ELLA documentation

# Evidence sections: Evaluate annotation and studies

In FULL (default) mode, the main view of the CLASSIFICATION page contains a number of different sections (shown here as collapsed):


The different sections presenting annotation evidence are described below, see the separate page for the Classification section.

# Analysis specific (ANALYSES only)

This section is for information and actions specific to the variant as observed in the current sample (analysis), and displays indicators of the quality of the variant calling (HTS data, ANALYSES mode only).

NOTE

This is the only section on this page that is specific to the sample (analysis), all other sections are tied to the variant interpretation, which is independent of samples. This also means that any comments you add here will be visible in this analysis only.

# Quality information

Quality information is tied to a particular sample, and is therefore only available in ANALYSES mode. In addition, only information from HTS data is provided. The following quality parameters from the VCF file (opens new window) are shown :

  • Filter: PASS if all filters have passed, otherwise list filters that failed

  • Quality: Variant calling quality

    • Phred-scaled probability that a REF/ALT polymorphism exists at this site given sequencing data
    • 10 = 1 in 10 chance of error, 100 = 1 in 10^10 chance of error

  • GQ: Genotype quality

    • Derived from PL (see below), comparing the difference between the second lowest PL to the lowest PL (0)
    • Value is capped at 99

  • PL: Phred-scaled genotype likelihood (opens new window)

    • Values are given for genotypes [REF/REF, REF/ALT, ALT/ALT]
    • Most likely genotype is 0; the higher the other values, the less likely they are true

  • Depth: Number of reads covering the site

  • Ratio: Number of reads covering each allele, with ratio for variant allele/total

# De novo variants

In addition to the above, for de novo variants (marked with a D tag in the sidebar), a P(de novo) value (0-1) is shown for the proband, indicating the posterior probability that the de novo call is true. For details, see Vigeland et al. 2016 (opens new window).

# Warning: Needs verification

HTS variants with quality issues are marked with NEEDS VERIFICATION in red in the QUALITY card. This is shown if any of the following conditions are true:

  • Filter not PASS
  • Depth <20 (if present)
  • Allele ratio ≤0.9 if homozygous, ≤0.3 or ≥0.7 if heterozygous (if present)

# Mark variant as not relevant

The header of the ANALYSIS SPECIFIC section provides options to mark a variant as not relevant for a particular analysis. Click the NOT RELEVANT (SELECT) dropdown to view the available options (see also table below). Hovering over each option also provides a description:

You can select multiple options; click again to unselect an option. Selecting one or more options moves the variant to the NOT RELEVANT VARIANTS side bar section. If a selected option is defined as one that "Needs review", a purple N tag or red N! tag (if a warning is also present) is added to the warning column for the variant in the side bar. Note that these variants usually do not need a classification, although it is possible to have both.

These are the available options:

# Choice Description (on hover) Needs review?
1 High frequency (dominant) Reported population frequency is too high for pathogenicity in a dominant gene. Use this in cases where gene is defined as AD/AR, but patient is likely AD and can therefore use the the lower AD frequency threshold. No
2 No direct phenotype match Gene is not known to be associated with the patient's given phenotype. Medical doctor should review patient medical records for any additional clues. Yes
3 VUS low Class 3 low in accordance with ACMG SVC 4.0 guideline. No
4 Carrier P/LP (recessive) Pathogenic variant in recessive gene (cannot explain phenotype), lab doctor decides inclusion in report as carrier finding. Yes
5 Alternate molecular cause found Similar to ACMG BP5 (without actual use) No
6 Gene of uncertain significance Gene has uncertain association with patient's given phenotype. Medical doctor decides inclusion in report. Yes
7 Incidental finding (Likely) pathogenic variant that is not directly related to the phenotype under investigation. Medical doctor decides inclusion in report. Yes
8 Other - Needs review Any other reason needing review. Yes
9 Other Any other reason that does not need review. No

# Technical variants and verification

In the header of the ANALYSIS SPECIFIC section, ELLA also provides the following features for handling variants with questionable quality in a particular analysis:

  • Use the VERIFICATION (SELECT) dropdown to mark variants in need of verification with a one of the alternative methods Sanger, MLPA, aCGH or ddPCR. This adds a yellow N tag in the "Q" column of the side bar. Depending on your setup, this information can automatically be passed on to external software such as a LIMS for further processing. Once verification has been performed, you can then mark the variant as VERIFIED or TECHNICAL based on the results. This will disable the VERIFICATION (SELECT) dropdown (unselect VERIFIED or TECHNICAL to enable it again).

  • Select VERIFIED to mark the variant as verified by an independent method in this analysis. This adds a green V tag in the "Q" column of the side bar.

  • Select TECHNICAL to mark the variant as a false variant call in this analysis. This adds a red T tag in the "Q" column of the side bar and moves the variant to the TECHNICAL VARIANTS side bar section. As with NOT RELEVANT variants, these usually do not need a classification, although it is possible to have both.

Note that it is possible to set VERIFIED or TECHNICAL without first choosing a verification method, e.g. based purely on quality information/raw data.

# Region

This section displays other, previously classified variants within a predefined, genomic distance from the selected variant. Click on the link to open the other variant evaluation in a new tab.

If there are more than a preset max number of variants (e.g., 10), only the closest variants are shown.

# Frequency

This section displays population frequencies reported in external or internal datasets (if any). Note that variants with a population frequency exceeding the threshold for ACMG criterion BA1 have already been filtered out.

# Included datasets

NOTE

Any dbSNP entry that has positional overlap with the current variant is reported, and therefore may not contain the exact same variant/allele. If you use this data, you should check the dbSNP entry to verify that it has the same variant as in your sample.

Variants with quality issues reported in gnomAD display a warning in the gnomAD cards.

# Prediction

This section displays various predicted effects of the variant.

# Included predictions

# VEP consequence

Variant Effect Predictor (VEP) (opens new window) provides basic information about the location and expected effect of a variant within a transcript and protein. VEP uses Sequence Ontology terms (opens new window).

By default, only effects in the default transcript(s) specified in the gene panel are shown. However, if there are worse consequences in other, alternative RefSeq (NM_) transcripts, this will also be displayed together with a warning. To view consequences in all alternative RefSeq transcripts, click the given consequence(s).

# Add predictions

Add other types of predictions by clicking the ADD PREDICTION button:

The choices you make here are meant as “keywords”, and any details important for evaluating the choices made should always be added to the PREDICTION-COMMENTS field. Choices made here result in suggestions for relevant ACMG codes in the CLASSIFICATION section, but you still need to approve/add individual codes before final classification.

NOTE

The DOMAIN and REPEAT options are also available in the reference evaluation module (see below), and should be added there if you have a specific reference to attach this information to.

The manual options are:

  • ORTHOLOG CONSERVATION

    • CONSERVED: the amino acid is highly conserved in orthologs
    • NON-CONSERVED: the amino acid is not conserved in orthologs

  • PARALOG CONSERVATION

    • CONSERVED: the amino acid is highly conserved in paralogs
    • NON-CONSERVED: the amino acid is not conserved in paralogs

  • DNA CONSERVATION

    • CONSERVED: the DNA nucleotide is highly conserved
    • NON-CONSERVED: the DNA nucleotide is not conserved

  • DOMAIN

    • CRITICAL FUNCTIONAL DOMAIN: the variant is located in a critical functional domain
    • CRITICAL FUNCTIONAL AMINO ACID: the function of the reference amino acid is known to be critical

  • REPEAT

    • REPEAT REGION: the variant is located in a repeat region.
    • NON-REPEAT REGION: the variant is located in a non-repeat region

  • SPLICE SITE EFFECT

    • SPLICE SITE LOST: the variant is predicted to cause the loss of an authentic splice site
    • DE NOVO SPLICE SITE: the variant is predicted to cause the creation of a novel splice site
    • NO SPLICE SITE EFFECT: the variant is predicted to have no effect on splicing

Remember to SAVE (top right) once you are done.

# External

This section shows annotation from external databases, currently including HGMD Pro and ClinVar.

# Included databases

# Add data from other external databases

If you want to add results from other databases (various LSDBs, depending on the gene), press the ADD EXTERNAL DB button:

In the pop-up, select from the dropdown for each database you want to add from. For pre-specified databases, there will be a Visit database link to the right, which will take you to the corresponding database and gene. Choices for all databases except OTHER (which is free text) are:

  • Unambiguous classification:

    • Pathogenic: Class 5 or similar
    • Likely pathogenic: Class 4 or similar
    • Uncertain significance: Class 3 or similar
    • Likely benign: Class 2 or similar
    • Benign: Class 1 or similar

  • Other:

    • Conflicting: Conflicting classes reported
    • Indirectly relevant: Variant is not the same, but has indirect relevance
    • Nothing found: No entries found in database
    • Other: None of the choices above are applicable

Remember to SAVE (top right) once you are done.

# Studies & references

This section contains literature references describing the selected variant. The references have been automatically retrieved from the annotation, but you can also add studies yourself.

# Reference evaluation

Evaluate or ignore a reference by pressing one of the corresponding buttons:

TIP

Recurring, generic or otherwise irrelevant references from the annotation can automatically be ignored by changing the configuration by a system administrator.

When a choice/evaluation already has been made, the EVALUATE button changes to RE-EVALUATE.

The STUDIES-COMMENTS field at the top should be used for summarizing findings and comments from the individual reference evaluation forms:

# Reference evaluation form

The EVALUATE button will bring up the reference evaluation form. Help text for the different options provided here is available by holding the mouse cursor over the header/question.

The options in the form are meant as a guide, and you should always make a comment (at the bottom of the form) summarizing any points from the reference that will be important for the classification of the variant. This comment is also visible in the reference list and can be directly edited here once an evaluation has been performed.

# Section cards

Evaluated/new references are placed in separate cards (only non-empty cards are shown):

  • EVALUATED: References marked with RELEVANCE: YES or INDIRECTLY in the evaluation form.

  • PENDING: New references (no evaluation has been performed).

  • NOT RELEVANT: References marked with RELEVANCE: NO in the evaluation form. Note that this should only be used when the reference has actually been reviewed, but found not to contain relevant evidence.

  • IGNORED: References marked with either the IGNORE button or RELEVANCE: IGNORE in the evaluation form. This card is hidden by default; To review, press the SHOW IGNORED button:

    These references may be added back by clicking the `RE-EVALUATE` button and changing to RELEVANCE: `YES` or `INDIRECTLY`.

# Add studies

If you have found other studies/references that aren’t already in the list, you can add them by clicking the ADD STUDIES button:

The resulting dialogue lets you add studies in one of three ways:

  • SEARCH: Search the internal database for studies that have already been added but not connected to the current variant, then click ADD next to a positive search result.

  • PUBMED: Add reference data from PubMed. To add a reference:

    • On the PubMed page for that reference, download the details with Save - "Format: PubMed" and Create file:

    • Open the file that is downloaded, copy the entire contents and paste it into the RAW PUBMED DATA field in ELLA.

  • MANUAL: Add studies manually, either PUBLISHED or UNPUBLISHED (e.g., in-house) studies. Fields marked with a * are mandatory.

Last Updated: 4/22/2026, 10:07:50 AM

← Classification section →