You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Additionally, splice variants are scored according to `Using the ACMG/AMP framework to capture evidence related to
17
+
predicted and observed impact on splicing: Recommendations from the ClinGen SVI Splicing Subgroup <https://doi.org/10.1016/j.ajhg.2023.06.002>`_.
18
+
In this case Exomiser will assign the PVS1, PS1, PP3, BP4, BP7 categories.
19
+
16
20
It is important to be aware that these scores are not a substitute for manual assignment by a qualified clinical geneticist
17
21
or clinician - The scores displayed utilise the data found in the Exomiser database and are a subset of the possible
18
22
criteria by which to assess a variant. Nonetheless, in our benchmarking on the returned cases from the 100K Genomes Project,
@@ -25,89 +29,215 @@ recall from 91% to 75%).
25
29
26
30
Exomiser is capable of assigning the following ACMG categories:
27
31
32
+
Variant Classification Evidence
33
+
===============================
34
+
35
+
These categories are taken from Figure 1. of `Richards et al. 2015 <https://doi.org/10.1038/gim.2015.30>`_, those in **bold**
36
+
are assigned by Exomiser.
37
+
38
+
39
+
Population Data
40
+
===============
41
+
42
+
For the population data criteria, all frequencies are considered using the populations set by the user in the
43
+
:ref:`frequencysources`, apart from any bottle-necked populations not recommended for frequency filtering from gnomAD
44
+
according to their `filtering allele frequency <https://gnomad.broadinstitute.org/help/faf>`_ document. This excludes
45
+
the Ashkenazi Jewish (ASJ), European Finnish (FIN), Other (OTH), Amish (AMI) and Middle Eastern (MID) populations. In
46
+
addition the LOCAL frequency will also not be used.
47
+
48
+
BA1
49
+
---
50
+
`MAF is too high for disorder`
51
+
52
+
Given Exomiser will filter out alleles with an allele frequency of >= 2.0%, this is unlikely to be seen. However, alleles
53
+
with a maximum frequency >= 5.0% in the frequency sources specified will be assigned the `BA1` criterion. Variants listed
54
+
as being excluded from this category by the ClinGen SVI working group `BA1 exclusion list <https://www.clinicalgenome.org/site/assets/files/3460/ba1_exception_list_07_30_2018.pdf>`_
55
+
will not be marked as `BA1`, assuming they survived variant filtering.
56
+
57
+
58
+
BS1
59
+
---
60
+
`MAF is too high for disorder`
61
+
62
+
Assigned to autosomal variants with a maximum non-founder population allele frequency >= 1.5% or mitochondrial variants
63
+
with MAF >= 5%
64
+
65
+
66
+
BS2
67
+
---
68
+
`Observation in controls inconsistent with disease penetrance`
69
+
70
+
Variants with a MAF < 5% (not BA1) which have a gnomAD non-founder population `hom` count > 5 when considered under an
71
+
autosomal dominant model (gene score) or > 2 when considered under autosomal recessive, X-recessive or X-dominant are
72
+
assigned a `BS2`.
73
+
74
+
75
+
PM2
76
+
---
77
+
`Absent in population databases`
78
+
79
+
In accordance with the `updated PM2 guidance <https://clinicalgenome.org/site/assets/files/5182/pm2_-_svi_recommendation_-_approved_sept2020.pdf>`_, variants absent from all of the user-defined :ref:`frequencysources`
80
+
will be assigned the `PM2_Supporting` criterion. Additionally, for variants considered under a recessive mode of inheritance they
81
+
can have a frequency of < 0.01% (0.0001) in all non-bottlenecked populations to be assigned `PM2_Supporting`.
82
+
83
+
84
+
PS4
85
+
---
86
+
`Prevalence in affecteds statistically increased over controls`
87
+
88
+
Not assigned
89
+
28
90
Computational and Predictive Data
29
91
=================================
30
92
PVS1
31
93
----
94
+
`Predicted null variant in a gene where LOF is a known mechanism of disease`
95
+
32
96
Variants must have a predicted loss of function effect, be in a gene with known disease associations and have a gene
33
97
constraint LOF O/E < 0.7635 (gnomAD 4.0) to suggest that a gene is LoF intolerant. Variants not predicted to lead to
34
98
NMD (those located in the last exon) will have the modifier downgraded to Strong.
35
99
36
100
PS1
37
101
---
102
+
`Same amino acid change as an established pathogenic variant`
103
+
38
104
Variants with the same amino acid change as previously reported P/LP missense or in-frame indel ClinVar variants will be
39
105
assigned `PS1` with a strength of `Strong` for variants >= 2 stars, `Moderate` for variants with 1 star or `Supporting`
40
106
for those without a ClinVar start rating.
41
107
108
+
Splice variants will be assigned PS1 with Strong, Moderate or Supporting modifiers,
109
+
according to table 2 of `Using the ACMG/AMP framework to capture evidence related to
110
+
predicted and observed impact on splicing: Recommendations from the ClinGen SVI Splicing Subgroup <https://doi.org/10.1016/j.ajhg.2023.06.002>`_.
111
+
42
112
PM4
43
113
---
114
+
`Protein length changing variant`
115
+
44
116
Stop-loss and in-frame insertions or deletions, not previously assigned a `PVS1` criterion are assigned `PM4`.
45
117
46
118
PM5
47
119
---
120
+
`Novel missense change at an amino acid residue where a different pathogenic missense change has been seen before`
121
+
48
122
Variants having a novel missense change to an amino acid where a previously reported ClinVar P/LP variant has been seen
49
123
will be assigned `PM5` with a strength of `Moderate` for those with >=2 stars or `Supporting` otherwise.
50
124
51
125
PP3 / BP4
52
126
---------
127
+
`Multiple lines of computational evidence support a deleterious effect on the gene/gene product (PP3)`
128
+
`Multiple lines of computational evidence suggest no impact on gene product (BP4)`
129
+
53
130
If REVEL is chosen as a pathogenicity predictor for missense variants, `PP3` and `BP4` are assigned using the modifiers
54
131
according to table 2 of `Evidence-based calibration of computational tools for missense variant pathogenicity classification
55
132
and ClinGen recommendations for clinical use of PP3/BP4 criteria <https://www.biorxiv.org/content/10.1101/2022.03.17.484479v1>`_.
56
133
Note that this suggests the use of modifiers up to Strong in the case of pathogenic or Very Strong in the case of benign predictions.
57
134
Otherwise, an ensemble-based approach will be used for other pathogenicity predictors as per the original 215 guidelines.
58
135
It should be noted we found better performance using the REVEL-based approach when testing against the 100K genomes data.
59
136
137
+
For splice variants outside of the +/-1,2 canonical splice donor/acceptor site are assigned `PP3` with a SpliceAI score
138
+
>= 0.2 or `BP4` with a SpliceAI score < 0.1
139
+
140
+
141
+
BP1
142
+
---
143
+
`Missense in gene where only truncating cause disease`
144
+
145
+
Missense or inframe indels found in a gene in which >=75% of ClinVar P/LP variants are loss of function variants are
146
+
assigned `BP1`.
147
+
148
+
BP7
149
+
---
150
+
`Silent variant with non predicted splice impact`
151
+
152
+
For synonymous variants, `BP7` is assigned if the variant is not reported as P/LP in ClinVar, with a 1+ star rating and
153
+
has a SpliceAI score < 0.1
154
+
155
+
For splice region variants Exomiser follows the recommendations made in `Using the ACMG/AMP framework to capture evidence related to
156
+
predicted and observed impact on splicing: Recommendations from the ClinGen SVI Splicing Subgroup <https://doi.org/10.1016/j.ajhg.2023.06.002>`_
157
+
158
+
In this case, non-donor/acceptor splice region variants with a SpliceAI score < 0.1 which have been assigned BP4 at or
159
+
beyond the +7 and −21 positions (conservative designation for donor/acceptor splice region) are also assigned a `BP7`.
160
+
161
+
60
162
Functional Data
61
163
===============
164
+
165
+
PS3
166
+
---
167
+
`Well-established functional studies show a deleterious effect`
168
+
169
+
Not assigned
170
+
171
+
62
172
PM1
63
173
---
174
+
`Mutational hot spot or well-studied functional domain without benign variation`
175
+
64
176
Missense and inframe indels are assigned `PM1` if the surrounding region of 25 nucleotides either side of the variant
65
177
contain at least 4 reported P/LP variants in ClinVar and no B/LB variants. If the number of P/LP variants is greater
66
178
than the number of VUS in the region the strength will be assigned `Moderate` but regions containing P/LP <= VUS
67
179
(and no B/BL) will have the strength downgraded to `Supporting`.
68
180
69
181
182
+
PP2
183
+
---
184
+
`Missense in gene with low rate of benign missense variants and path. missenses common`
185
+
186
+
Missense or inframe indels in genes where >= 75% of ClinVar P/LP variants are missense and ClinVar benign missense
187
+
variants make up <=5% of all reported missense ClinVar variants (P/LP/VUS/B/LB) in that gene are assigned a `PP2`
188
+
189
+
190
+
BS3
191
+
---
192
+
`Well-established functional studies show no deleterious effect`
193
+
194
+
Not assigned
195
+
196
+
70
197
Segregation Data
71
198
================
199
+
200
+
PP1
201
+
---
202
+
`Co-segregation with disease in multiple affected family members`
203
+
204
+
Not assigned
205
+
206
+
72
207
BS4
73
208
---
74
-
If a pedigree with two or more members, including the proband is provided, Exomiser will assign `BS4` for variants not
75
-
segregating with affected members of the family.
209
+
`Non-segregation with disease`
210
+
211
+
Not assigned. Previously this was assigned but the pedigree assignment of 'affected' status was not always reliable so
212
+
this is no longer applied automatically.
213
+
76
214
77
215
De novo Data
78
-
===========
216
+
============
79
217
80
218
PS2
81
219
---
220
+
`*De novo* (paternity & maternity confirmed)`
221
+
82
222
Exomiser assigns the `PS2` criterion for variants compatible with a dominant mode of inheritance, with a pedigree containing
83
223
at least two ancestors of the proband, none of whom are affected and none of whom share the same allele as the proband.
84
224
85
-
Population Data
86
-
===============
87
-
88
-
For the population data criteria, all frequencies are considered using the populations set by the user in the
89
-
:ref:`frequencysources`, apart from any bottle-necked populations not recommended for frequency filtering from gnomAD
90
-
according to their `filtering allele frequency <https://gnomad.broadinstitute.org/help/faf>`_ document. This excludes
91
-
the Ashkenazi Jewish (ASJ), European Finnish (FIN), Other (OTH), Amish (AMI) and Middle Eastern (MID) populations. In
92
-
addition the LOCAL frequency will also not be used.
93
-
94
-
BA1
225
+
PM6
95
226
---
96
-
Given Exomiser will filter out alleles with an allele frequency of >= 2.0%, this is unlikely to be seen. However, alleles
97
-
with a maximum frequency >= 5.0% in the frequency sources specified will be assigned the `BA1` criterion. Variants listed
98
-
as being excluded from this category by the ClinGen SVI working group `BA1 exclusion list <https://www.clinicalgenome.org/site/assets/files/3460/ba1_exception_list_07_30_2018.pdf>`_
99
-
will not be marked as `BA1`, assuming they survived variant filtering.
In accordance with the `updated PM2 guidance <https://clinicalgenome.org/site/assets/files/5182/pm2_-_svi_recommendation_-_approved_sept2020.pdf>`_, variants absent from all of the user-defined :ref:`frequencysources`
104
-
will be assigned the `PM2_Supporting` criterion. Additionally, for variants considered under a recessive mode of inheritance they
105
-
can have a frequency of < 0.01% (0.0001) in all non-bottlenecked populations to be assigned `PM2_Supporting`.
106
231
107
232
Allelic Data
108
233
============
109
234
PM3 / BP2
110
235
---------
236
+
`Observed in *trans* with a dominant variant`
237
+
`Observed in *cis* with a pathogenic variant`
238
+
239
+
For recessive disorders, detected in *trans* with a pathogenic variant *PM3*
240
+
111
241
If Exomiser is provided with a phased VCF and a variant is found to be *in-trans* with a ClinVar Pathogenic variant and
112
242
associated with a recessive disorder, the `PM3` criterion will be applied. However, in cases where variant is being
113
243
considered for a recessive disorder and is *in-cis* or a dominant disorder and *in-trans* with another pathogenic variant
@@ -118,14 +248,26 @@ Phenotype
118
248
=========
119
249
PP4
120
250
---
251
+
`Patient's phenotype or FH highly specific for gene`
252
+
121
253
Given Exomiser's focus on phenotype-driven variant prioritisation, variants in a gene associated with a disorder with a
122
254
phenotype match score > 0.6 to the patient's phenotype are assigned the `PP4` criterion at the Moderate, rather than
123
255
Supporting level.
124
256
257
+
258
+
BP5
259
+
---
260
+
`Found in case with an alternate cause`
261
+
262
+
Not assigned
263
+
125
264
Clinical
126
265
========
127
266
PP5 / BP6
128
-
--------
267
+
---------
268
+
`Reputable source w/out shared data = benign (BP6)`
269
+
`Reputable source = pathogenic (PP5)`
270
+
129
271
If a variant is previously reported as P/LP in ClinVar with a 1-start rating, it will be assigned `PP5`, those with >= 2
130
272
stars (multiple submitters, criteria provided, no conflicts / reviewed by expert panel / practice guideline) will be
131
273
assigned a Strong level. Conversely, if the variant is previously reported as B/LB it will be assigned `BP6` with the same
@@ -138,7 +280,8 @@ Transcript Selection
138
280
139
281
Transcripts will be selected using the most deleterious predicted variant effect from `Jannovar <https://doi.org/10.1002/humu.22531>`_
140
282
according to the `transcript-source` property set in the `application.properties`. We recommend using the Ensembl
141
-
transcript datasource as the Exomiser build uses the GENCODE basic set of transcripts. Future versions should use MANE transcripts.
283
+
transcript datasource as the Exomiser build uses the GENCODE basic transcripts for hg19/GRCh37 and MANE transcripts for
284
+
hg38/GRCh38.
142
285
143
286
ACMG assignments will be reported for a variant on a transcript consistent with a particular mode of inheritance in
144
287
conjunction with a disorder, the assigned criteria with any modifiers and the final classification e.g.
0 commit comments