bioRxiv preprint doi: [Link] this version posted September 24, 2021.

The copyright holder for this preprint

(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under aCC-BY-ND 4.0 International license.

piRNAdb: A piwi-interacting RNA database

Ricardo Piuco1,2 and Pedro A. F. Galante2,*

1
Centro de Oncologia Molecular, Hospital Sirio-Libanes, São Paulo, SP 01308-060,
Brazil
2
Programa Interunidades de Pós-Graduação em Bioinformática

*To whom correspondence should be addressed: pgalante@[Link]

Abstract

Motivation: Found in several metazoan species, piwi-interacting RNAs (piRNAs)

regulate the expression of a wide variety of transposable elements and genes
associated with cellular development, differentiation, and growth. Despite their
importance, piRNAs are not well known and are still underexplored. To facilitate
piRNA research, a comprehensive and easy-to-use piRNA database is still needed.
Results: Here, we present piRNAdb, an integrative and user-friendly database
designed to encompass several aspects of piRNAs. We selected piRNAs from four
reliable human RNA-sequencing datasets to start our database. After data
processing, we displayed these sequences, their genomic location, clustering
information, putative targets on known genes and transposable elements, as well as
direct links to other databases. In this first release, piRNAdb catalogues 27,329
piRNAs, as well as 23,380 genes that are putative targets and 47,060 associated gene
ontology terms, both of which are organized and linked to each respective piRNA.
Finally, to improve information exchange and increase the confidence of sequences,
a feedback system is provided to users of piRNAdb. Conclusion: The inclusion of
new features to facilitate piRNA analyses, data visualization, and integration is the
major pillar of piRNAdb. Our main goal was to make this database an easy interface
between the data and the user. We believe that this web tool achieves this objective
by providing a streamlined and well-organized data repository for piRNAs and that it
will be extremely useful to those already studying piRNAs and to the broader
community. Availability: piRNAdb is available freely and is compatible with
smartphones and tablets: [Link]
bioRxiv preprint doi: [Link] this version posted September 24, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under aCC-BY-ND 4.0 International license.

Introduction
Piwi-interacting RNA (piRNA) is a class of small non-coding RNAs that range
from 26 to 31 nucleotides (1). piRNAs are involved in the regulation of several
biological processes, including i) control of transposable element (TE)
expression (2); ii) renewal and differentiation of stem cells (3); iii) animal
development (4); and iv) invasion, differentiation, and growth of breast
tumors, gastric tumors, and hepatocellular carcinomas (5).
piRNAs are found in almost all metazoans with surprisingly diversity,
even for closely related species (6). Most piRNAs originate from large genomic
regions, named clusters (7), that have a high density of piRNA sequences.
These clusters are known ‘traps’ for TEs; after a truncated form of this element
is inserted into a piRNA cluster, piRNA transcription can repress the parental
element. This event increases the repertory and diversity of piRNAs that
control the expression of cell molecules (7). The bound piRNA-PIWI protein
acts inside the cell nucleus, recruiting methylation or breaking the
non-processed mRNAs of their targets (8). As an increasing number of diverse
piRNAs are identified, a comprehensive and user-friendly database is needed.
Although some piRNA databases are available, none of them provide a
complete, updated, integrated, and user-friendly set of information about
piRNAs. For example, piRNABank (9) stores only data from NCBI and does not
include additional information relevant to these short RNAs. piRBase (10) has
the greatest number of piRNA sequences; however, it also presents suspicious
piRNAs that lack alignment to the human genome.
Here, we aimed to develop a comprehensive database of piRNAs. Our
resulting piwi-interacting RNA database (piRNAdb) groups all known piRNA
sequences in humans and includes information about alignments to the
genome, putative targets of piRNAs, clusters of piRNAs, gene ontology
classification for target genes, and hotlinks to other databases. Importantly,

1
bioRxiv preprint doi: [Link] this version posted September 24, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under aCC-BY-ND 4.0 International license.

our freely available web-based tool uses modern programming patterns and
architectures, allowing easy, intuitive, and responsive browsing with all
devices, including mobiles.

Data Implementation
We created and implemented a MySQL ([Link]) relational database
to store all of the data presented in the piRNAdb web interface. We developed
the back-end interface using the Model-View-Controller architecture in PHP
language ([Link] due to its ideal data separation and system
developer group workability. We developed our front-end interface using
HyperText Markup Language (HTML) 5, Asynchronous JavaScript, and
Cascading Style Sheets 3.
To assign specific pages and information sections more relevance, we
incorporated search engine optimization techniques and semantic HTML in
order to improve the ranking on search engines. In an effort to make data
retrieval faster and more accessible, we developed piRNAdb following
responsive rules, which are specific instructions given to different devices to
display the website content in correct proportions. As a result, our database is
functional on the majority of modern browsers, tablets, and smartphones
(mobiles).

Data Sources
Our complete database includes all piRNA sequences already reported in
humans according to datasets from (1,11–13). We built a PHP script to process
these piRNA data and upload the information to piRNAdb (see next section
for details). Of the 34,322 piRNA sequences, 1,496 were found to be duplicated
in two or more datasets; therefore, we merged these sequences.
Nevertheless, we kept references to all original data sources, even for merged
piRNAs. We downloaded the human genome (versions: hg19 and hg38) and

2
bioRxiv preprint doi: [Link] this version posted September 24, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under aCC-BY-ND 4.0 International license.

transcriptome (RELEASE 84) from the University of California, Santa Cruz

(UCSC) genome browser (http:/[Link]) and ENSEMBL
([Link] respectively.

Mapping piRNAs
For gaining information about piRNA genome location, we first aligned the
total set of 32,826 (34,322 – 1,496) piRNAs against the human genome (hg19
and hg38). We used the Burrows-Wheeler Aligner (parameters: -n 0 -t 8) (14),
which does not allow mismatches or gaps, and kept only full-length
sequence alignments against the genome. Using this approach, we identified
approximately 1,600,000 alignment positions (789,637 on hg19 and 794,576 on
hg38).
Human chromosomes 15 (37,938 alignments), 17 (37,174 alignments) and
19 (34,183 alignments) had approximately the same amount of alignments as
chromosome 4 (38,812 alignments), despite the fact that they (15, 17, and 19)
are shorter in length than chromosome 4. Curiously, chromosomes 19 and 17
had the highest density of coding genes in our genome (ENSEMBL RELEASE
87). Likewise, chromosome 15 had the highest density of small nucleolar RNAs
among all human chromosomes (ENSEMBL RELEASE 67).
Next, we developed an in-house pipeline to process and upload the
alignments to piRNAdb. We found that 26,879 (82%) of all piRNAs perfectly
matched to the human genome. In total, 437 piRNAs (1.34%) presented one or
two mismatches within polymorphic sites (based on NCBI’s dbSNP, version
150). The remaining 5,510 piRNAs presented one or two mismatches at other
genomic regions (hg19 or hg38). We only kept piRNAs with perfect matches
against the human genome in piRNAdb. However, the full list of piRNAs can
be found and downloaded directly on the piRNAdb website.

3
bioRxiv preprint doi: [Link] this version posted September 24, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under aCC-BY-ND 4.0 International license.

Clusters
We next determined the piRNAs clusters using a similar approach as (15). In
brief, we searched for 20kb genomic windows (with a 1kb step) that contained
at least seven piRNA alignments. We grouped piRNAs matching this cut-off
into clusters. In this step, we used only those piRNAs with unique alignment
to the genome. In total, we detected 217 clusters using this methodology. The
number of clusters found did not correlate with the chromosome size; for
example, chromosome 15 had more clusters (28) than chromosome 1 (10),
despite being greater in length and having a larger number of piRNAs.
Chromosome 17 and 19, which have the highest number of piRNAs, showed a
similar number of clusters as other chromosomes.

Targets
The last step of our data processing pipeline was the detection of piRNA
targets. We searched for piRNA targets within the lists of all known TEs and
coding and non-coding genes. To be considered a target, we required the
genomic coordinates to overlap between a piRNA and at least one transcript
from a gene. We allowed overlapping on both exonic and intronic regions, as
piRNAs are known to bind non-processed mRNA (16).
We found a total of 262,458 alignments from 18,799 unique piRNAs in
which the genomic coordinates overlapped with transcripts; 14,801 were
located on protein-coding regions and 6,163 on non-coding regions. This data
suggests that, similar to microRNAs (miRNAs), a piRNA may have several
genes as targets and a gene may be targeted by several piRNAs
(Supplementary Figure 1 and Supplementary Figure 2). For example,
DNM1P47 is a potential target for 251 piRNAs, and piRNA hsa-piR-20792 has
targets in 26,533 genes. Interesting, among the hsa-piR-13284 targets, we

4
bioRxiv preprint doi: [Link] this version posted September 24, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under aCC-BY-ND 4.0 International license.

found genes already described as being expressed in few tissues or being

broadly expressed. For example, CNTNAP2 is known to be expressed during
nervous system development (Poot, 2015) and EIF4G3 is a validated miRNA
target that is involved in mRNA cap recognition and transport to ribosomes
(17). Therefore, our results show that piRNAs have targets not only in TEs (see
below), but also within coding and non-coding genes, including those widely
expressed and those with known functions in specific organs/tissues.
We applied a similar methodology to find piRNA targets within TEs. We
found 228,869 piRNAs alignments (representing 3,680 unique piRNAs)
overlapping with TEs. Interesting, 1,076 (4%) piRNAs had targets in both
protein-coding genes and TEs elements, whereas other piRNAs (1,400; 5.20%)
bound only to TEs. Coding/non-coding gene and TE targets for piRNAs are
displayed in the same section (piRNA Alignments) on the piRNAdb web tool.
We next aimed to identify ontology terms related to piRNA targets.
Approximately 47,000 gene ontology terms (18) were available to the piRNA
targets. The top three most recurrent biological process terms were (1)
transcription, (2) regulation of transcription, and (3) signal transduction (1,366,
992, and 820 genes annotated, respectively). The three most frequent
molecular function terms were (1) protein binding, (2) metal ion binding, and
(3) ATP binding (7,135, 1,858, and 1,292 genes annotated, respectively). These
findings are shown in Supplementary Figure 3. To the best of our knowledge,
piRNAdb is the first piRNA database to display gene ontology information.
We believe that this information will be advantageous to users in performing
networking and integrative analyses using piRNAs’ targets.

Database Query and Interface

Currently, the field lacks a standard nomenclature for piRNAs. In order to
contribute to a standardization, we named piRNAs following a similar pattern
used for miRNA nomenclature. We assigned the piRNAs a unique code

5
bioRxiv preprint doi: [Link] this version posted September 24, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under aCC-BY-ND 4.0 International license.

formed by the initial related to the organism (hsa, for human), the word "piR"
(meaning piRNA), and a numerical code (e.g., resulting in hsa-piR-1 or
hsa-piR-5). The identified clusters also received similar codes formed by the
initial organism, the word "clu", and a numerical code (e.g., hsa-clu-7 and
hsa-clu-17).
We also developed a tool to link piRNAs named in other databases to
our system and vice versa. In the “Cross Codes” section, the user is able to
place a piRNA named in another database; our system then finds and
displays the access code in piRNAdb. Beyond the piRNA name, retrieving
information can be performed using the chromosome position, cluster name,
ENSEMBL code ID, or gene name of a desired target. For example, if the user
searches for a piRNA target (Figure 1a), a list of piRNAs (if they exist) will be
retrieved (Figure 1b). By clicking on a piRNA, a webpage displays further
information about the selected piRNA, such as its nucleotide sequence
(Figure 1c), alignments to the genome, and information about genes and
transposable elements that may be associated with each alignment (Figure
1d). In order to better integrate our data with other databases, we included
hotlinks on piRNAdb. For instance, clicking on one of piRNAs alignments will
redirect the user to the UCSC genome browser page, showing the respective
piRNA genomic coordinates and other information.
Also unique to this database, piRNAdb provides an interactive system
where users may give feedback about the reliability of a piRNA (Figure 1c). At
least for now, we will analyze and curate all feedback. It is also possible for the
user to be directed to Wikipedia ([Link]) to describe the piRNA
there as well. These features will enable faster communication among
researchers and enthusiasts of discoveries related to piRNAs.
Aiming for an easier visualization of the targets related to piRNAs for
the user and search engines, we developed a section below the alignments to
show all the targets for a specific piRNA in the tag cloud format. This feature

6
bioRxiv preprint doi: [Link] this version posted September 24, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under aCC-BY-ND 4.0 International license.

shows the most recurrent genes in a larger font size and stronger color to
highlight them to the users. Associated gene ontology terms to these gene
targets are displayed below (Figure 1e); information about which genes have
each term and the frequency of occurrence is available as well.
At the end of the page, the "Datasets" section displays information about
the methodology and tissues studied for the piRNA characterization. Finally, a
list of all the references linked to each specific piRNA is shown. All raw data
displayed in our web interface can be downloaded ("Downloads" page) in
various formats. A frequently asked questions and help section (FAQ/Help
page) provides information to the users about using the web interface, the
data version, and potential browsing problems. As this is the first piRNAdb
version (RELEASE 1), it only provides information about humans. We
anticipate that in the next release, piRNAdb will include piRNAs and their
related information for an additional organism as well.

Conclusion
Our system provides a powerful and user-friendly web tool for studying
piRNAs. piRNAdb uses modern programming architecture and tools, as well
as techniques to ensure scalability, usability, and better visualization for users.
Our web tool displays only relevant information about piRNAs, within a
streamlined and well-organized interface that is compatible with almost all
devices, including mobiles. We also provide information that is currently
lacking in other piRNA databases, such as gene ontology terms. In addition,
as we performed a strict selection of the piRNAs stored on our database, our
system is more robust and reliable for use by researchers.
In conclusion, piRNAdb provides reliable information about piRNAs and
several downstream analyses to help facilitate piRNA studies. Our findings on
the amount of alignments and putative gene targets associated with gene
ontologies provide unique information about our piRNAs. In the future, strong

7
bioRxiv preprint doi: [Link] this version posted September 24, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under aCC-BY-ND 4.0 International license.

candidates could be further explored or identified by utilizing the piRNAdb

system.

Figure 1. Schematic Search process and information accession about a

specific piRNA (has-piR-4346). a) Search form providing some options. Circle on
top right displays the search by official gene name to find piRNAs that possibly
target it. b) piRNA list that fulfills the term on the search step. c) Information
page for a specific piRNA with feedback system on the bottom right circle. d)
Section listing all the alignments to the genome (version hg38), putative gene
and transposable elements associated with the hsa-piRNA-4346. e) Gene
Ontology terms associated with the gene that are (putative) targets of
hsa-piR-4346.

8
bioRxiv preprint doi: [Link] this version posted September 24, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under aCC-BY-ND 4.0 International license.

Acknowledgements

We thank members of the Galante lab for suggestions and fruitful discussions.
R.P. is supported by a Coordenação de Aperfeiçoamento de Pessoal de Nível
Superior fellowship. This work was in part funded by the Fundação de Amparo à
Pesquisa do Estado de São Paulo (2012/24731-1 to P.A.F.G.) and CNPq (to P.A.F.G.).

Conflict of Interest
None declared.

REFERENCES

1. Aravin, A., Gaidatzis, D., Pfeffer, S., et al. (2006) A novel class of small RNAs bind to
MILI protein in mouse testes. Nature, 442, 203–207.
2. Watanabe, T. and Lin, H. (2014) Posttranscriptional regulation of gene expression by
Piwi proteins and piRNAs. Mol. Cell, 56, 18–27.
3. Gangaraju, V. K. and Lin, H. (2009) MicroRNAs: key regulators of stem cells. Nat. Rev.
Mol. Cell Biol., 10, 116–125.
4. Stefani, G. and Slack, F. J. (2008) Small non-coding RNAs in animal development.
Nat. Rev. Mol. Cell Biol., 9, 219–230.
5. Tan, Y., Liu, L., Liao, M., et al. (2015) Emerging roles for PIWI proteins in cancer. Acta
Biochim. Biophys. Sin. , 47, 315–324.
6. Siomi, M. C., Sato, K., Pezic, D., et al. (2011) PIWI-interacting small RNAs: the vanguard
of genome defence. PIWI-interacting small RNAs: the vanguard of genome defence.
Nature Reviews Molecular Cell Biology (2011) , 12, 246–258.
7. Yamanaka, S., Siomi, M. C. and Siomi, H. (2014) piRNA clusters and open chromatin
structure. Mob. DNA, 5, 22.
8. Liu, J. J., Shen, R., Chen, L., et al. (2010) Piwil2 is expressed in various stages of breast
cancers and has the potential to be used as a novel biomarker. Int. J. Clin. Exp.
Pathol., 3, 328–337.
9. Sai Lakshmi, S. and Agrawal, S. (2008) piRNABank: a web resource on classified and
clustered Piwi-interacting RNAs. Nucleic Acids Res., 36, D173–7.
10. Zhang, P., Si, X., Skogerbø, G., et al. (2014) piRBase: a web resource assisting piRNA
functional study. Database , 2014, bau110.
11. Girard, A., Sachidanandam, R., Hannon, G. J., et al. (2006) A germline-specific class of
small RNAs binds mammalian Piwi proteins. Nature, 442, 199–202.
12. Krawetz, S. A., Kruger, A., Lalancette, C., et al. (2011) A survey of small RNAs in human
sperm. Hum. Reprod., 26, 3401–3412.
13. Li, Y., Wang, H.-Y., Wan, F.-C., et al. (2012) Deep sequencing analysis of small
non-coding RNAs reveals the diversity of microRNAs and piRNAs in the human
epididymis. Gene, 497, 330–335.
14. Li, H. and Durbin, R. (2009) Fast and accurate short read alignment with
Burrows-Wheeler transform. Bioinformatics, 25, 1754–1760.
15. Lau, N. C., Seto, A. G., Kim, J., et al. (2006) Characterization of the piRNA complex from

9
bioRxiv preprint doi: [Link] this version posted September 24, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under aCC-BY-ND 4.0 International license.

rat testes. Science, 313, 363–367.

16. Zhong, F., Zhou, N., Wu, K., et al. (2015) A SnoRNA-derived piRNA interacts with
human interleukin-4 pre-mRNA and induces its decay in nuclear exosomes. Nucleic
Acids Res., 43, 10474–10491.
17. Mazan-Mamczarz, K., Zhao, X. F., Dai, B., et al. (2014) Down-regulation of eIF4GII by
miR-520c-3p represses diffuse large B cell lymphoma development. PLoS Genet., 10,
e1004105.
18. Ashburner, M., Ball, C. A., Blake, J. A., et al. (2000) Gene ontology: tool for the
unification of biology. The Gene Ontology Consortium. Nat. Genet., 25, 25–29.

10