hours only. property of centogene ag. unauthorized distribution …€¦ · order to maximize...

Title: Curation of GBA variants detected in Gaucher screened patients

Document type: SOP (english) IT support

Document ID: SOPeIT-85

Author: Digital Data Products & Curation

Owner: Digital Data Products & Curation

Approver(s): Ellen Kargesapproved at 2020-01-28 15:22 (UTC +0100)

Approval date: 2020-01-28

Effective date: 2020-01-28

SOP (english) IT supportCentogene AG SOPeIT-85Curation of GBA variants detected in Gaucher screened patients Version: 2.0

Property of Centogene AG. Unauthorized distribution or copying prohibitedGenerated by Digital Data Products & Curation at 2020-02-21 14:06:09 (UTC +0100) - Uncontrolled copy. Valid for 24hours only.

of75

1. Purpose and ObjectiveThis Standard Operating Procedure (SOP) describes the process of collection, association, update and review ofGBA variants detected in patients screened within Gaucher disease clinical context at CENTOGENE AG, into astructured and standardized format. It utilizes a combination of computer-based tools and manual review in orderto assure the accuracy, efficiency and quality of the curation process.

2. Area of ApplicationThis SOP applies to Curation departments at Centogene.

3. Terms and AbbreviationsAA: amino acid

AR: autosomal recessive

CentoLSD: Centogene’s Lysosomal Storage Disease database

CentoMD: Centogene’s Mutation Database

CI: Clinical informatin

CRV: clinically relevant variants

CuRepo: Curation repository

G2P: Genotype – to – Phenotype

GBA: b-glucocerebrosidase

GD: Gaucher Disease

HGNC: HUGO Gene Nomenclature Committee

HGVS: Human Genome Variation Society

HPO: Human Phenotype Ontology

MOI: Mode of inheritance

STD: Standard deviation

VUS: variant of uncertain significance

WES: whole exome sequencing

WGS: whole genome sequencing

4. Applicable DocumentsSOPeIT-28 Curation Repository Data Submission

SOPeIT-32 Curation Repository Software Description

SOPeIT-46 SeqPilot Import into Curation Repository

SOPeIT-48 Quality checks - CuRepo

SOPeIT-55 Curation Repository MLPA/qPCR Import

SOPeIT-56 Curation repository Biochemistry import

SOPeIT-62 Curation Repository Other Import



of75

SOPeIT-64 Curation Repository Illumina HTS Import

SOPeIT-72 Curation Repository General Quality Checks

SOPeIT-76 GBA gene- Gaucher disease association and curation

SOPeIT-78 Curation of GBA screened cases

SOPeIT-81 Classification of GBA variants following the ACMG guidelines

SOPeIT-83 Reclassification of GBA variants following the ACMG guidelines

SOPeR-36 Variant classification following ACMG guidelines at Centogene

VALeIT-58 Automatic Variant classification based on ACMG guidelines

5. ResponsibilitiesThis SOP applies to all employees responsible for curating GBA variants screened within Gaucher disease clinicalcontext.

6. Reagents, materials and devicesSoftware:

UniDB: http://ts0001.russ.CENTOGENE.internal/unidbweb/variantsearch

CentoMD®: www.centomd.com

Curation Repository: https://srv-centomd.CENTOGENE.internal/curation-repo

OMIM:https://www.omim.org

Gepado: https://gepado-prod.centogene.internal/Xpro/

CentoLSD: https://www.centogene.com/centolsd.html

7. ProcedureContents

7.1. Data organization and available options under Unique variant view

7.2 Curation of GBA variants

7.3 Curation of GBA variants by warning

7.4 Track case history

Before proceeding

Scope and goal of the case curation process

Curators are responsible for collection, association, update and review of genetic and phenotypic data ofcases analyzed at Centogene (or externally) into the structured format of the Curation repository database.

During curation process, curator utilizes a combination of computer-based tools and manual review inorder to maximize curation accuracy, efficiency and assurance of the highest level of data quality inCentoMD and CentoLSD database.



of75

http://ts0001.russ.centogene.internal/unidbweb/variantsearch

http://www.centomd.com

https://www.omim.org

https://gepado-prod.centogene.internal/Xpro/

https://www.centogene.com/centolsd.html

Data gathering and curation are procedures developed and implemented in the Curation Repositorysystem, that is compliant with the HGNC, HGVS and HPO nomenclatures

Curators undergo extensive training to ensure curation consistency and standardization. They confirm thatdatabase is error-free (items properly associated and interpreted, no undetected inconsistencies, and / ordiscrepancies against detected observations in house and external sources), and close the curation processby manual approval that reviewed and curated data comply with standard procedures established in house.

The data is gathered by a combination of manual submission and data import following a case-orientedmodel where characteristics belonging to a particular individual (patient information, clinical data,methodology and detected genetic variants) are stored and associated together. Data gathering process isinfluenced by already existing and processed master data, where genes-diseases- MOI are pre-linked.

The process is finalized by manual confirmation that reviewed and curated data comply with standardprocedures established in house (see additional documents). At database level, the reviewed item changetheir status from “pending”, “marked” into “approved” item.

In order to start curation by case, all variants detected in this case of interest must be approved. It aims atassuring that the entries belonging to an individual follow the rules for final statement closely, and that allassociated data is in agreement with the agreed guidelines. The following factors are considered as criticalfor the final statement: variant significance, patient genotype (number of clinically relevant changes, theirzygosity and location -i.e. cis vs. trans), inheritance pattern of the disorder, the sex of the patient (for X-linked diseases), the phenotypic description, and if available- levels of biomarkers.

NOTE: during variant curation, a deep understanding of ACMG guidelines is a must. Whenever the variantcuration ends with a discordant conclusion, inform immediately your supervisor!

Workflow description

Curation of variants starts with ascertaining reference sequences, standard ontology and nomenclature system. Tomaintain uniformity for variants naming, the curation scientists are following closely the HGVS nomenclature. Theuse of g. for genomic, c. for coding DNA seq, p. for protein and m. for mitochondrial is mandatory. The referencetranscript is represented by longest transcript at amino acid level. Variants described on historical nomenclatureare matched to current transcript/ nomenclature. At CENTOGENE, NCBI transcripts are used as standard referencetranscript (RefSeq Gene) and the hg19 genomic build is considered. To standardize the use of gene symbols, thecuration scientists follow the HGNC guidelines.

HGVS-compliant variants are then subjected to variant classification following the ACMG recommendations.Curation scientist evaluates all available criteria under the following categories: general population and sub-population allele frequencies and observations, computational and predictive tools, functional assays performed inhouse and / or obtained from literature, segregation and de novo observations, annotations in literature and / orexternal and available databases (including disease-specific, locus-specific and variant databases), family historyand allelic and genotyping data.

Variant-related information is predominantly extracted automatically, and when the case extracted / managedmanually.

CENTOGENE has been optimized the variant classification, implementing a new standalone criterion, namely PVS2.This very strong pathogenic criterion is assigned to variants that are confirming a deleterious effect via in vivomeasurements of biomarker levels.



of75

At CENTOGENE variants are classified using five-tiered scheme: pathogenic, likely pathogenic, variant of uncertainsignificance (VUS), likely benign and benign. An additional clinical class, namely risk factor, represents a raresituation.

Variants are in general quarterly reviewed; however, if the system notifies curation scientists on new evidences tobe available, the re-evaluation based on the new knowledge is prioritized and performed immediately.

The curation process of GBA variants within GD clinical context takes place in CuRepo software. Data are importedweekly as briefly described below:

All new patients screened for GBA and linked with the status “completed” / “finalized” in Gepado aretransferred in CuRepo (see documents related to data import in CuRepo: SOPeIT- 56, SOPeIT- 62; SOPeIT-55; SOPeIT- 46; SOPeIT- 64). During this transfer, patient details, clinical information and family data areretrieved. Additionally, the biochemical analyses are transferred and stored in the structured formatCuRepo provides. On the imported cases, the GBA genetic variants are transferred from UniDB, and ifSanger sequencing performed, SeqPilot is used as most reliable source. Once import complete, patients aregetting the status “pending” in CuRepo.

GBA variants are subjected for independent review in CuRepo

During variant curation process, ALWAYS check the issued medical report. When inconsistencies / errorsdetected, inform your supervisor!

A DETAILED DESCRIPTION OF DATA ORGANIZATION IN CUREPO IS PROVIDED IN SOPeIT-32 CURATIONREPOSITORY SOFTWARE DESCRIPTION

7.1. Data organization and available options under Unique variant view

Below a detailed overview of the available data during variant curation is provided.

Curation of GBA variants is performed as described in 7.2 subchapter under the module Unique variants;https://srv-centomd.centogene.internal/curation-repo/curation/variant-unique

1. By default, under Unique variants view, the result table is by default empty. The following items are included:



of75

https://srv-centomd.centogene.internal/curation-repo/curation/variant-unique

a. Reserve: option to reserve the variants

b. Details/ ID: the unique identifier for each CuRepo variant; this number can be selected, and curator jumps on eachvariant for more details and to perform curation (see 7.2)

c. Is Pub?: information if variant already described in external databases (i.e. HGMD, ClinVar)

d. Gene: the corresponding gene, in this situation GBA

e. Chrom: the chromosome where gene located, in this situation chromosome 1

f. gDNA: the genomic coordinates, according to hg19 build

g. cDNA: the cDNA change/ coordinates on transcript of interest, for GBA the NM_001005741.2 is used

h. Location: location within exon/ intron, according to transcript of interest, for GBA the NM_001005741.2 is used

i. Protein: the protein change according to transcript of interest, for GBA the NM_001005741.2 is used

j. Reference: publication linked to variant of interest

k. dbSNP ID: the rs ID when variant identified in NCBI

l. HGMD Accession: HGMD Accession number, if variant detected in HGMD database

m. Type on DNA Level: the change at DNA level

n. Clinical significance: the manual curated clinical significance

o. ACMG class: the variants class according to ACMG related SOPs

p. QC class: the Quality Check class, where manual curated clinical class (i.e. clinical significance) is compared againstACMG class

q. PathoScore: the highest pathogenicity score of Lyso-Gb3 biomarker level (not applied to GBA)

r. Coding effect: the effect on protein level

s. Submitter: user submitting for the first time variant into CuRepo (via manual submission or automated import)

t. Comment available: Indicates if variant is linked or not with a comment (See 7.2 below)

u. Status: indicates the quality of variant (pending, marked or public)

v. Approve option: curator has here the option to mark or approve a variant

w. Delete Option: curator has here the option to remove a variant from database

x. History: curator can see here all applied changes to a variant

Select the variant of interest

By selecting any variant ID, curator jumps on Variant view, and a new tab opens

The variant view is divided into the following categories:

1. Variant information: by default all details shown

2. External information: by default all details shown

3. Curated ACMG Rationale: by default all details hidden

4. Patients: by default all details hidden



of75

Screen shot of the GBA variant c.1604G>A, p.R535H, on by default Variant View:

On variant view, the variant of interest follows the Transcript (Gene): cDNA change rule. For the exampleabove (i.e. GBA c.1604G>A), Curation of variant NM_001005741.2(GBA):c.1604G>A is indicated.

Under this information, two search boxes (rectangles with green lines) can be used to initiate anothersearch: Gene and cDNA change. To search here for another GBA variant, go under cDNA change search box(green line) and add cDNA change following HGVS nomenclature.

c.108G>A variant search initiated from Variant View of c.1604G>A GBA variant:

Press Search

The Variant view is updated, and c.108G>A details are indicated:



of75

NOTES:

The search box of cDNA change follows strictly the HGVS nomenclature. Therefore, system notifies the user whenthe change does not follow the HGVS guidelines. Below two examples:

Example 1: cDNA change is not complete; c.108G> used instead of c.108G>A

Example 2: cDNA change is not correct: c.108T>A instead of c.108G>A

Each category of the Variant View is below in details characterized:

a) Category Variant Information:

Once the variant is selected/ searched, the following details are seen under Variant information category:

cDNA change: editable, following HGVS nomenclature

gDNA change: editable, following HGVS nomenclature

Protein change: editable, following HGVS nomenclature



of75

Location: editable, automatically indicated by the system based on Gene- Transcript- cDNA change asstored under master data

Pathogenicity score: by default on 0 for GBA gene

Type on DNA level: editable, drop-down list

Coding effect: editable, drop-down list

Clinical significance: editable, drop-down list

ACMG class: automatically calculated based on assigned rules under Curated ACMG Rationale (seeCategory Curated ACMG Rationale)

QC class: automatically calculated based on clinical significance and Curated ACMG class

Gene

Transcript: automatically based on Gene: Transcript association under Master data

Chromosome: automatically based on Gene: Chromosome location under Master data (chromosome 1for GBA gene)

Publication: automatically indicated based on metadata (see External information category)

Published status: automatically indicated based on metadata (Published when variant previouslyreported in the literature; Unpublished when variant not reported in literature)

Last edited: last user adding corrections/ updates and automatically tracked based on log in credentials

Last justification: user adding rationale and automatically tracked based on log in credentials

No of patients: number of patients carrying the variant of interest; automatically calculated based oncase- variant associations

All mandatory fields are highlighted with the red asterisk:

Example of Variant Information category for GBA c.1604G>A, p.R535H variant. Data organization and mandatoryfields are indicated:



of75

Under Variant Information, curator has the following options:

Review on one page all positive patients by clicking on the [cDNA change] option from the grey bar.Selection of the [cDNA change] opens a new tab (see below the screenshots- left and right).

Left screen: Patients view for variant GBA, c.1604G>A. The variant annotations are indicated and number of co-concurrent CRV (same gene and other genes are indicated, if the case). The result table contains one positiveindividual per one row.

The Patients details can be here reviewed to understand case- related evidences (left side):



of75

Right side_ Patients view for variant GBA, c.1604G>A:



of75

Retrieve coordinates on transcript of interest (on cDNA or gDNA level; see 7.2 below); green symbol:

Editing options: all editable fields can be here processed (see below and chapter 7.2.1)

View History: curators sees here all applied changes (if any)

Check for duplicates: under this option, system checks if variant may be twice imported / submitted

Delete variant: curator can remove the variant, if variant is confirmed to be an artefact.

View and add comment: curator can, if necessary, add comments, which will be visible to other curators. Ifat least one comment available, system notifies via a red message, on the top of Variant view

Comment notification:



of75

To see comment, select the View and add comment option: a new window opens where user can read the availablecomment(s).

View and Add comment:

Mark/ Approve: curator can mark an already approved variant, or can approve an variant, warning freevariant on status pending or marked (see 7.2)

If any change applied, Update option is ready for approval. Note that changes without pressing Updatebutton will not be saved!



of75

b) Category External information

Under this category are indicated the metadata used to review during ACMG assignment or rejections.

Under this category, data is divided into sub-category:

All transcript annotations: the cDNA changes of the variant of interest are displayed on all knownNCBI transcripts, using hg19 assembly. Curator must quickly see if the variant type / coding effect staysunchanged on all transcripts, as at Centogene the “uniform model” is applied automatically.

Example screen shot below: the GBA variant c.1604G>A is annotated against the five known GBA NCBI transcriptsusing hg19 build. Note that variant is located on all transcripts in exons, and the same protein change is observed(i.e. Arg into His)

Frequencies: the following tools are represented: ESP, 1000 Genomes, gnomAD genex (genomes andexomes), gnomad genex Subpop (subpopulations), CentoMD. Versions of each tool are indicated. Whenvariants not detected, the system provides No Record Present message.



of75

Predictions:

For missense variants the following in silico predictions tools are represented: SIFT, Polyphen,Mutationtaster

For splicing variants: ada and rf_Score

For all variant types: the conservation score using PhyloP100Way_ vertebrate to understand if thechange is conserved or not at nucleotide level

rsID indicates if the variant is known in NCBI database

External databases

HGMD is automatically retrieved showing the variant class, variant HGMD accession number and thephenotype for which variants was described. If variant not detected the HGMD, No Record Present is



of75

indicated

ClinVar is automatically retrieved showing the link (if the case) and the clinical class(es). When variantnot detected in ClinVAr, No record Present is indicated.

User cannot here edit any item; this category is only for visualization of the metadata.

User has the option to jump to original source, when information available (rsID, HGMD and ClinVar) andindicated as links

c) Category Curated ACMG Rationale

By default, this category is hidden.

Click on Curated ACMG Rationale, and the data is organized as following (for visualization see the screenshot below):

Category bar indicating the pre calculated ACMG class (indicated by class and color code; red forpathogenic/ likely pathogenic; dark orange for Uncertain; green for benign and likely benign)

Assigned rules (following the ACMG ID and the color code),

Rejected rules (following the ACMG ID and no color code; all black letters),

Supportive pathogenic evidences (including all ACMG rules supportive for pathogenic effect; red colorcode)

Supportive benign evidences (Including all ACMG supportive for tolerated effect; green color code)

Report summarizing all assigned rules (as ACMG rule ID, rule name, Strengths, Definitions andDescriptions).

Example, GBA c.1604G>A; p.R535H



of75

Curated ACMG Rationale bar . Indicates the pre-calculated ACMG class. This class automatically updateswith the classification scheme (see APPX1) via assignments and rejections actions (see 7.2.2 and 7.2.3).

Supporting pathogenic evidences: this part is in red and refers to the assigned pathogenic rules , if thecase. Under this part there is a text box (see 7.2) where user can add description during rule assignment.

Each pathogenicity supporting ACMG rule has its own place indicated by a rectangle with red line andcorresponding ACMG ID and name (for example: PVS1= Null/ truncating). A not yet assigned rule, isindicated by a gray fill. An assigned rule is indicated by a red fill. See screen shot below and compare PVS1=Null/ Truncating versus PVS2= Internal biomarker.



of75

Supporting benign evidences, this part is in green and refers to the assigned benign rules , if the case.Under this part there is a text box (see 7.2) where user can add description during rule assignment.

Each benign supporting ACMG rule has its own place indicated by a rectangle with green line andcorresponding ACMG ID and name (for example: BA1= Allele frequency >5%). A not yet assigned rule, isindicated by a gray fill. An assigned rule is indicated by a green fill. See screen shot below and compareBA1= Allele frequency >5% versus BP6= reputable source benign.



of75

Report of the assigned rules where ACMG rule ID, Rule name, Strength, Definition and Description aresummarized.

At this category, user has here the following options (described in detail under 7.2.2. and 7.2.3 subchapters):

Assign ACMG rules

Add / Edit descriptions for assigned rules

Reject assigned rules

Perform quick QC via report review

Update ACMG class

Correct QC class via assignment and rejection

Review applied changes via History option

d) Category Patients

All patients carrying the variant of interest are indicated under Patients. The number of total patients isindicated between the squared brackets, and by default no extra details are indicated.

Click on patients, and the patient ID (following Gepado Patient ID terminology) are indicated



of75

User has the option to click here on each Patient ID. Once Patient ID is clicked, a new tab opens, with UniqueIndividuals view, where patients’ details can be reviewed.

7.2 Curation of GBA variants



of75

Scope: assure that all entries belonging to a GBA variant are according to ACMG guidelines and internal establishedSOPs (following the GBA- GD specifics).

Procedure:

Go on https://srv-centomd.centogene.internal/curation-repo/ and log in with your credentials

Under Unique variants module search for GBA variants (as described above under 7.1)

To prioritize variants for curation, review the QC class (see screen shot below). The QC class calculatesautomatically the clinical significance of the variants against the ACMG pre-calculated class (see SOPeR- 36and SOPeIT-81 GBA variant classification).

The QC class indicates one of the following items: Perfect match, Partial match, Medium match, No match

Perfect match: the clinical significance of the variant matches the ACMG class

Partial match: the clinical significance of the variant matches the super-class (i.e. Pathogenic or Benign;here are expected the following variations: likely pathogenic vs pathogenic; and likely benign vs benign,respectively)

Medium match: the clinical significance of the variants detects one class under a super- class and oneunder VUS class. Expected are: Likely pathogenic / pathogenic vs VUS; or Likely benign/ benign vs VUS

No match: the clinical significance of the variant follows under discordant super- classes. Expected areto be seen variants pathogenic/ likely pathogenic vs benign/ likely benign

Under Reserve? Press and note that system informs that variant is reserved by you, and all other curatorsare notified, thus avoiding to work on the same variant simultaneously

As next step, you need to open the variant view. To do so, select under Details / ID the variant ID (blue,marked with an blue up arrow)



of75

https://srv-centomd.centogene.internal/curation-repo/

Subjected for curation are Variant information details and Curated ACMG Rationale.

7.2.1 Curation under Variant information

Curator is responsible to bring all fields subjected for curations according to HGVS terminology, CuRepoterminology and closing this step of the process with a Public variant

CuRepo system is HGVS complaint and variants must follows throughout this submodule the qualityimplemented as described in CuRepo- software related documents. When terminology does not followthese quality standards, notifies the user by highlighting thedeviating fields in red and by notifying the issueand which steps must be processed.

Below a such GBA variant where system detects empty fields:

In order to retrieve the genomic coordinates and the integrated rules for variant annotation, press thegreen symbol under cDNA Change.



of75

System automatically retrieves the genomic coordinates and fills accordingly the protein change andlocation

Process Type on DNA level and Coding effects according to SOPeIT- 28 Data submission from the availabledrop-down list



of75

Note that warnings are deactivated and all problematic fields processed. Once at least one change / actionby curator detected by the system, the option to Save the changes is active (note the Update blue option)

Press Update. The Update button changes from blue color into green color, and user is informed thatVariant is updated

The clinical significance of the variant is initially the variant class used by reporters on the medical issuedreports (via import). When clinical significance is not retrieved from the original source (UniDB, Gepado orSeqPilot), it will be empty, and system notifies that Clinical significance is empty and it must be selected.

Therefore, initially the variant is on pending and represents the reported clinical class. Curator must openand check the report in Gepado before approving the variant. At this step, curator ensures that Variantinformation in CuRepo is consistent with the reported details.

Once curator agrees with all fields under Variant information, press Approve variant.



of75

Variant changes the status from Marked into Public.

NOTE: Curator cannot approve a variant with Warnings which are not yet resolved and will inform user accordingly.This implies all fields where system has the mandatory symbol and where terminology invalid or field is blank

Once this step is complete, curator proceed with variant curation following the ACMG guidelines, andconsidering the specifics established and implemented at Centogene.

7.2.2 Curation under Curated ACMG Rationale

Subjected to this curation are variants with inconsistent Clinical significance and ACMG class. System



of75

compares the clinical significance against the ACMG class (if any), and the aim of this step is to have a QCclass of Perfect match.

Remember that following ACMG rules are pre-calculated according to ACMG automation at Centogene:PVS1, PVS2, PS1, PM1, PM2, PM4, PM5, PP2, PP3, PP5, BA1, BS2, BP3, BP4, BP6 and BP7

Curator can manually reject or edit the automatically generated rules.

The following rules are subjected to manual assignment: PS2, PS3, PS4, PM3, PM6, PP1, PP4, BS1, BS3, BS4,BP1, BP2, BP5

During variant curation following ACMG guidelines, follow closely the appropriate related GBA SOP (seeapplicable documents)

Going through ACMG class evidences, curator reviews all most up to date available information and decidesif clinical significance class must be corrected or not.

Below specific curator actions and challenges using examples as available are described. If any uncertaintyencountered during this process, contact your supervisor.

One variant curation example is provided under 7.2.3 In this chapter are provided examples by ACMGevidence type.

In the situation that ACMG class after completing the steps below are not supporting the clinicalsignificance, and QC class is not on Perfect match, follow the SOPeIT- 83 Reclassification of GBA variants.

Review of PVS1 / Null or truncating for GBA gene:

GBA variants assigned for PVS1 must follow under truncating effect definition as implemented (see VALeIT-58 Automatic Variant Classification based on ACMG guidelines (AVCA2.0)).

Here is one example where PVS1 is automatically assigned (see screenshot below). Note that PVS1 isindicated under assigned rules, the PVS1 rectangle is red- filled by default. See below c.1265_1320del;p.L422Hfs*28. Selecting the PVS1 red filled rectangle under Supporting pathogenic evidences/ SelectedPathogenic ACMG Rule, you can read the avaible description of the evidence of interest.



of75

Limitation (I): this limitation refers to the following mutation types, which are not yet included in theautomatized PVS1 execution: gross and gene rearrangements, conversions and inversions.Such variant willnot have a PVS1 assigned, and under External information no data to display (the most gross / generearrangements do not follow HGVS guidelines in the original sources in order to be automaticallyretrieved). Therefore, curator must manually act here as described below for the GBA variant c.28_115del,where deletion of exon 3 has been confirmed:



of75

In order to assign PVS1 for this variant, click on PVS1 rectangle (by default no red fill). The two options A(assign) or R (reject) are indicated. Note that system denies A or R selection without a description (Saveoption not active until description/ rationale entered)



of75

Add your rationale under Selected Pathogenic ACMG Rule: PVS1. Note that Save option is now active! Checkthe rationale against accuracy and against grammatical errors (if any; remember that grammar check isintegrated into the free text box). Once ready, and no other correction/ adjustment required, press A, andSave.

After pressing Save option, system calculates automatically the ACMG class (note that from Not classified,ACMG class changes into Uncertain), and informs that data was saved successfully.



of75

Limitation (II): this limitation refers to those truncating variants and localized within the last exon or intron.

Therefore, GBA variants localized in exon 12 (GBA transcript NM_001005741.2 contains twelve exons), arenot automatically assigned for PVS1.

Curator must evaluate the impact of the truncating mutation type on protein by reviewing functionalevidences: this can refer to other already classified pathogenic GBA truncating variant at the same codon ordownstream from the last codon of interest; or levels of Lyso-Gb1 pathological biomarker in vivo.

For example, GBA variant c.1545_*102del, p.V516X is a nonsense variant localized in exon 12. By default,PVS1 is not assigned (due to its location within last GBA exon).



of75

From the report of the assigned rules, PVS2 has been assigned based on clear Lyso-Gb1 pathological levelsobserved internally in vivo. Curator can check accuracy of this observation in CuRepo and agree to assignPVS1 ACMG criterion based on functional impact observed Lyso-Gb1 levels.

This observation is important to understand the complete types of evidences for both the variant ofinterest, and for those variants having a truncating impact on this codon (i.e. 516) or downstream but notyet associated with functional evidences.



of75

Review of PVS2 / Internal biomarker for GBA gene:

The PVS2 rules has been developed and implemented at Centogene as pathogenic / very strong type ofevidence for variants which are BA1 free.

It belongs to variant based on evidences at case level (ONLY INFORMATIVE cases are considered) on Lyso-Gb1 biomarker

The PVS2 minimum cutoff is >=10 ng/ml with a reference of <=4.8 ng/ml. Note that 4.8 ng/ml it representsthe pathological diagnostic cutoff (not the PVS2 acceptable minimum cutoff).

For GBA informative is a patient with clinical picture suggestive for GD, or no info/ known, with a hom orcomp het genotype, and Lyso-Gb1 of min 10 ng/ml, when reference for pathogenic diagnostic cutoff is<=4.8

One informative case is considered as enough to assign PVS2

Treat with caution the “asymptomatic” label, since it might me “on treatment” or “atypical/ mild” (i.e. notinformative case)

Do not consider biomarker results from follow up patients

Do not consider patients with other genotype as specified under “informative case” definition

Exception: you might apply for a PVS1 variant in presence of a missense (but not for the missense untilinformative case available to further clarify the impact on biomarker levels)

Do not consider patients with no GD suggestive picture (like Parkinson, dementia)

To curate PVS2, repeat steps as described under PVS1: select the PVS2 rectangle, edit description (note thatPVS2 has by default the description template available), assign and save.

Below the GBA variant c.1265_1320del; p.L422Hfs*28

Before editing (note that under Selected Pathogenic ACMG Rule: PVS2, the description indicates the standardtemplate for PVS2 rationale:



of75

Since PVS2 is already assigned, you need only to edit the description (replace the ADD MAX VALUE accordingly). TheSave button is thus activated, and press Save! In this situation, no ACMG class change expected to occur.

NOTES:

·Variants which are linked with Lyso-Gb1 levels between the minimum pathological diagnostic cutoff (i.e. >4.8ng/ml; reference <=4.8 ng/ml) and not reaching a maximum of PVS2 cutoff (i.e. <10 ng/ml; reference <=4.8 ng/ml)are evaluated as PS3 (see below the PS3 criterion).

·In presence of PVS2, no need to assign PS3 based on internal observations.

·A GBA variant pre-linked to PS3 due to no informative case reaching the minimum PVS2 cutoff of 10 ng/ml(reference <=4.8 ng/ml), must be assigned to PVS2 and rejected for PS3 (assigned only based on internalobservations) when all assignment PVS2 criteria are fulfilled!



of75

Review of PS1 / Same Amino acid for GBA gene:

This rule is automatically assigned, and curator should check accuracy of the variant evaluation in theoriginal source(s), when other than CentoMD.

Remember that the same amino acid change must be previously established as pathogenic regardless ofnucleotide change. Beware of changes that impact splicing rather than the predicted amino acid change.

Based on the manual evaluation, curator can approve or reject this pre-assignment. In both circumstances,curator must add a summary of the findings from literature.

Below such example for the GBA variant c.882T>G; p.H294Q. The PS1 rule has been automatically assigned,based on the information that HGMD classifies the same amino acid as DM (for the cDNA change c.882T>G).

In order to check accuracy of this statement, click on HGMD link under External information (note thatClinVar annotates the variants as VUS):

The link leads to a new tab, where the internal HGMD interface opens and details are ready for evaluation:



of75

From this view, observe that under Literature citation, the last three PMIDs are linked with Parkinsondisease, not with Gaucher disease. For the first PMID, there is no Comment / notes. Click on PMID10649495. The PubMed view opens. Note that this mutation was analyzed in the context of type 2 Gaucherdisease (highlighted in blue, your variant of interest H255Q) and this publication must be reviewed to clarifythe pathogenic impact on protein.

Remember that variants described as DM/ risk factor for other than GD phenotype must be excluded!

Identify the key findings, summarize under the PS1 text field via editing, and save. For this variant, there ispathogenic supportive evidences and therefore, curator agrees with PS1 assignment.



of75

Review of PS2 / Confirmed de novo for GBA gene

As GD is AR, the de novo inheritance is not expected to occur in GD patients typically

There are few GD patients described in the literature carrying one GBA genetic variant inherited from oneparent, and one GBA genetic variant arose by either germline mosaicism or as de novo mutation, in theother parent (for review see Alfonso et al., 2008; PMID 18313951; Saranjam et al., 2013 PMID 22713811)

Such rare situation can be identified by screening of the index patient and his/ her parents.

So far, at Centogene there is no such case identified. Nevertheless, there is a theoretical possibility, andtherefore curator must be aware of such uncommon situations.

Review of PS3 / Functional assay for GBA gene

This ACMG rule referring to functional studies supportive of a damaging effect implies observations ofenzymatic activity in vitro and/ or in vivo. GBA variants associated already with PVS2 rule go through PS3evaluation in a manual manner whenever a PMID is linked to the variant of interest (meaning that varianthas been at least one time reported in the literature). The observations from literature are processedaccording to the SOPeIT- 81 GBA classification SOP and are restricted to those characterized in GD-patients.

Internal observations (+PVS2) are concordant with published statements, and no conflict identified

i) Internal observations (+PVS2) are concordant with published statements, and no conflict identified

ii) Internal observations (+PVS2) are concordant with published statements, and partial conflictidentified

There are different situations expected to happen:7.1.



of75

iii) Internal observations (+PVS2) are discordant with published statements (including full conflict)

iv) No internal observation available (-PVS2), and concordant published statements

v) No internal observation available (-PVS2), and discordant published statements

vi) Internal observations (-PVS2) supporting PS3 (enzyme activity, pathological Lyso-Gb1 withinpathological range but below PVS2 cutoff

Subjected to PS3 review are all GBA variants linked with an external publication. Curators receive exportsbased on CuRepo content with the pre-associated PMID (s). At database level, curator sees these variantswith entries under References (note the screenshot below; three GBA variants are linked to correspondingpublication)

Below, indications how to proceed for each situation are summarized; in case of unclear or not yet describedsituations, contact your supervisor.

i) Internal observations (+PVS2) are concordant with published statements, and no conflict identified

This situation refers to GBA variants pre-assigned with PVS2 (i.e. internal pathogenic Lyso-Gb1 levels observed invivo) and supportive damaging effects in the literature. It is important that all publications evaluated, are linked tothe GD phenotype and all experiments performed in vitro or in vivo are conflict- free.

Although, the PS3 ACMG rule will not be considered for calculating ACMG class (PVS2 is internally established andvalidated as pathogenic very strong criterion), it is important to gathered as much knowledge as possible, toanalyze and understand how Centogene’s classification scheme performs compared with other laboratories.

Example: GBA variant c.1604G>A¸ p.R535H. For this variant PVS2 is assigned. Under External databases, HGMDlabels the variant DM in the context of Gaucher disease (see hgmd_ phenotype), and ClinVar as pathogenic.



of75

Open the HGMD and ClinVar links and summarize the functional findings accordingly. For this variant there is noconflict and all sources support a damaging effect of this variant in GD context.

Therefore, select the PS3, add your description and save. Variant is updated, and ACMG class unchanged:

ii) Internal observations (+PVS2) are concordant with published statements, and partial conflict identified

This situation refers to conflicting observations in the literature, where Centogene’s internal data (i.e. PVS2) confirmthe damaging effect described by some authors. The right action here is adding description as identified in theliterature (i.e. discordant results and summarizing the main findings), and reject the PS3 (based on the internal



of75

SOPs, we do not score inconsistent or conflicting information). The rejection of PS3 will not impact the ACMG class(PVS2 is pathogenic very strong) but offer a complete analysis and understanding of the up-to-date knowledgegathered internally and in the literature.

Example: GBA variant, c.485T>C; p.M162T. For this variant PVS2 is assigned. Under External databases, HGMDlabels the variant DM in the context of Gaucher disease (see hgmd_ phenotype), and ClinVar as VUS.

Open the HGMD and ClinVar links and summarize the functional findings accordingly. For this variant there isconflict identified, and therefore add description accordingly, and reject the ACMG rule. Variant is updated, andACMG class unchanged. System indicates the rejected rule as white fill and selecting the PS3 rule, the description isdisplayed:



of75

iii) Internal observations (+PVS2) are discordant with published statements (including full conflict)

This situation refers to the theoretical case that GBA variants are linked to PVS2 based on internal evidences, butthe most/ all of the available observations in the literature are against damaging effect (i.e. BS3). In this situation,BS3 (see below a detailed description of the BS3 evaluation ) is rejected; the stored information under BS3 is importantfor further analysis and validation of PVS2 superiority compared with external findings.

iv) No internal observation available (-PVS2), and concordant published statements

Some GBA variants identified in GD context cannot be yet linked to GD- informative cases. This refers usually to thescreening of GD- carrier individuals (parents or relatives of GD- confirmed cases; GD- positive family history), andwhere Lyso-Gb1 measurements are not informative to further clarify the variant.

Therefore, PS3 evaluation is mandatory and review of the existing literature is initiated. PS3 is assigned when onlydamaging effects are described in vitro or in vivo (no conflicting or tolerated effect reported). In this circumstance,PS3 counts for the calculation of ACMG class.

Example: GBA variant c.1609T>C; p.*537Rext*15. This variant is free of PVS2 and the only patient positive for thisvariant is a GD carrier individual. Therefore, the enzymatic activity assays and biomarker analysis in house are notinformative to clarify further the variant class. According to External information, this variant has been previouslyreported, and is annotated as disease causing in HGMD for GD phenotype (reason for PP5 assignment). Thisvariant is not identified in ClinVar database.



of75

Review the publication and summarize the functional characterization of this variant. Once the supportive findingsare complete, press A (assign) and save.

v) No internal observation available (-PVS2), and discordant published statements

As described under iv, in absence of internal observations and literature findings are not consistent, concluded orenough, PS3 is not assigned, but summary of the literature evaluation is documented under description, and PS3 isrejected.

Example: GBA variant c.1483G>C; p.A495P. For this variant are under External information discordant informationdetected in HGMD (DM) and ClinVar (Benign).



of75

From literature review based on HGMD inclusions, there are conflicting information which do not support the DMclass: it was initially identified as part of a complex disease allele by Latham in 1990 and reclassified as possiblebenign variant by Jesus in 2016. ClinVar lists this variant as benign (clinical testing, variation ID 93450). Based onthese observations add the summary and reject (press R) the PS3 ACMG rule (rectangle turns into white).

vi) Internal observations (-PVS2) supporting PS3 (enzyme activity, pathological Lyso-Gb1 within pathological rangebut below PVS2 cutoff)



of75

Variants linked at case level with biomarker values between pathological diagnostic value and PVS2 cutoff areassigned for PS3

Lyso-Gb1 levels: >4.8 but <10 ng/ml (when reference is <=4.8 ng/ml)

Variants linked only to internal enzymatic assay are assigned for PS3 when pathological. Do not assign forborderline, slightly decreased and analytical failure. At least one informative case (definition is kept) must beidentified to fulfill all requirements for PS3 in this situation.

When variant is linked already with a publication, perform literature review as described above. Document thefindings as described above under this type of evidence.

Review of PS4/ Affected > controls for GBA gene

This ACMG rule is not evaluated for GBA gene in the context of GD.

Review of PM1/ Hot spot- critical domain for GBA gene

This rule runs automatically, and does not require manual curator interaction.

Review of PM2/ Absent in controls for GBA gene

This rule runs automatically, and does not require manual curator interaction

Review of PM3= AR in trans for GBA gene

This rule is subjected to manual review by the curator.

CuRepo system automatically calculates the CRV variants within GBA gene based on the previously curatedcases (see SOPeIT- 78 Curation of GBA screened cases). Remember that number and zygosity of the GBACRV variants leads to genotype call.

To view information on c-concurrent CRV, from Variant view, select the cDNA change. Below the GBA variantc.1504C>T; p.R502C

New tab opens where presence of co-concurrent CRVs is indicated initially by number of genetic variants(see screenshot below; under Number of associated CRV is 4). The result table indicates all patients, andunder Patient Genotype focus on Compound heterozygotes.



of75

Select and open the Number of associated CRV

From this view you identify the number of co-concurrent and clinically relevant GBA variants (in the examplebelow the c.1504C>T co-localizes in trans with other four GBA variants – c.1448T>C; c.259C>T; c.115+1G>Aand c.367G>A) and number of corresponding patients (in the example below are in total 7 compoundheterozygotes individuals) is indicated:

Remember that for GBA variants, Lyso-Gb1 levels are the indicator (in absence of family members) if thetwo heterozygous GBA genetic variants are most likely in cis or trans. When the two CRV variants are in cis,the Lyso-Gb1 levels are outside pathological range (i.e. <10 ng/ml). Such individual has the genotype on“other/ complex” and evaluated as “carrier”. When the two CRV variants are in trans, the Lyso-Gb1 levels are



of75

within pathological range (i.e. >=10 ng/ml). Such individual has the genotype on “compound heterozygote”and evaluated as “affected”. In the presence of parents, the phase of the two GBA variants are additionallysupporting the Lyso-Gb1 levels.

Searching under the main result table under Genotype for compound heterozygote and under Statementfor affected, 7 individuals out of 19 (see screenshot above) are identified. Note that Lyso-Gb1 biomarkerlevels are pathological. These individuals are informative for PM3 assignment. To do so, select from variantview the PM3 and add the rationale accordingly. Assign and save.

Review of PM4/ Protein length for GBA gene


Review of PM5/ Different AA for GBA gene

This rule must be reviewed as described under PS1 ACMG rule. The system automatically retrievesinformation if a different amino acid at the same codon has been previously reported as DM in HGMD,pathogenic/ likely pathogenic in ClinVar and CentoMD.

When other variant is reported pathogenic/ likely pathogenic according to CentoMD, no evaluation isexpected.

When variant is detected in the external sources only (HGMD and/ or ClinVar), no information about qualityof annotation is provided. To avoid an annotation in these sources which is not according to the initialpublication, curator needs to check the publication and at the end to decide if agrees or not with PM5 pre-assignment. In case curator does not agree with pathogenic/ likely pathogenic statements, rejects PM5 andadjusts the description (summarizing the identified evidences).



of75

Curator needs to adjust the rationale according to the identified findings if agrees with PM5 assignmentsbased on supportive pathogenic/ likely pathogenic statements authors provides in their work.

The aim is to adjust if the case the accuracy of PM5 assignment/ rejection and bring the rationality up-to-date to be used then in downstream processes.

Example: GBA variant c.1292A>G; p.N431S. This variant bas been pre-associated with PM5 and PP2 ACMG rulesbased on pre- assignment process. In HGMD this variant (p.N431S) has been classified as DM? in HGMD in a non-GD clinical context. However, based on HGMD content, another amino acid (i.e. p.N431I) has been published as DMfor GD phenotype.

Search on for GBA and p.N431I mutation.

Open and evaluate the two available publications, summarizing the findings and deciding if DM annotation fits tothis variant class.



of75

Based on the supportive evidence, the variant p.N431I is accurately annotated as DM in HGMD. Adjust rationaleaccording to the findings and save (compare the automatically generated rationale against the manually adjusteddescription).



of75

Review of PM6/ Assumed de novo for GBA gene

This rule does not apply to GBA- GD association. Be aware of the rare situation described under PS2=Confirmed de novo.

Review of PP1/ Segregation for GBA gene

This rule is manually to be adjusted by curator if the case. Most likely this rule is supposed to be processedfor patients screened initially against WES or WGS (not Gaucher screening).

To assign this rule you need either at least two independent trios (parents and index) or at least one familycomprising minimum five- family members but three affected.

A perfect genotype- phenotype correlation must be detected in order to assign PP1. Family members whereCI are missing or are not informative, are excluded. Be careful with “asymptomatic”status (i.e. healthy or ontreatment)

Example: GBA variant, c.1342G>C; p.D448H. This variant has been identified in 100 individuals. Among them, thereare six independent families: 2061090, 2027809, 2083650, 216663, 2166692 and 2348084 respectively. Below the



of75

family ID 206109 is depicted to describe a specific familial observation. The index patient was clinically andbiochemically suspected for GD. A positive genotype – phenotype correlation has been observed by genetic testingof the patient and her parents. Additionally, the pregnant mother run a prenatal testing of the fetus, which wasidentified homozygote for the GBA variant identified in the index patient.

The other five families are represented by trios (index patient and parents), where genotype- phenotypecorrelations were identified. Below the family ID 2166663:

Based on these six independent observation, you assign PP1 for this variant, summarize the findings and save.



of75

Review of PP2/ Low rate benign missense for GBA gene


Review of PP3/ In silico or computational for GBA gene


Review of PP4/ Phenotype – family history for GBA gene

In some instances, when enzymatic analysis and Lyso-Gb1 biomarker not possible to measure, other typesof evidences shall be considerd. This ACMG rule is such criterion, when detailed information about positiveGD family history received, variant not described elsewhere and only clarification about carrier statusrequested.

Therefore, at Centogene the PP4 criterion is applied in the absence of GD- informative cases in house, and /or in absence of any external databases to further clarify the variant clinical class. Most likely, PP4 isevaluated for carriers with positive GD- specific phenotype in the family.



of75

Example: GBA variant, c.358T>C; p.F120L. This variant is automatically assigned for PM2, PP2 and PP3; moreover,not described in HGMD and ClinVar.

At Centogene there is only one patient requesting a carrier test for GD. Curator needs to investigate the “reason”for such a request. Therefore, search in Gepado / Documents / Request form and understands if any family historyof GD indicated, and summarize the findings in both CuRepo/ Clinical information module and under PP4

From the documents we understand that patient is 58 yrs. old, healthy with two siblings (one brother and onesister) confirmed with GD and deceased.



of75

Curator can assign (and additionally requests approval from medical director) PP4 based on the evidences abovedescribed, with no against pathogenicity observation.

Assigning in this case PP4 will not affect the ACMG class (variant follows under Uncertain definition), but will informabout the strong positive GD- history in the family of this individual.



of75

Positive family history (relatives confirmed to be Gaucher patients).

Review of PP5/ Reputable source for GBA gene

This rule is automatically generated, however requires some additional evaluation by curator. As reputablesource three databases are called: CentoMD, HGMD and ClinVar. Some circumstances can be encounteredby curator:

The only reputable source is CentoMD (including when CentoMD is discordant to HGMD/ ClinVar). It isexpected that assignment of PP5 based on CentoMD is associated with other types of evidences(biochemical, phenotype related, segregation etc.). Therefore, GBA variants linked with PP5 based onlyon CentoMD but no other evidence (to easily understand why CentoMD calls pathogenic/ likelypathogenic) implies searching for additional criteria (before agreeing with PP5 assignment). If curatordoes not find that piece(s) of evidence(s) based on which CentoMD says pathogenic/ likely pathogenic,its rejection is expected.

Only one external database (HGMD or ClinVar) contributes to PP5. In this situation, curator mustanalyzed the evidences for pathogenicity statement (to avoid any confusion or miss-classification).When those evidences are available (for example PS3 assigned and functional evidences summarized)and consistent with PP5, no additional evaluation is required. As in situation above, there is a need tounderstand the underlying evidence for “that reputable source” to conclude variant pathogenicity.



of75

Example: GBA variant c.203del; p.P68Rfs*23. Automatically, PVS1 and PP5 are assigned.Under PP5 is indicated thatCentoMD classifies the variant as pathogenic. No other evidence is activated, and therefore the reason forclassifying this variant as pathogenic in CentoMD is not yet transparent.

In order to agree or not with PP5 assignment, curator needs to understand the type of evidences (except PVS1) tobe assigned (if any) in order to accurately use CentoMD as reputable source to support pathogenicity of thisvariant.

The only patient linked with this variant is a carrier individual, and enzymatic activity or biomarker measurementswill not be informative.

Going through the other types of evidences, note that variant is absent in all external databases and below thecutoff expected for AR genes (i.e. 0.005) in CentoMD.



of75

Based on this internal observation, assign PM2

Under External databases, HGMD classifies this variant as DM for Gaucher disease (GD).

Review the publication and summarize the findings. This variant was observed in trans with another GBApathogenic variant (1226G) in one patient with moderately severe type I GD, and was considered by authors aslethal variant. No functional study was conducted. Based on these observations, you can assign PM3.



of75

Once you agree on accuracy on PP5 (based on PM2 and PM3 assignments), update the rationale, including asreputable sources, both CentoMD and HGMD databases.



of75

Review of BA1/ Allele frequency >=5% for GBA gene


Review of BS1/ Allele frequency < disease frequency for GBA gene

This rule is not at the moment evaluated for GBA- GD associations

Review of BS2/ Observed in healthy for GBA gene

This rule runs automatically, and does not typically require manual curator interaction.

The only specific information refers to the genotype of the “healthy” individuals (homozygotes, since GD isAR disease) and “healthy” definition (i.e. no misunderstanding in asymptomatic translation as healthy, whenpatient already on ERT / or other GD- specific treatments).

Review of BS3 / Functional assay against for GBA gene



of75

This evidence type is subjected to manual evaluation.

In order to assign BS3, the results must originate from GD- informative cases (i.e. homozygote orcompound heterozygote patients) with Lyso-Gb1 levels within normal range (i.e. <=4.8 ng/ml).

As functional evidences are used enzymatic activity, Lyso-Gb1measurements and if the case findings fromliterature. Note that benign functional observations must originate in GD- informative cases.

The same cautions (as under PVS2) are here kept:

Treat with caution the “asymptomatic” label, since it might me “on treatment” or “atypical/ mild” (i.e notinformative case)

Do not consider biomarker results from follow up patients

Do not consider patients with other genotype as specified under “informative case” definition

Do not consider patients with enzymatic results which are borderline, slightly decrease or preanalyticalfailure.

Only consistent findings (i.e. benign/ tolerated) are used to assign BS3. Inconclusive or conflicting findingsare documented, but BS3 rejected.

Internal observation on beta-glucocerebrosidase and / or Lyso-Gb1 are treated as most reliable functionalevidence.

In the presence of PVS2 assignment, but discordant evidences from literature (see PS3 ACMG rule, point iii),curator adds under BS3 the external findings and rejects the rule.

In presence of normal beta- glucocerebrosidase activity and / or Lyso-Gb1 within normal range (i.e. <10ng/ml) curator must ensure that individual is not on ERT/ or other GD- specific treatment, before closing thevariant evaluation.

Example: GBA variant c.1223C>T; p.T480M. This variant is associated with BS2 and BP6 evidences, in anautomatized way. This variant was assigned for BS2 based on observation of this variant in homozygote state intwo sources: CentoMD and gnomAD. Under reputable source there are two databases indicated CentoMD andClinVar. Note that HGMD classifies this variant as DM? for GD phenotype.

There are 115 patients carrying this variant at Centogene. Therefore, to understand the type of evidences (exceptthe BS2) for CentoMD to classify it as likely benign curator needs to evaluate the case-specific observations like



of75

internal functional evidences (or co-occurrences; this evidence is described under BP5 below for the same variant).

To review the functional assay, select the cDNA change from Variant information:

From result table with positive patients for this variant, use under Zygosity “hom” and under Genotype “wild” torestrict the 115 patients to those being homozygotes for this variant and carrying no other GBA CRV variant.

From 115 patients, four are matching the search criteria:

Activate the option “Show more information on biochemical”:

The result table extends and the results from enzyme and biomarker are indicated:



of75

Three homozygotes with no other GBA variant identified (i.e. with wild-type genotype) have normal Lyso-Gb1levels. Two of them with a reference of <=4.8 ng/ml. Therefore, assign BS3 based on Lyso-Gb1 measurements, andsave. Thus, there is a clear understanding based on which evidences CentoMD classifies this variant as likelybenign.

Review of BS4 / No segregation for GBA gene

This rule is reviewed manually by curators and refers to the patients screened via WES or WGS with aphenotype GD- related or not.

In order to assign this rule, at least two trios are required, or a family comprising at least 5 members.

All GBA variants not segregating within these families with GD- phenotype are assigned for BS4 andrationales generated; remember that inconsistent observations are not expected to be assigned; but can bedocumented and BS4 rejected

No example can be provided at this moment.

Review of BP1 / Missense in LOF for GBA gene


Review of BP2/ Co-occurrences cis/ trans (AD/AR) same gene for GBA gene



of75

This rule requires manual curator interaction

For GD as AR disease, segregation data is required to evaluate a variant that possible in cis with anotherpathogenic GBA variant. Therefore, in order to assign this rule, parents must be screened and the phase ofthe GBA variant of interest confirmed.

No example can be provided at this moment.

Review of BP3 / Repeat in frame for GBA gene


Review of BP4/ In silico or computational benign for GBA gene


Review of BP5/ Diagnosis confirmed by other variant or gene for GBA gene

This rule implies manual evaluation by curator

Curator gathers information on two different aspects: i) genetic diagnosis of GD was clarified by anotherGBA variant(s) or/ and ii) genetic diagnosis was confirmed by other gene (no GD confirmed). For the 2nd

diagnosis precaution must be taken when dual diagnosis possible.

Example: GBA variant c.1223C>T, p.T480M (previously introduced under BS3 ACMG rule). Initially, this variant wasassigned for BS2 and BP6 in an automatized way. Based on the internal function assay, curator manually assignedthe BS3 rule (see above curation of BS3 evidence).

To further gather information and summarize all type of evidences, proceed with co-occurrences findings. To do so,from the individual view, focus on associated CRV. Note that this variant co-localizes with 8 other GBA variants, andwith other 4 CRV in other genes.

Open the Number of associated CRV (i.e. all GBA co-localizing CRV variants):



of75

To understand if homozygotes patients for this variant, the GD- phenotype was confirmed by other GBA CRVvariant, go under the result table, and under Statement “affected” and under Zygosity “hom”. Six patients have theGD- confirmed phenotype with other GBA CRV variants, and pathological Lyso-Gb1 levels.

Open the Number of associated CRV other genes. Four patients carrying CRV in four different genes: GLA, NPC1,LRRK2 and SMPD1.



of75

Open each patient (by selecting Number of patients and reviewing in database if the patient linked with a geneticdiagnosis. Below are the steps of first patient (i.e. GLA gene, patient 1211196) shown.

After selecting “1” for GLA under Number of patient, you land here:

After selecting the 1211196, new tab opens. From this view, you can see if any diagnosis confirmed. For this patient,Fabry disease was the genetic diagnosis confirmed.



of75

Repeating these steps for the remaining patients, note that for NPC1 gene, patient is confirmed with Niemann -Pick type C1; for LRRK2 gene the patient is confirmed with Parkinson type 8; and for gene SMPD1, the patient isconfirmed with Niemann- Pick type A/B.

Based on this observation, assign BP5, summarize the findings and save.



of75

Review of BP6/ Reputable source benign for GBA gene

The BP6 rule is automatically generated, however requires manual evaluation by curator. The BP6 rulesmust associate with other benign types of evidences to understand the underlying reasons for the reputablesource to classify it as benign/ likely benign.

As mentioned under its compatible rule, i.e. PP5, the reputable source (CentoMD, HGMD and ClinVar) issubjected to extract the supportive benign findings

As example you can use the GBA variant c.1223C>T as evaluated for BS3 and BP5. This variant was initiallyassigned for BP6 based on two sources: CentoMD and ClinVar. To understand the evidences leading to thebenign class by CentoMD, curator performed case- specific evaluation for functional assays and co-occurrences.

All GBA variants linked to BP6 but with no other / not enough evidences returns via QC checks back tocurators.

Review of BP7/ Synonymous for GBA gene


7.2.3. Example of GBA variant curation

Below it is described the process of complete GBA variant curation using one example

The final outcome of variant curation is to obtained a variant on status Public, and Perfect match for QCclass

GBA variant c.731A>G, p.Y244C

From Variant information category, there is a discrepancy between the class used for reporting (i.e. Pathogenic)and the ACMG pre-calculated class (i.e. Uncertain). The QC class is Medium match. This is a missense variant.



of75

Open the Curated ACMG Rationale: Three pathogenic rules are pre-assigned: PP2, PP3 and PP5

Review the assignment report: this missense variant is predicted to be deleterious and two reputable sources areannotating this variant as disease causing (CentoMD and HGMD).



of75

Review under External information the Frequencies: this variant does not exist in any included external database(only CentoMD), and consider to assign PM2 after gathering case-specific evidences.

Review the data from External databases: This variant is not detected in ClinVar, and in HGMD is annotated as DMfor GD type 1.



of75

Only one patient is identified at Centogene.

Therefore, open the patient and first understand if this individual is informative for GD phenotype.

This patient is a carrier for GD, with one heterozygote pathogenic GBA variant (i.e. the variant of interest).

The Lyso-Gb1 levels are < 10 ng/ml (i.e. normal).



of75

Therefore, this patient cannot be used as GD- informative case.

In order to clarify further the variant according to ACMG terminology, proceed with literature check.

To do so, select the HGMD link from External information/ External databases:

New tab opens with HGMD details for the variant of interest:

Since no information about evidences documented in HGMD, select and open publication. Search for the variant ofinterest (i.e. 205; see under Codon number that in the paper instead of codon 244, 205 is used).

A search in the publication reveals the following findings:

One patient carries this variant in trans with another GBA pathogenic variant (p.L444P).

For this patient a clinical suspicion of GD disease (type 1) was suggestive based on clinical picture, deficientenzyme in cultured fibroblast (all mutations characterized in this paper ranging from 5.5 to 16% of normalwild type), and by presence of Gaucher cells in the bone marrow aspirates.



of75

Based on this publication, curator assigns PS3 and ACMG class changes from Uncertain into Likely pathogenic. TheQC class changes from Medium class into Partial match.

Based on the same publication, assignment of PM3 is expected, since the confirmation of the clinical andbiochemical suspicion was via in trans location with another GBA variant, to clarify the GD which is an AR disease.Assign and save.



of75

The assigned rules update, and the ACMG class based on ACMG rules combinations remains within Likelypathogenic class.

Finally, assign PM2 (see above the Frequencies) based on its very rare frequency for an AR disease. Then pressSave. Assigning this rule leads to update under ACMG class (Pathogenic) and QC class (Perfect match).



of75

Review once again the remaining unprocessed rules:

PVS1 does not apply, since this mutation is not a truncating (but a missense).

PVS2 cannot be applied due to no GD-informative case at hand.

No other AA or different amino acid at this codon has been described to date as being pathogenic/ likelypathogenic in any of the analyzed database (PS1; PM5).

PS2 and PM6 do not apply (GD is AR, and de novo is expected to occur rarely; no parents avaible)

PS4 is not at the moment a key type of evidence for GD

Variant is not detected to be located in any hot spot or critical domain (PM1).

The variant is not an in frame change (PM4) and therefore does not apply for missense.

Due to no family at hand, PP1 is not applicable.

Although the phenotype of the published patient is specific to GD, at least two independent observationsare required (therefore, PP4 does not apply here).

Look then over benign evidences. No benign evidence is pre- assigned. Re-evaluation if any evidence might



of75

support a benign class, leads to no evidence against pathogenicity (frequencies, functional assays, no sourceindicating a benign / tolerated impact; no families for segregation analysis).

To close the variant review, analyze once again the assignment report. Every assigned rule is associated with adescription, and description matches the definition.

When no change / edit required and variant associated with Public status and Perfect match, variant curation isconsidered complete.



of75

7.3 Curation of GBA variants by warning

Bioinformaticians are regularly running scripts to detect any deviations from the current SOPs, process known asQC checks (see SOPeIT-72). The QC are run as following:

1. Quarterly, before CentoMD and CentoLSD content release

2. Monthly / quarterly according to the metadata version updates

Curators may expect to receive re-evaluation notifications from the bioinformaticians whenever one of thefollowing situations is identified:

The GBA variant does not anymore follow the rare definition

The GBA variant is with the new metadata update discordantly annotated

The GBA variant is now linked to an informative case

The GBA variant does not correlate with phenotype (no segregation)

The CRV GBA variant is identified in healthy/ asymptomatic

The CRV GBA variant is now linked with normal Lyso-Gb1 levels in informative GD patient

ACMG pre-annotated rules change the status based on fulfill criteria (PS1, PM1, PM2, PM4, PM5, PP3, BS2,BP3, BP4, BP7)

All these warnings are manually reviewed by curators within CuRepo system and acts according to all related GBAprocedures.

7.4 Track case history

System track automatically all changes applied during variant curation.

In order to understand the applied changes, select the History option, under the assignment report

Once History selected, a new window opens. The changes are highlighted in yellow.

For example, the GBA variant c.371A>G used as example, is linked with revisions IDs whenever Save actionhappens. Under the most-up-to- date revision, the PM2 rules changed the status from rejected intoassigned; and this assignment led to change under the Classification.



of75

On the right side of the screenshot, note that system automatically tracked the editor (i.e. curator), the date andtime and the change type (i.e. update)

8. References

9. Appendices1. SOPeIT-85 APPX1_Classification scheme v4 pba.docx

2. SOPeIT-85 APPX2_Training module GBA variant curation pba.xlsx



of75

https://app.qualio.com/attachments/427869?download

https://app.qualio.com/attachments/427870?download