TN-Grid

TN-Grid
TN-Grid logo
Project
Status	Active
Category	Bioinformatics, Genetics, Drug discovery
Compute	CPU
Development
Developer	Department of Information Engineering and Computer Science (DISI), University of Trento
Sponsor	University of Trento, National Research Council of Italy (CNR), BOINC.Italy
Maintainer	TN-Grid team
Initial release	January 1, 2014  (12 years ago)
Repository	https://gene.disi.unitn.it/test/
Software
Written in	C, C++
Operating system	Windows, Linux, macOS
Size	~10 MB
BOINC statistics
Stats as of	May 23, 2026  (0 years ago)
Performance	94.71 GigaFLOPS
Active users	63
Total users	3,513
Active hosts	165
Total hosts	69,613
Metadata
Website	https://gene.disi.unitn.it/test/
License	Mixed; based on BOINC infrastructure

BOINC based TN-Grid is a volunteer computing project focused on bioinformatics, computational biology, and gene network analysis. The project uses the BOINC distributed computing platform to harness unused processing power donated by volunteers around the world. TN-Grid is operated by the Department of Information Engineering and Computer Science (DISI) at the University of Trento in Italy in collaboration with several scientific institutions and research groups.

The project is best known for its gene@home sub-project, which studies gene regulatory networks and causal relationships between genes in order to improve biological understanding and support applications such as disease research and drug repositioning.

History

TN-Grid was launched during the 2010s as an academic volunteer computing initiative developed at the University of Trento. The project was designed to provide researchers with access to large-scale computational resources without requiring dedicated supercomputers. Instead, computations are distributed among thousands of volunteer computers connected through the Internet using BOINC.

The project gained visibility within the BOINC community because of its focus on bioinformatics and causal inference in genetics, areas that require substantial computational resources for statistical analysis and network reconstruction.

During the COVID-19 pandemic, TN-Grid received support from the AMD HPC Fund, which provided computing resources for scientific research initiatives.

Why TN-Grid?

Modern biology produces enormous quantities of genomic and transcriptomic data. Understanding how genes interact with one another is one of the major challenges in bioinformatics and systems biology. Many diseases, developmental processes, and responses to environmental stress are controlled not by single genes, but by large interacting networks of genes and proteins.

Constructing and analyzing these networks requires large-scale statistical computation. The computational complexity of many network inference algorithms grows rapidly with the number of genes involved. In simplified form, the number of possible interactions between genes may scale approximately as:

<math>\frac{n(n-1)}{2}</math>

where <math>n</math> represents the number of genes being analyzed.

For modern genomic datasets involving thousands of genes, the number of possible relationships becomes extremely large. TN-Grid distributes these calculations across volunteer computers, significantly reducing the time required to analyze complex biological systems.

Goal

The primary goal of TN-Grid is to support scientific research in bioinformatics through volunteer distributed computing. The project focuses on identifying causal relationships among genes, expanding gene regulatory networks, and improving computational methods for biological data analysis.

TN-Grid researchers aim to:

• discover new relationships between genes
• improve understanding of cellular regulatory systems
• assist drug repositioning and biomedical research
• analyze large biological datasets efficiently
• develop scalable algorithms for systems biology

The project also serves as an example of citizen science and volunteer computing in academic research, allowing members of the public to contribute directly to scientific discovery.

Volunteer computing

Gene expression matrix

TN-Grid operates using the BOINC middleware platform developed at the University of California, Berkeley. Volunteers install the BOINC client software, attach to the TN-Grid project, and receive computational work units from project servers.

The computations are generally CPU-based and run in the background while the volunteer's computer is idle. Results are returned to project servers for validation and scientific analysis.

Like many BOINC projects, TN-Grid awards participants credit points based on completed work units. These credits are used primarily for community statistics and competition among volunteers and teams.

Sub-projects

gene@home

gene@home is the principal scientific application of TN-Grid. The project studies gene regulatory networks (GRNs), which describe causal and regulatory relationships among genes inside living organisms.

Every living organism contains genes that encode the information necessary to produce proteins. Gene expression involves the transcription and translation of genetic information into functional biological molecules. Regulatory proteins and signaling pathways influence when genes are activated or suppressed.

Gene regulatory networks are often represented mathematically as graphs:

<math>G = (V,E)</math>

where:

• <math>V</math> represents genes
• <math>E</math> represents regulatory or causal relationships

The goal of gene@home is to expand known GRNs by identifying additional genes that may participate in regulatory interactions.

The project uses an algorithm called PC-IM, an iterative implementation derived from the PC algorithm used in causal inference and probabilistic graphical models.

The algorithm partitions candidate genes into blocks, merges them with existing regulatory networks, and evaluates possible causal relationships. Iterative refinement is used to improve prediction quality and reduce false positives.

Research has included studies on:

• human gene interaction networks
• grapevine gene regulation
• plant biology
• drug repositioning
• causal inference in genomic data

The project has collaborated with Fondazione Edmund Mach (FEM) and other Italian research institutions.

Scientific results

TN-Grid uses the BOINC volunteer computing platform.

gene@home

The PC-IM algorithm has been evaluated using both synthetic and real biological datasets, including expression data from Arabidopsis thaliana and grapevine regulatory networks.

Several experimental evaluations have been reported:

Preliminary evaluation

Researchers compared in silico generated datasets with in vivo biological data from public databases. While simulated data provided higher sensitivity under some conditions, real biological datasets were considered more reliable for practical analysis.

The project also compared the PC algorithm with ARACNE, another network inference algorithm. The PC algorithm demonstrated better performance on real expression datasets and improved positive predictive value (PPV).

PC-IM evaluation

Experiments analyzed several factors affecting PC-IM performance:

• block size optimization
• iteration count
• robustness against non-real GRNs
• comparison with competing methods such as GENIES

The best performance was obtained using approximately 1000 genes per block and around 100 iterations.

Biological validation

Researchers validated many predicted gene relationships through bibliographic analysis and comparison with known biological literature. Significant enrichment was observed relative to randomly selected genes, suggesting that the algorithm successfully identifies biologically meaningful relationships.

COVID-19 research

During the COVID-19 pandemic, TN-Grid participated in computational efforts related to biomedical research and received support through AMD's HPC Fund initiative.

The availability of volunteer computing resources allowed researchers to continue large-scale computational analysis during a period of increased global scientific collaboration.

Project team / Sponsors

TN-Grid is operated primarily by researchers affiliated with:

• National Research Council of Italy (CNR)
• University of Trento (UniTN)
• Department of Information Engineering and Computer Science (DISI)
• BOINC.Italy
• Fondazione Edmund Mach (FEM)

The project also received support from AMD through the COVID-19 AMD HPC Fund initiative.

Scientific publications

gene@home

1. Pilati, Stefania.(2021}).Vitis OneGenE: A Causality-Based Approach to Generate Gene Networks in Vitis vinifera Sheds Light on the Laccase and Dirigent Gene Families. Biomolecules. DOI: 10.3390/biom11121744.
2. Blanzieri, Enrico.(2021}).A Computing System for Discovering Causal Relationships Among Human Genes to Improve Drug Repositioning. IEEE Transactions on Emerging Topics in Computing. DOI: 10.1109/TETC.2020.3031024.
3. (2019})."OneGenE: Regulatory Gene Network Expansion via Distributed Volunteer Computing on BOINC".DOI: 10.1109/EMPDP.2019.8671629.

1. Malacarne, Giulia.(2018}).Discovering Causal Relationships in Grapevine Expression Data to Expand Gene Networks. Frontiers in Plant Science. DOI: 10.3389/fpls.2018.01385.
2. Asnicar, Francesco.(2018}).NES 2 RA: Network expansion by stratified variable subsetting and ranking aggregation. The International Journal of High Performance Computing Applications. DOI: 10.1177/1094342016662508.
3. (2015})."Discovering Candidates for Gene Network Expansion by Distributed Volunteer Computing".DOI: 10.1109/Trustcom.2015.640.

1. Erculiani, Luca.(2015}).Discovering candidates for gene network expansion by variable subsetting and ranking aggregation. F1000Research. DOI: 10.7490/F1000RESEARCH.1110311.1.
2. (2015})."TN-Grid and gene@home project: volunteer computing for bioinformatics".

See also

• BOINC
• Volunteer computing
• Distributed computing
• Bioinformatics
• Gene regulatory network
• Citizen science

External links

• Official TN-Grid website
• gene@home
• BOINC