Astrid Research is looking to differentiate itself in the market for next-generation sequence analysis tools by offering users a combined hardware/software solution that it hopes will appeal to users with limited IT resources.
The Debrecen, Hungary-based company recently released GenoMiner, a next-generation sequencing data-analysis system that includes proprietary algorithms as well as open source software designed to help users process and analyze their sequence data.
Among other features, GenoMiner includes an algorithm for performing reference assemblies using "color space" data generated on Applied Biosystems' SOLiD sequencer, as well as data generated on Roche and Illumina sequencers.
Other functionalities of the system include tools to preprocess raw short reads and assemble de novo sequences; and an analysis pipeline that lets users analyze ChIP-sequence data.
The company offers three versions of the solution: GenoMiner W4, priced at €20,000 ($27,853); GenoMiner W9, priced at €29,000 ($40,361); and GenoMiner S, priced at €100,000 ($139,270) or more.
W4 and W9, which are geared towards "standalone users," both operate on 2x2x4 Intel CPU cores and provide 48 and 96 gigabytes of RAM and 8 TB and 10 TB of disk capacity, respectively. GenoMiner S, which provides higher performance, operates on 8x6 CPU cores and can include up to 500 GB of RAM and 900 TB of disk.
Zoltan Kovacs, Astrid's chief executive officer, told BioInform that the company chose to develop a combined hardware/software system to target users with limited IT resources who may have difficulty selecting the necessary parameters needed to analyze their data in most software packages.
"What we did is create the framework and put the most useful free and proprietary software pieces [on it] and create pipelines for next-generation sequence reference assembly and so on," he said.
The Path to Bioinformatics
Born in 2004 as a spinoff from the University of Debrecen in Hungary, Astrid Research began its days developing tools for crystallography and medical informatics before setting its sights on the bioinformatics industry, Kovacs said.
During its crystallography days, the company licensed a hashing algorithm and a random-number-generating algorithm from its parent university and planned to implement both algorithms in data security software, but chose to abandon the project because "there was no demand" for it, he said.
Next the company turned its attention to medical informatics and began developing several tools for that market including an automated image processing and machine learning software aimed at helping clinicians speed up the process of screening eye images from large numbers of patients to identify those that are at risk for diabetic retinopathy, a diabetes-related eye disease. According to Kovacs, the tool is still in the development phase.
The shift towards bioinformatics began in early 2009, Kovacs said, when the company partnered with a Hungary-based SOLiD sequencing service provider and realized that there was a "huge need" for data analysis tools to handle the data generated on SOLiD sequencers. Furthermore, Astrid was also approached by Richter-Pharma AG, a Hungarian-based pharmaceutical company that was looking for a tool to analyze its microarray data.
Another problem that users identified, Kovacs said, has to do with a recurring issue in the bioinformatics field that's yet to be resolved, which is that a lot of open source software isn't maintained and supported long term.
In response to these needs, Kovacs said, Astrid's developers came up with GenoMiner.
GenoMiner is the company's first bioinformatics product, but it had previously offered analysis services for biobanking, oligonucleotide design, and gel electrophoresis. Kovacs noted that prior to the development of GenoMiner, the company offered sequence analysis services but decided that it was "time to sell the software instead of doing services."
At its core, GenoMiner incorporates several free open source tools including the European Bioinformatics Institute's Velvet for sequence assemblies, and the Short Read Mapping Package, or SHRiMP, an alignment algorithm developed by the University of Toronto's Michael Brudno and others.
The package also includes several of the company's in-house algorithms, as well as GenoViewer, a free visualization browser for the SAM/BAM format that lets users generate and export mutation tables and consensus sequences — two functionalities that the company claims no other browsers support. Among other features, users can also visualize read errors, SNPs, insertions, and deletions in both color and base space.
A key feature of GenoMiner, which Kovacs said distinguishes it from the offerings of companies like CLC Bio, is that it contains the company's in-house algorithm for handling short-read datasets, also called tags, from the SOLiD sequencer, which are color coded.
According to the company, where most reference assembly algorithms have difficulty working in color space, its algorithm "aligns the raw output tags to the color-coded transformation of the reference sequence" and in this way is able to detect polymorphisms, insertions, and deletions in the data.
Astrid is not the only provider offering combined hardware/software solutions for the next-gen sequencing market. CLC Bio — which Kovacs said is Astrid's main competitor — offers a range of hardware-based options for customers, while other firms like Biomatters have also bundled NGS analysis algorithms onto dedicated hardware offerings (BI 7/23/2010).
In the face of a highly competitive marketplace, Astrid's marketing strategy is to be "faster and more accurate as well as to offer more affordable tools," Kovacs said. He added that GenoMiner's ability to handle SOLiD data is also a key selling point.
Furthermore, Astrid's business model "is not to sell software but to try to sell the software with the hardware," he said. This way, researchers receive the software they need along with sufficient memory space to run experiments in a single package.
The next version of GenoMiner, which is scheduled for release early next year, will have additional functionalities including RNA-seq analysis, Blast, and multiple sequence alignment.
Currently the market for Astrid's bioinformatics tools includes researchers in academic laboratories, universities, and research institutions. While its current customers are all based in Hungary, the company is looking to move its products into the international life sciences market.
To that end, the company recently presented GenoMiner at the Genome Informatics conference hosted by the Wellcome Trust Sanger Institute in Cambridge, UK, and plans to present the platform at the Bio-Europe conference in Munich, Germany, in November.
Currently Astrid has 36 employees, including software engineers, mathematicians, molecular biologists, and medical doctors. In addition to revenue generated from the sales of GenoMiner and the provision of informatics services, the company is funded by grants from the Hungarian government and the European Union.