NEW YORK (GenomeWeb) – Following months of intensive development and testing with several early-access users, bioinformatics software provider Genialis is gearing up for the public launch of its cloud-based solution for analyzing and visualizing next-generation sequencing data sometime in 2018.
Genialis, which officially opened its doors in 2013 in Ljubljana, Slovenia, recently raised $2.3 million in a seed round that it is using to finance commercialization efforts, as well as to expand the US-based operations that it started in 2016, according to company CEO Nejc Škoberne. He said in an interview that the company now has four employees in its Boston office and plans to add at least two more employees to its roster in the next several weeks including a business development manager as well as sales staff. This team will handle the company's commercial operations in the US moving forward.
Commercialization efforts will focus on the most recent iteration of the Genialis platform which the company has been developing for the last two years. "We anticipate doing a public launch early next year," Škoberne said. He noted that the platform is already being used by researchers in a few universities and research institutes in the US. With an expanded US-based team, Genialis should be able tap into the US market more easily and "in a more systemized way," something which has been a challenge since most of the commercial team has hitherto been based in Europe, he said.
Recently, Genialis rolled out several software-as-a-service products based on it core bioinformatics solution. The company is marketing what it describes as a modular platform that modernizes how research groups store and interact with their data.
Specifically, it provides a comprehensive collection of visual analytics and bioinformatics tools and third-party integrations that let users curate, manage, and visualize their data. This includes the company's open source Resolwe data engine which offers a forum for the company to share analysis pipelines with collaborators and crowdsource bioinformatics expertise from the community for future development efforts.
Once they are fully developed on the backend, those pipelines are then exposed to biologists in a simplified manner so that they can be run without additional configuration. Once their data is processed, biologists can interact with the information via user-friendly interfaces and interactive visualization tools such as heat maps. "Basically, you need no training as a biologist to start using it. That's the main goal," Škoberne said.
Much of the complexity underlying the Genialis platform is hidden from the end users, according to its developers. Underneath the hood, the solution leverages powerful algorithms that learn how users prefer to interact with their data and work to optimize their analysis activities and experience.
"Our company is a spinout from a data science lab in Slovenia which is very strong in artificial intelligence and machine learning," Škoberne said, and they continue to leverage this background to make improvements to the solution. For example, "right now we are implementing some very advanced algorithms that will advance the user experience of the workflow even further," he said. Specifically, "we are moving from only presenting the data … to [in a sense] interpreting the data for the end users at least to some extent," he explained. The benefit to users will be "that they get hints on what they should be looking at," which should help simplify their analysis further.
The most recent tools that the company has begun offering on its platform include the Expressions suite for RNA-seq data analysis and exploration, as well as a variant discovery application for analyzing targeted genome sequencing data which was developed in collaboration with Swift Biosciences.
"The main difference [between our platform] and other similar platforms is that we separate bioinformatician role and biologist role quite clearly," Škoberne said. "We don't anticipate that biologists will be able to tinker with workflow parameters or various bioinformatics concepts that they typically are not familiar with."
The list of early-access users of these and other bespoke applications that the company has developed for individual lab use include researchers from Harvard University, California Pacific Medical Center Research Institute, Johns Hopkins University, and the Jackson Laboratory.
Genialis also provides an enterprise iteration of its full platform for customers that require combined workflows and additional data management and dissemination tools. One such customer is Baylor College of Medicine with whom the company signed a licensing agreement last month. At the time of the announcement, Baylor noted that six labs at the school were actively using the platform, including contributing data, pipelines, and providing feedback to the company. The goal is to have a dozen or more labs at the school using the platform by the end of the year with the aim of opening enrollment to other labs shortly thereafter, the school said.
"The feedback we've got typically is that it is really easy for a biologist without a bioinformatics background to use the workflow [and] the visualizations to get from raw data to publishable figures," Škoberne said. At the same time, it is not a black box so "a biologist [can] look at how the specific result was computed … and see all the technical details that are required" if that’s of interest to them.
At Baylor, researchers in the laboratory of Gad Shaulsky, a professor of molecular and human genetics, are using one of Genialis' bespoke application to explore gene expression data from the soil amoeba Dictyostelium discoideum. The lab studies both gene expression and social behavior in the soil amoeba. Genialis President Rafael Rosengarten worked as a postdoctoral student in Shaulsky's lab prior to joining Genialis where he explored patterns of "leaps and lulls" in the development of Dictyostelium discoideum, Shaulsky explained in an interview.
What this platform does is allow pure biologists to "have access to our data in a way that allows us to explore the data and communicate it with both scientists who are true biologists and scientists who are computational and every shade of grey in between," Shaulsky said. For example, he and Rosengarten used the platform to generate a multidimensional picture that showed gene expression levels at various time points in Dictyostelium discoideum development.
"Dictys has something like 12,000 genes, and so for each one of them we can assign a number that is a reflection of how abundant the messenger RNA was at a particular time" using the Genialis infrastructure, he explained. This makes it possible to, for example, compare different conditions at different time points during the development of the protozoa and explore how they relate to each other.
Furthermore, users can share their data with collaborators via the Genialis platform and allow them to interact with the data for themselves, Shaulsky said. "They can click on a point … and explore the genes that are responsible for the difference between two different time points or two different conditions, and they can use tools like gene ontology enrichment or graphs that show heatmaps of gene expression or things like that."
For more complex analyses, biologists will still need the help of bioinformaticians and statisticians, Shaulsky noted. However, the Genialis platform provides a solid foundation from which to have those more complex conversations. "One of the limitations for me to collaborate with a statistician or data miner is that it may sometimes take a week or so for an idea to form in my lab [and] be transported to the person who can analyze the data and then back to me in a way that I can understand," he said. "Now it is at the tip of my fingers [and] I can do it immediately."
Another benefit of the solution is the ease with which Shaulsky's lab can process data. "Before Genialis came on board, it was fairly difficult for us to close the loop from when we produced the data and submitted it to the sequencing facility [and they] gave us back files," he said. "Unless you know what to do with these files, you've got nothing. What Genialis has generated is a pipeline where as soon as the sequencing service tells me that my files are deposited, I push a button and it allows me to annotate my experiments and everything else is taken care of."
Besides convenience, the Genialis platform is structured to support reproducible research. "Genialis opens up the black box and tells you exactly what was done to the data from A to Z and anybody can repeat it," Shaulsky said.
For Jernej Ule, a group leader at The Francis Crick Institute and a professor of molecular neuroscience at the University College London's Institute of Neurology, working with Genialis has been very positive. "We have set a timeline made of three stages, which have been achieved in time and recently successfully completed," he said in an email. "The interaction with the developers has been quick and professional."
Ule's group studies the assembly and function of ribonucleoprotein complexes (RNP) with the focus on neuronal biology and neurodegenerative diseases. They developed the individual-nucleotide resolution UV crosslinking and immunoprecipitation, (iCLIP), method for identifying in vivo protein-RNA interactions in a transcriptome-wide manner. His team has used iCLIP and related methods to address the mechanisms controlling alternative splicing and to characterize the function of several RNPs that are implicated in motor neuron disease.
"Most members of my group lack sophisticated bioinformatics expertise, which is required to analyze iCLIP data," Ule said. "[We needed] a solution for more efficient data processing and organization. Genialis platform offered such a solution and provided a toolbox for more intuitive and collaborative data analysis."
One of the tools Ule's lab uses is called iMaps which is used for analyzing high-resolution sequencing data. "The web platform enables [my group] to independently organize, analyze, and visualize their own data, and make discoveries without any additional help from bioinformaticians," he said.
Furthermore, Genialis offers well-documented software, which improves research reproducibility, according to Ule. In one experiment, the results obtained using the software "reproduce findings that we previously published using custom-made scripts," he said. "Moreover, the new software allows use of an updated genomic annotation, and improves the peak finding algorithm, which increases the robustness of findings."
Moving forward, Genialis plans to incorporate additional RNA-seq functionality in its platform next year. Specifically, "we are going to support transcript level expression," Škoberne said. "That's going to be quite an important feature a lot of people are asking for." The company also plans to include functionality for analyzing single cell RNA-sequencing data at some point in the future, he added.
Along with the commercial launch slated for next year, Genialis also intends to roll out a new per-sample price point instead of offering a purely licensing model, Škoberne said. He declined to disclose how much the company will charge per sample, but he did say that it would be both affordable and competitive compared to existing offerings.
Presently, Genialis supports some on-premise installations in addition to the cloud-based version of the software, Škoberne said but it hopes to onboard most of its users on the cloud.