My concern is that i think is still valid is the rigor of the health-related results.For example, i see percentiles (for what i believe are polygenic risk scores, even if only a limited number of variants are calculated).I think the percentiles are likely to confuse customers (since risk may be mostly normal, even with relatively high and low percentiles), and i am also not sure what is the predictive power/ reproducibility of what is being reported for the health results (or whether this has been checked by the fda).For earlier customers, i have specific concerns about the accuracy of the imputed genotypes from low-coverage whole genome sequencing data (which may not be a concern for new customers).I think the link will be removed from maude; however, in terms of providing info to the fda, there is add'l info available here: (b)(6), i also uploaded data to my personal genome project page on 07/11/2019: (b)(6), if the fda would like more info, then i would be glad to provide it.However, these are essentially my concerns: i have one copy of the apoe e4 allele for alzheimer's disease risk, but the imputed genotypes said that 2 wt/e3 alleles with higher confidence, and the accuracy when i compared my higher coverage wgs genotypes to my icwgs imputed genotypes (with custom code as well as in precision fda) was lower than you might expect at being closer to approx 90% than 99% (but still higher than i would get when i tested generating gencove imputed variants for my cat's sample).I think the imputed genotypes were acceptable for broad ancestry and close family relationships, but i don't think they should be acceptable for health results (since i don't think the genotypes for an individual variant are sufficiently accurate).Nebula genomics.Fda safety report id # (b)(4).

Event Description

Additional info received from reporter on 08/10/2021 for mw5093887.I received an e-mail on (b)(6) 2020 indicating that there was a "new" 0.4x low coverage whole genome sequencing (lcwgs) option available from nebula.After ordering another lcwgs sample and receiving my results, the company confirmed this was not true (it is really the same product).However, there was a period of time after my previous fda medwatch report (mw5093887) when lcwgs data was not available from nebula.Also, the exact points from this report are a little different than before.This time, the coverage for my individual sample was approximately 2.5x higher, and my apoe genotype was called correctly.Nevertheless, even if you only focus on my newer sample, you can see the table with percentiles for my new samples.So, i am not sure if some of the reports should be considered "medical".In general, you can see the new results and a comparison to earlier results here: https://github.Com/cwarden45/dtc_scripts/tree/master/nebula/sample2 for example, the pearson correlation coefficient between the prs percentiles for the 2 replicates is 0.8.However, some differences were very large: for example, for the "bone mineral density (kemp, 2017)" prs, the largest percentile difference being 2% for one replicate and 99% for the other.I double-checked to confirm that this was not a typo on my end (for the prs with the difference in percentiles of 97% for the 2 replicates).To be fair, nebula describes the "basic" lcwgs coverage as "medium" accuracy, and the regular wgs coverage as "high" accuracy.However, i was not sure how clear this was made to customers, or how others intended to use the results.Fda safety report id# (b)(4).

• Brand Name: NEBULA EXPLORE (0.5-30X WHOLE GENOME SEQUENCING)
• Type of Device: AUTOSOMAL RECESSIVE CARRIER SCREENING GENE MUTATION DETECTION SYSTEM
• MDR Report Key: 9865419
• MDR Text Key: 184905541
• Report Number: MW5093887
• Device Sequence Number: 1
• Product Code: PKB
• Combination Product (y/n): N
• Reporter Country Code: US
• Number of Events Reported: 1
• Summary Report (Y/N): N
• Report Source: Voluntary
• Reporter Occupation: Patient
• Type of Report: Initial,Followup
• Report Date: 08/07/2021

1 Device was Involved in the Event

1 Patient was Involved in the Event

• Date FDA Received: 03/20/2020
• Is this an Adverse Event Report?: No
• Is this a Product Problem Report?: Yes
• Device Operator: No Information
• Was Device Available for Evaluation?: Yes
• Was Device Evaluated by Manufacturer?: No Information
• Type of Device Usage: N
• Patient Sequence Number: 1
• Patient Age: 34 YR
• Patient Weight: 77