A Flexible and Open-Source Tool for Genetic Variant Annotation
Andrea Bombarda
1 a
, Matteo Bellini
2 b
, Maria Iascone
2 c
and Domenico Fabio Savo
1 d
1
Department of Management, Information, and Production Engineering, University of Bergamo, Bergamo, Italy
2
Medical Genetics Lab., ASST Papa Giovanni XXIII, Bergamo, Italy
Keywords:
Medical Genetics, Variant Annotation, Rare Genetic Disease Research.
Abstract:
Advances in genomic research have significantly enhanced our understanding of the genetic factors influencing
human health. A key output of this research are VCF (Variant Call Format) files, which document genetic
variations detected through DNA sequencing. These files, however, provide limited information, making it
challenging to interpret the biological significance of the variants without additional data. Annotation, the
process of enriching VCF files with information from publicly available biomedical datasets, is essential for
facilitating variant interpretation in research. In this paper, we present VCFAnnotator, a tool developed to
adapt ANNOVAR software used in genetic research, enabling the annotation of entire directories with a single
command and facilitating the use of any relevant external database. Additionally, VCFAnnotator offers the
ability to scrape the various websites of the biomedical databases in use, ensuring that the researchers remain
informed of any updates.
1 INTRODUCTION
Genomic research has transformed our understand-
ing of human biology, providing crucial insights into
the genetic factors that influence human health (My-
ers et al., 2001; Battista et al., 2011). The study of
genetic variants is now pivotal in medical research,
helping to identify new biological pathways, elucidate
disease mechanisms, and inform the development of
innovative therapeutic approaches. As the volume of
genomic data continues to grow, the demand for effi-
cient tools to process, analyze, and interpret this infor-
mation has become increasingly critical to advancing
research in medical genetics.
A crucial output of genetic analysis process are
the VCF (Variant Call Format) files. These files, pro-
duced through DNA sequencing and alignment, doc-
ument genetic variations in comparison to a reference
genome. Some variants may be associated with the
development of diseases, but not all are pathogenic.
It is up to the geneticist, based on current medical re-
search, to evaluate whether specific symptoms can be
linked to these genetic variants.
To assist the genetics researchers in their work, the
a
https://orcid.org/0000-0003-4244-9319
b
https://orcid.org/0009-0001-4297-9160
c
https://orcid.org/0000-0002-4707-212X
d
https://orcid.org/0000-0002-8391-8049
data contained in the VCF file are enriched with ad-
ditional information about each variant by using data
from publicly available biomedical datasets. This pro-
cess, called annotation, is considered crucial for inter-
preting genetic data and assessing the potential impact
of variants on human health (Salgado et al., 2016).
In fact, apart from detailing the chromosome po-
sition and the nucleotides in a specific variant, VCF
files provide very little information. As a result, with-
out performing annotation, it becomes challenging to
address key questions, such as whether a variant has
already been identified in other studies, what its ef-
fects might be, or whether it is widespread in the pop-
ulation and therefore unlikely to be pathogenic. In the
context of genetic research, where new correlations
between variants and diseases are continuously dis-
covered, leveraging up-to-date information is imper-
ative for increasing the chances of achieving prompt
and accurate results. For this reason, several tools that
help researchers to annotate VCF files have been pro-
posed (Wang et al., 2010; Yang and Wang, 2015; Ped-
ersen et al., 2016; Cingolani et al., 2012; McLaren
et al., 2010). However, despite their efficiency, most
of them lack in providing information to the user
about the up-to-date status of the used data sources
and are limited to annotations performed using only a
subset of the available biomedical databases.
In this paper, we present VCFAnnotator, a tool de-
406
Bombarda, A., Bellini, M., Iascone, M. and Savo, D. F.
A Flexible and Open-Source Tool for Genetic Variant Annotation.
DOI: 10.5220/0013111600003911
Paper published under CC license (CC BY-NC-ND 4.0)
In Proceedings of the 18th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2025) - Volume 2: HEALTHINF, pages 406-413
ISBN: 978-989-758-731-3; ISSN: 2184-4305
Proceedings Copyright © 2025 by SCITEPRESS – Science and Technology Publications, Lda.

veloped in collaboration with the medical genetics de-
partment of Papa Giovanni XXIII Hospital in Berg-
amo, to address the issues they observe in existing
annotation tools. Specifically, VCFAnnotator acts as
a wrapper for the ANNOVAR (Wang et al., 2010)
tool, enabling the automatic verification of the up-to-
date status of the data sources used to perform the an-
notation. Moreover, it allows for working with any
database that is not natively compatible with ANNO-
VAR by automatically converting selected resources
into a format accepted by the ANNOVAR annotation
tool. With these features, VCFAnnotator integrates
seamlessly with ANNOVAR, allowing researchers to
perform VCF annotation more quickly and efficiently,
both for new cases and for those previously analyzed
but considered inconclusive.
The rest of the paper is structured as follows.
Sect. 2 provides the background on the variant iden-
tification and annotation process, as well as the most
used biomedical databases, and presents the ANNO-
VAR , which is at the base of VCFAnnotator. Then,
Sect. 3 and 4, present the requirements and the design
of VCFAnnotator and its implementation details and
functionalities, respectively. Finally, Sect. 5 discusses
related works, and Sect. 6 concludes the paper.
2 BACKGROUND
In this section, we describe the process of analysis
and investigation that, from the biological material of
a subject, allows for the extraction of his/her genetic
variants, and the following annotation process. More-
over, we provide an overview of the software tools
currently used to support the annotation phase and
highlight their limitations.
2.1 Variant Identification
In Next Generation Sequencing (NGS), the journey
from biological material to a VCF file is a complex,
multistep process that integrates both wet-lab proce-
dures and computational bioinformatics. This pro-
cess aims to identify genetic variants, such as Sin-
gle Nucleotide Variants (SNVs), insertions, deletions,
Copy Number Variations (CNVs), and other muta-
tions, which are fundamental for genetic analysis and
therefore for formulating a diagnosis. In cases involv-
ing pediatric subjects, analyses typically involve data
from the child, mother, and father, sometimes includ-
ing other relatives.
Fig. 1 depicts a general overview of the process. In
the first step, biological material (e.g., blood, saliva,
or tissue) is collected, and DNA is extracted, frag-
Raw Biological
Material
Sequencing Reads
Mapping and
Alignment
BAMVariant CallingVCF File
Annotation Tool +
Variant Quality
Control
Annotated
VCF File
Clinically Relevant
Variants
Conclusive
Analysis
Inconclusive
Analysis
Figure 1: Overview of the Genetics Process.
mented, and tagged with adapters for sequencing. The
fragments are then processed through a sequencing
platform, producing raw data (short or long reads)
in FASTQ format with nucleotide sequences and qual-
ity scores. These reads are aligned to a reference
genome, generating a binary alignment map (BAM)
file with alignment details and quality metrics.
Then, variant calling is performed, and statisti-
cally significant differences (SNVs, insertions, dele-
tions, CNVs, and structural variants) between the
sample’s DNA and the reference genome are identi-
fied. The result is a VCF
1
file (Danecek et al., 2011),
listing all detected variants along with auxiliary in-
formation. An example of a VCF file is reported in
Listing 1. It is a tab-separated file containing chromo-
some, variant position, ID, reference and altered nu-
cleotides, quality score, filter info, read details (e.g.,
depth, genotype), field format, and case-specific data
for each individual.
To aid researchers, VCF files undergo further pro-
cessing, including variant quality control to remove
low-quality variants and false positives, and annota-
tion to enrich data with information about identified
variants. During annotation, information taken from
relevant biomedical databases (see Sect. 2.2) with the
potential functional effects, known associations with
diseases, or spread in the population for each variant
is added. This phase is carried out through the use of
annotation tools such as ANNOVAR (see Sect. 2.3).
As a result, a VCF file similar to the one we report in
Sect. 4 is produced. The annotation of VCF files is
essential because it transforms a simple list of genetic
variants into a tool rich in clinical and biological in-
formation, enabling their filtering and prioritization.
Given the importance of the annotation process
in research, several biomedical databases have been
released by the scientific community (see Sect. 2.2).
Moreover, several tools are available to perform this
activity. While most sequencing machine manufac-
1
The complete v4.3 specification of the VCF is avail-
able at https://samtools.github.io/hts-specs/VCFv4.3.pdf
A Flexible and Open-Source Tool for Genetic Variant Annotation
407

Listing 1: Exerpt of a VCF file.
1 #CHROM POS ID REF ALT QUAL FILTER INFO FORMAT W3HH−AFA−I
2 chr1 941119 . A G 568.23 PASS AC=2;AF=1;AN=2 ;DP=117;FS=0;MQ=250;QD=4.86 ;SOR=1.981 GT:AD: AF :DP :GQ: FT : F1R2 : F2R1 : PL
:GP: PP:DN 1 / 1 : 0 , 2 8 : 1 : 2 8 : 8 1 :PASS:0 ,1 3:0 , 1 5 : 169 , 8 4 ,0:1 3 0 . 8 2 ,8 0 . 816 , 0 : 329 , 1 6 4 , 0 : I n h e r i t e d [ . . . ]
turers offer proprietary analysis tools, these often re-
quire subscriptions and incur per-use fees. Conse-
quently, open-source and free alternatives, e.g., AN-
NOVAR (Sect. 2.3), have emerged as viable options.
2.2 Genetics Databases
As discussed in Sect. 2.1, the usefulness and ef-
fectiveness of the annotation process depend on the
databases used. Each of the available databases fo-
cuses on specific information, such as the frequency
of a variant in the population or the predicted effect of
the variant on the proteins. In the following, some of
the most common databases are introduced.
OMIM: The Online Mendelian Inheritance in Man
(OMIM) database (Hamosh, 2004) is a comprehen-
sive resource that catalogs human genes and genetic
disorders. It focuses on the relationships among genes
and the diseases they cause, as well as their inheri-
tance patterns. More specifically, it is used to indi-
viduate the inheritance patterns (dominant, recessive,
etc.) thanks to the link of each genetic variant to the
scientific literature, helping clinicians and researchers
better understand the genetic basis of diseases.
GnomAD: The Genome Aggregation Database
(GnomAD) (Chen et al., 2024) is a large public
dataset cataloging human genetic variants. It is
widely used in genomics research to help understand
the spectrum of genetic variants across diverse popu-
lations and their frequency. Given that most genetics
laboratories work on rare diseases, GnomAD is used
to exclude all variants with high occurrence rates in
the population, which are therefore not pathogenic.
Clinvar: ClinVar (Landrum et al., 2015) is a public,
freely accessible database archiving reports on human
genetic variants linked to diseases,along with support-
ing evidence. It enables easy access to information
about the relationships between genetic variants and
specific health conditions. ClinVar processes submis-
sions including variants identified in patient samples,
classifications related to diseases and drug responses,
and additional supporting data, and provides a classi-
fication based on the predicted biological effect of a
mutation (benign, pathogenic, etc.).
Gencode: The Gencode database (Pei et al., 2012)
is a collection of annotations of human and mouse
genomes. It provides detailed information on gene
structures, including protein-coding genes, noncoding
RNA genes, and other functional elements. It is used
for linking variants (both the position and the specific
nucleotide alteration) to specific genes and regions.
RefSeq: RefSeq (Reference Sequence) (O’Leary
et al., 2015) is a database curated by the National
Center for Biotechnology Information (NCBI) that
provides complete, annotated reference sequences for
genomes, transcripts, and proteins. It is used as a stan-
dard for genomic research and comparative studies,
offering accurate and non-redundant representations
of genetic sequences from humans and other species.
HGMD: The Human Gene Mutation Database
(HGMD) (Cooper, 1998) is a comprehensive resource
collecting data on clinically relevant variants that are
associated with genetic disorders. It can be used as
a key reference for researchers by providing detailed
information on variants that cause or may cause in-
herited diseases. The key advantage of this database
is that it contains entries that come only from scien-
tifically proven sources. Unlike the other previously
presented databases, HGMD is not freely available.
SIFT: SIFT (Sorting Intolerant From Tolerant) (Ng,
2003) is a database used to determine whether an
amino acid substitution in a protein will affect its
function. It analyzes sequence homology and the
physical properties of amino acids to assess whether a
mutation is likely to be deleterious (damaging) or tol-
erated (neutral). It is commonly used to evaluate the
potential impact of genetic variations.
2.3 ANNOVAR
ANNOVAR (Wang et al., 2010) is a tool that can be
used to annotate Single Nucleotide Variants (SNVs)
and insertions/deletions, such as examining their
functional consequences on genes, reporting func-
tional importance scores, or finding variants in con-
served regions. It is available as a CLI software tool
and can be used as a standalone application on sys-
tems in which standard Perl is supported. ANNO-
VAR works with text-based input VCF files (e.g., the
example reported in Listing 1), whereby each line cor-
responds to a genetic variant and reports its character-
istics. Then, to annotate variants, ANNOVAR needs
to download gene annotation databases, such as those
described in Sect. 2.2, and to save them to a local disk.
ANNOVAR is not the only tool available for an-
notating VCF files, and we describe most of the others
in Sect. 5. However, being open-source and supported
HEALTHINF 2025 - 18th International Conference on Health Informatics
408

by the community makes ANNOVAR one of the most
chosen options by researchers. Nevertheless, it still
has some limitations, which we try to address with
VCFAnnotator and discuss in the following section.
2.3.1 Limitations
Despite being very powerful, during our experiments,
we discovered several limitations of ANNOVAR that
our work aims to solve. First, the efficacy of the anno-
tation process is closely linked to the use of up-to-date
biomedical databases (see Sect. 2.2). Indeed, because
new correlations between gene variants and patholo-
gies are frequently discovered, the databases must be
in their latest available versions for the annotation to
be effective. However, ANNOVAR lacks a method to
automatically check for newer database versions.
Additionally, ANNOVAR requires the databases
used for annotating VCFs to be in a specific format,
which is not always the one used by those databases’
creators. To solve this issue, ANNOVAR provides a
set of already adapted databases,
2
but most of these
are not up-to-date and some required by the genet-
ics laboratories are not available (e.g., OMIM; see
Sect. 2.2). Thus, to get all needed and up-to-date in-
formation, working with “external” (and possibly in a
different format) databases is required. Another sim-
ilar limitation of ANNOVAR is that it can perform
the annotation by exploiting only databases with the
same format. This means that if one of the biomed-
ical databases used for the annotation is stored in a
txt file, all the other databases must be provided as
txt as well. In practice, every biomedical database
is provided in many possible formats and this limits
the usability of ANNOVAR as it is. Finally, ANNO-
VAR was designed as a tool that is to annotate only a
single VCF file per time, making it unsuitable for the
re-analysis process, in which or a set of previously
analyzed cases require reanalysis at the same time.
3 SOFTWARE DESIGN
In this section, we begin by introducing the re-
quirements we identified for VCFAnnotator that were
specifically tailored for overcoming the limitations
identified for ANNOVAR. Then, we present the ar-
chitecture we developed to allow the highest config-
urability and flexibility for the tool.
2
https://annovar.openbioinformatics.org/en/latest/
user-guide/download/
3.1 Software Requirements
At the beginning of the project, thanks to meet-
ings with genetics researchers we collaborated with,
we identified several requirements for VCFAnnota-
tor. More specifically, we identified some general re-
quirements and three modes.
In terms of general requirements, VCFAnnotator
should be as much configurable as possible because
of the evolving environment in the genetics labora-
tory and scientific state-of-the-art. This means that
the databases used for VCF annotation should be set
by the user, allowing for different operations in dif-
ferent formats. Similarly, the paths in which the anno-
tated VCF and biomedical databases are stored should
be configurable to make VCFAnnotator usable even
when additional storage is added and the file system
structure changes. In terms of modes, we identified
scraping, DB preparation, and annotation modes:
• When in scraping mode, VCFAnnotator should
be able to automatically check the websites of the
chosen biomedical databases, to discover whether
new versions are available. This operation should
be set as manual or automatic every time a new an-
notation is performed, but the update should always
be manual to avoid unwanted overwrites or having
databases in an inconsistent state, such as incomplete
downloads.
• When in DB preparation mode, the downloaded
databases should be adapted to make them suitable
for use with ANNOVAR. This phase is pivotal, as
biomedical databases are available in three different
formats (vcf, txt, and gff3) and adopt different en-
coding. More specifically, VCFAnnotator should be
able to remove all useless comments from txt files,
convert some vcf file into a txt file, and substitute the
<DEL> string, used when the variant implies a dele-
tion, with a dot. The user should be able to set for
each database, the required preparation operations.
• When in annotation mode, VCFAnnotator
should annotate the input VCF files using the infor-
mation taken from selected databases. More specifi-
cally, VCFAnnotator should support working at least
with the Gencode, Clinvar, GnomAD, OMIM and
HGMD databases and, in general, with databases in
the vcf, gff3, and txt format. Furthermore, VC-
FAnnotator should annotate a single file, a portion of
a VCF file, or all files in a specified folder to make
re-analyses possible on previously inconclusive cases.
At the end of annotation, VCFAnnotator should pro-
duce a single file with database information stored in
separate columns, enabling easier and up-to-date vari-
ant analysis.
A Flexible and Open-Source Tool for Genetic Variant Annotation
409

VCFAnnotator
«Annotation tool»
Annovar
«Component»
Input Parser
«Component»
WebScraper
«Component»
DBPreparation
«Database»
GnomAD
«Database»
HGMD
«Database»
Clinvar
«Database»
Gencode
«Database»
OMIM
Figure 2: VCFAnnotator software architecture.
VCFAnnotator
Functionalities
Scraping
Annotation
DatabaseAdaptation
DB
Preparation
Check Database
Update
Automatic
Scraping
Inputs
VCF Folder or
VCF File
Databases
Folder Path
(Optional)
Output Path
(Optional)
Output File for
Each Database
Annotation
Annotated VCF
File Merging and
Column
Formatting
Ready-to-Use
Databases
Figure 3: VCFAnnotator usage flow.
3.2 Software Architecture
Fig. 2 shows the VCFAnnotator software ar-
chitecture. The tool, implemented in Python
and available at https://github.com/ANTARES-PRJ/
VCFAnnotator, provides an Input Parser compo-
nent, which parses the input parameters and configu-
ration files. Depending on the configuration and the
user request, one of the three functionalities is started.
The annotation functionalities are carried out by
ANNOVAR (see Sect. 2.3). This takes as input the
VCF file (or files) to be annotated and, exploiting
the selected biomedical databases, performs the an-
notation. Note that ANNOVAR is executed as a
command-line tool. In this way, we kept the architec-
ture as modular as possible, and substituting it with
a different (and possibly more powerful or efficient)
annotation tool is possible in the future. In Fig. 2,
only five databases (i.e., those identified in the soft-
ware requirement analysis and previously explained
in Sect. 2.2) are reported, but the architecture is flexi-
ble, and new databases can be added on the fly.
The web scraping functionalities are carried out
by the WebScraper component, which takes advan-
tage of the functionalities offered by the beautifulsoup
Python library (Hajba, 2018). It is devoted to fetching
database websites and signaling whether new versions
for each of the set databases are available.
Finally, the database preparation is performed by
the DBPreparation component. As for the ANNO-
VAR tool, despite only five databases being connected
to the module in Fig. 2, the component can interact
with all databases, depending on user configuration.
4 APP PROTOTYPE
In this section, we present the usage of VCFAnnota-
tor, as shown in Fig. 3. VCFAnnotator supports three
different types of operations, with each fulfilling one
specific task - namely the automatic scraping of the
database websites, the preparation of the databases,
and the actual annotation.
Configuration. The properties of VCFAnnotator
and the configurations used in each of the three
supported functionalities are set by using the
config.yaml file reported in Listing 2. More specif-
ically, the file contains two relevant paths: The
db path, where the databases used during the anno-
tation phase are stored, and the destination path,
where the annotated files are saved (lines 1 and 2).
Then, the configuration information for each database
is described. As previously discussed, VCFAnnota-
tor supports databases in the txt, vcf, and gff3 for-
mats. Thus, according to the format, each database is
reported in a different list (databasesTXT at line 3,
databasesVCF at line 10, and databasesGFF3 at
line 17). For each entry, the configuration file con-
tains the id, file name, and the operation type to
be used while annotating a VCF file with the ANNO-
VAR subcomponent.
3
Automatic Scraping. The purpose of the automatic
scraping procedure is to discover, for each of the
databases used during the annotation phase, whether
new versions are available. This is pivotal for the ac-
curacy and effectiveness of the diagnosis provided by
a genetics laboratory. The scraping operation can be
performed automatically whenever a new annotation
is launched (autoCheck parameter at line 27 in List-
ing 2) or launched manually. To manually execute
this operation, VCFAnnotator can be called by using
the --checkDB (or -c) option:
$ python vcf annotator.py −−checkDB
In this way, the tool automatically checks for
database updates and prints a table in which updates
3
The following values can be used as operation: g for
gene-based, gx for gene-based with cross-reference annota-
tion, r for region-based, and f for filter-based.
HEALTHINF 2025 - 18th International Conference on Health Informatics
410

Listing 2: Configuration file for VCFAnnotator.
1 db path: "humandb/"
2 destination path: "result/"
3 databasesTXT: # Name of the DBs
4 - id: Clinvar
5 file: "clinvar"
6 operation: "f"
7 - id: OMIM
8 file: "omim"
9 operation: "f"
10 databasesVCF:
11 - id: gnomAD
12 file: "hg38_gnomad.vcf"
13 operation: "f"
14 - id: HGMD
15 file: "hg38_hgmd.vcf"
16 operation: "f"
17 databasesGFF3:
18 - id: Gencode
19 file: "hg38_gencode.gff3"
20 operation: "r"
21 convertFromVCFToTxt: # convert from vcf to txt
22 - id: Clinvar
23 removeDEL: # substitute <DEL> with the .
24 - id: HGMD
25 clean: # remove the comments from txt files
26 - id: OMIM
27 autoCheck: false # Scraping
28 scraping:
29 - id: Gencode
30 release: "44"
31 website: "https://www.gencodegenes.org/human"
32 textToSearch: "Release "
33 tag: "h1"
34 - id: Clinvar
35 date: "2024-04-22"
36 website: "https://ftp.ncbi.nlm.nih.gov/pub/clinvar
37 /vcf_GRCh38/"
38 textToSearch: "clinvar_"
39 tag: "a"
40 - id: GnomAD
41 release: "4.1"
42 website: "https://gnomad.broadinstitute.org/news/
43 category/release/"
44 textToSearch: "gnomAD"
45 tag: "h2"
46 - id: OMIM
47 date: "2024-09-20"
48 website: "https://omim.org"
49 textToSearch: "Updated "
50 tag: "h5"
are indicated with a [!], as illustrated by the screen-
shot in Fig. 4. Note that at the moment, the update
of the databases is not automatic, and the user has to
download the new files and manually substitute the
old versions. More specifically, after having down-
loaded the new files and substituted them in the cho-
sen db path (see line 1 in Listing 2), the user may
update the information in the config.yaml file (see
Listing 2, starting from line 28) by adding for each
database id the relevant version information (e.g.,
release number or release date), the website that
has to be scraped to find whether new versions are
available, and the text to be searched in a specific
HTML tag. The update process was made manual to
ensure researchers know the current database version
and avoid annotations with inconsistently updated
databases, which risk incomplete information. Fur-
thermore, we emphasize that in the screen in Fig. 4,
Figure 4: Output table for the automatic scraping procedure.
only four of the five databases used by VCFAnnotator
are reported. Indeed, looking at the scraping section
in Listing 2, the HGMD database is not set as one of
those to be scraped because, as explained in Sect. 2.2,
it is the only one not freely available.
Database Preparation. As introduced in Sect. 3,
VCFAnnotator allows for annotating VCF files by us-
ing different databases. This operation is made pos-
sible by its integration with ANNOVAR. However,
we have found that the formatting conventions used
by ANNOVAR and those used by major databases
may differ and, thus, a preparation of each database
is needed. To execute this operation, VCFAnnotator
can be called by using the --prepare (or -p) option:
$ python vcf annotator.py −−prepare
This operation encompasses three different steps
and is performed depending on the configuration
set in the config.yaml file reported in Listing 2
(from line 21 to line 26). The databases in the
convertFromVCFToTxt list are translated into the
.txt format, which is supported by ANNOVAR. De-
spite ANNOVAR is supposed to work correctly with
vcf files, we have found that for some of them, it does
not. For example, in our experiments, we have seen
that the Clinvar database is not supported by ANNO-
VAR if used as a vcf file, while the GnomAD one
works fine. Thus, for those databases ANNOVAR
does not work with, they are converted into txt files.
Then, if some <DEL>s are present, indicating a dele-
tion in the databases in the removeDEL list, they are
converted into a dot. Finally, all lines starting with #
are removed from the databases in the clean list as
they may be interpreted as the header by ANNOVAR.
VCF Annotation. This mode provides the core
functionalities of VCFAnnotator- i.e., the annotation
of VCF files. To execute this task, VCFAnnotator can
be called using the --annotateVCF (or -a) option:
$ python vcf annotator.py −−annotateVCF input.vcf
This operation requires the user to specify an
input vcf file or the path of a folder containing
multiple files. In both cases, for each input file,
A Flexible and Open-Source Tool for Genetic Variant Annotation
411

Listing 3: Example of an annotated VCF file.
1 Chr S t a r t End Ref A l t CLNALLELEID CLNDN CLNDISDB CLNREVSTAT CLNSIG I d Qual F i l t e r I n f o Format W3HH−AFA−I Gencode
MIM Number Gene/ Locus And Other Rel at ed Symbols Gene Name Approved Gene Symbol Entr ez Gene ID Ensembl Gene ID
Comments Phenotypes Mouse Gene Symbol / ID AC AN AF CLASS MUT GENE STRAND DNA PROT DB PHEN RANKSCORE SVTYPE END
SVLEN
2 chr1 941119 941119 A G . . . . . . 568.23 PASS AC=2 ;AF=1;AN=2 ;DP=117;FS=0;MQ=250;QD=4.86;SOR=1.981 GT: AD: AF: DP:GQ:
FT : F1R2 : F2R1 : PL :GP: PP:DN 1 / 1 : 0 , 2 8 : 1 : 2 8 : 8 1 :PASS: 0 ,1 3 :0 , 1 5 : 1 69 , 8 4 ,0:13 0 . 8 2 ,8 0 . 8 16 , 0 : 3 29 , 1 6 4 ,0: I n h e r i t e d Name=
ENSG00000187634. 13 ,ENST00000618323 . 5 , ENST00000474461 . 1 , ENST00000478729 . 1 , ENST00000618181 . 5 , ENST00000618779 . 5 ,
ENST00000622503 . 5 , ENST00000342066 . 8 , ENST00000616125 . 5 , ENST00000341065 . 8 , ENST00000616016 . 5 , ENST00000455979 . 1 ,
exon : ENST00000474461 . 1 : 1 , ENST00000617307 . 5 . . . . . . . . . . . . . . . . . . . . . . . . [ . . . ]
the annotation is performed using all databases
reported in the config.yaml file and, for each of
them, with the specific operation. We emphasize
that we designed VCFAnnotator to work even with
noncomplete vcf files, and with files composed
by merging rows appertaining to multiple subjects.
In such a way, researchers can also annotate a
reduced set of variants, e.g., those under investiga-
tion for finding a specific pathology. To maintain
the traceability of all operations, VCFAnnotator
calls ANNOVAR iteratively on all databases, so
that, at the end of the annotation process, multiple
annotated vcf files are available with the name
DBName VCFInputName YYYY-mm-dd HH MM SS.
In addition, a merged file with the name
VCFInputName YYYY-mm-dd HH MM SS is pro-
duced. It contains all annotations from all databases,
each one in one or more columns with the same name
(or starting with a prefix) as the used database.
During annotation, the databases are taken from
the db path folder (line 1 in Listing 2), and the
results are stored in the destination path folder
(line 2 in Listing 2). However, different paths can
be specified when executing VCFAnnotator from the
command line. More specifically, the database path
can be specified in the command after the option
--DBPath (or -db), while destination path after the
option --DestinationPath (or -d). An example of
an annotated VCF obtained using VCFAnnotator and
with the VCF previously shown in Listing 1 is re-
ported in Listing 3. It can be seen that the annotation
process adds several columns (e.g., all those starting
with CLN for the Clinvar database and the one marked
as Gencode for the gencode database).
5 RELATED WORK
Several attempts to provide efficient and effective
tools are available in the literature. One of the most
used is ANNOVAR (Wang et al., 2010), which is
the tool underlying the annotation functionalities of
VCFAnnotator. Over the years, it has been continu-
ously updated and extended, including the addition
of a web-based annotation environment, wANNO-
VAR (Yang and Wang, 2015). Similarly, in (Ped-
ersen et al., 2016), the Vcfanno tool was proposed.
While powerful, its complexity makes it less suitable
for non-expert users. Future work could explore au-
tomating its configuration as our architecture (Sect. 3)
supports replacing the annotation component without
affecting others. The SnpEff and VEP tools were
proposed in (Cingolani et al., 2012) and (McLaren
et al., 2010), respectively. They annotate variants
by genomic location and predict coding effects but
lack additional needed information, making SnpEff or
VEP unsuitable replacements for ANNOVAR. Simi-
larly, BCFTools (Danecek et al., 2021) can perform
VCF annotation, but it works only by using a sin-
gle reference database, and only in the GFF3 for-
mat. Instead, VCFAnnotator allows users to use more
databases with multiple input formats. Finally, most
companies producing sequencers provide their own
annotation tools, such as for Illumina.
4
However, in
most cases, the tools are not open-source and are ex-
pensive and inflexible. Other tools, such as VCF-
Miner (Hart et al., 2015), was proposed in the liter-
ature, but they allow only for mining information into
already annotated VCFs.
In addition to what mentioned in Sect. 2.2, other
tools can enhance the annotated VCF and help pri-
oritize variants. GERP (Genomic Evolutionary Rate
Profiling) (Davydov et al., 2010) detects conserved
regions by analyzing evolutionary constraints across
species. CADD (Combined Annotation-Dependent
Depletion) (Kircher et al., 2014) predicts variant im-
pact by combining annotations into a score, identify-
ing those likely to affect gene function. PolyPhen-
2 (Polymorphism Phenotyping v2) (Adzhubei et al.,
2013) assesses the impact of amino acid substitutions
on protein structure and function, aiding in evaluating
genetic variant pathogenicity.
6 CONCLUSIONS
Genetic variant annotation plays a crucial role in
genetics-related diseases research. This process en-
ables genetics laboratory researchers to filter variants
4
https://emea.illumina.com
HEALTHINF 2025 - 18th International Conference on Health Informatics
412

and assess their potential impact on human beings.
However, the analysis is not always definitive due to
the evolving nature of scientific knowledge, making
the use of up-to-date databases critically important.
Various tools have been introduced in the litera-
ture to annotate VCF files, with ANNOVAR being
one of the most widely used. Despite its popular-
ity, ANNOVAR has certain limitations regarding the
range of databases it supports. In this paper, we
propose VCFAnnotator to address these limitations.
It facilitates the use of external databases, automati-
cally preparing them for compatibility with ANNO-
VAR, and includes a scraping feature to detect up-
dated databases.
This study marks an initial step towards devel-
oping an automated system for streamlining vari-
ous tasks in genetic research processes. As future
work, we aim to enable researchers to selectively re-
annotate specific entries of the VCF file based on pre-
defined criteria and to facilitate comparisons between
these new annotations and previous ones. Addition-
ally, we aim to explore steps required for VCFAnnota-
tor ’s certification for diagnostic use (Bombarda et al.,
2022; Bombarda et al., 2021).
ACKNOWLEDGEMENTS
This work was funded by PNRR - ANTHEM (Ad-
vaNced Technologies for Human-centrEd Medicine)
- Grant PNC0000003 – CUP: B53C22006700001 -
Spoke 1 - Pilot 1.4. We would like to thank Fabio As-
solari and Simone Ronzoni for the preliminary work
that they did for this project during their B.Sc. theses.
REFERENCES
Adzhubei, I. et al. (2013). Predicting functional effect of
human missense mutations using polyphen-2. Current
Protocols in Human Genetics, 76(1).
Battista, R., Blancquaert, I., et al. (2011). Genetics in health
care: an overview of current and emerging models.
Public health genomics, 15(1):34–45.
Bombarda, A., Bonfanti, S., et al. (2021). Lessons learned
from the development of a mechanical ventilator for
covid-19. In 2021 IEEE 32nd International Sym-
posium on Software Reliability Engineering (ISSRE),
page 24–35. IEEE.
Bombarda, A., Bonfanti, S., et al. (2022). Guidelines for the
development of a critical software under emergency.
Information and Software Technology, 152:107061.
Chen, S., Francioli, L. C., Goodrich, J. K., et al. (2024). A
genomic mutational constraint map using variation in
76,156 human genomes. Nature, 625(7993):92–100.
Cingolani, P., Platts, A., et al. (2012). A program for
annotating and predicting the effects of single nu-
cleotide polymorphisms, snpeff: Snps in the genome
of drosophila melanogaster strain w1118; iso-2; iso-3.
Fly, 6(2):80–92.
Cooper, D. (1998). The human gene mutation database.
Nucleic Acids Research, 26(1):285–287.
Danecek, P., Auton, A., et al. (2011). The variant call format
and vcftools. Bioinformatics, 27(15):2156–2158.
Danecek, P., Bonfield, J. K., et al. (2021). Twelve years of
samtools and bcftools. GigaScience, 10(2).
Davydov, E. V., Goode, D. L., Sirota, M., et al. (2010).
Identifying a high fraction of the human genome to be
under selective constraint using gerp++. PLoS Com-
putational Biology, 6(12):e1001025.
Hajba, G. L. (2018). Using Beautiful Soup, page 41–96.
Apress.
Hamosh, A. (2004). Online mendelian inheritance in man
(omim), a knowledgebase of human genes and genetic
disorders. Nucleic Acids Research, 33(Database is-
sue):D514–D517.
Hart, S. N., Duffy, P., et al. (2015). Vcf-miner: Gui-
based application for mining variants and annota-
tions stored in vcf files. Briefings in Bioinformatics,
17(2):346–351.
Kircher, M. et al. (2014). A general framework for estimat-
ing the relative pathogenicity of human genetic vari-
ants. Nature Genetics, 46(3):310–315.
Landrum, M. J. et al. (2015). Clinvar: public archive of
interpretations of clinically relevant variants. Nucleic
Acids Research, 44(D1):D862–D868.
McLaren, W. et al. (2010). Deriving the consequences of
genomic variants with the ensembl API and SNP ef-
fect predictor. Bioinformatics, 26(16):2069–2070.
Myers, C., Paulk, N., and Dudlak, C. (2001). Genomics:
implications for health systems. Frontiers of Health
Services Management, 17(3):3–16.
Ng, P. C. (2003). Sift: predicting amino acid changes
that affect protein function. Nucleic Acids Research,
31(13):3812–3814.
O’Leary, N. A., Wright, M. W., et al. (2015). Reference
sequence (refseq) database at ncbi: current status, tax-
onomic expansion, and functional annotation. Nucleic
Acids Research, 44(D1):D733–D745.
Pedersen, B. S., Layer, R. M., and Quinlan, A. R. (2016).
Vcfanno: fast, flexible annotation of genetic variants.
Genome Biology, 17(1).
Pei, B., Sisu, C., Frankish, A., et al. (2012). The gencode
pseudogene resource. Genome Biology, 13(9):R51.
Salgado, D., Bellgard, M. I., et al. (2016). How to identify
pathogenic mutations among all those variations: vari-
ant annotation and filtration in the genome sequencing
era. Human mutation, 37(12):1272–1282.
Wang, K., Li, M., et al. (2010). ANNOVAR: functional
annotation of genetic variants from high-throughput
sequencing data. Nucleic Acids Res., 38(16):e164.
Yang, H. and Wang, K. (2015). Genomic variant annotation
and prioritization with annovar and wannovar. Nature
Protocols, 10(10):1556–1566.
A Flexible and Open-Source Tool for Genetic Variant Annotation
413