Blog takeover: The EMBL-EBI team working on HoloFood

    19.01.23 09:15 AM By HoloFood-KU

    This blog post we head over to HoloFood partners EMBL-EBI where we meet up the Microbiome Informatics team. As part of the HoloFood project their hands-on tasks come into force after the HoloFood experiments and laboratory work is complete, whereby the HoloFood data is now in computational format. 


    We take a look behind the curtain and talk to them about their role and ideas of the new holo-omic solutions!


    Meet the team

    The team is a combination of software developers, web developers and bioinformaticians forming the service branch of the Microbial Informatics team. We use complex mathematical models tailored for life-science research with an overall aim is to simplify access to curated, complex data and to assist researchers with their questions.


    Seen here the full team assembled: Sandy Rogers (far left), Varsha Kale, Rob Finn, Germana Baldi and Lorna Richardson after a full day of teaching at the HoloFood training course in Bilbao, Spain September 2022. 

    What does your team do?

    Varsha & Germana - We analyse biological data after it has been sequenced. It can now be processed on a high performance computing cluster with the power and memory required to piece together the complex environments we investigate. Computational resources are crucial for discovering and exploring the microbial world, and our main goal is to find the right balance between precision of our results and adaptability of the same approach to different types of data. We realise more and more that relying only on bacteria that can be cultivated within a lab setting (known as ‘cultures’) doesn’t always encompass the novelty in complex organisms and environments, and culturing fussy microbes can be difficult. 


    In a project like HoloFood we have the unique option to have multiple data layers to work with - not just the DNA of the microbes and host but also of all the other biological molecules that have also been sequenced such as the transcriptome (mRNA) that are very useful to get the full range of not only who is in the environment we are investigating but also what they are doing. 


    The team at EMBL-EBI receives HoloFood data after it is has been generated in the laboratory from samples from the field - in the case of HoloFood from the chicken and salmon trials designed to optimised food production. Once in this computational format they use tools to organise and analyse these data for them to be used for research.

    Lorna - In general we build and create workflows and “toolboxes” to assist in the analysis of the datasets we work on, using tools that are frequently used by the research community. Of course, each project requires not only collaboration between us but also with the other beneficiaries of the project and as is normal with these big data projects to require some translation between the various needs and expertise of each partner. 


    Another big part of our role is to create and curate a HoloFood database - a portal where all the levels of biological data, from measurements in the field, the genomes, microbiomes and more are all organised and accessible in an easy to use interface. This requires a lot of input not only from the scientists generating the data but also from people on our team (such as Sandy) who know how to create these websites that can link all this information in a logical fit.  In the end we strive that our work for these projects can live on beyond the timeframe of the projects - making these results accessible allowing for future exploitation of the results. 


    Sandy – As Lorna said, I spend most of my time developing web resources. The HoloFood database is interesting because it is designed to house as little data as possible – instead we want the projects’ data to be archived in public repositories. The portal fills in some of the filtering and interlinking between those datasets that is otherwise difficult to do. In that sense I see some of it as a prototype of a world where this kind of linking (between experimental trials, animal hosts, samples etc) is possible at that repository level. 


    What excites you about what you do?

    We all agree : The team is what makes the job the best. The different backgrounds of the team are great to help us get a much better idea on how to optimise our workflows and ensure that our end results can be usable and useful to the community!

    Varsha Kale at work where she is analysing the datasets generated from the HoloFood project

    Varsha & Germana -  From a research standpoint we are excited that there is an increased interest in the microbiome and appreciation that the function of the microbiome is important. We look forward to seeing how these large scale omic data projects, like HoloFood, open the door to more projects with a focus on building statistical models or machine learning algorithms to help us disentangle the level of data complexity we are encountering.  As tools are developed we look forward to when more overlooked members of the microbiome such as viruses and eukaryotes get some much needed attention as well. 


    There are still many challenges, where every unique dataset requires a unique set of skills to tackle the differences of each animal or plant biome but open and good collaboration with all the experts from the start of these projects pays off in the long run. Just the differences between salmon and chicken projects within HoloFood were quite surprising.

    Outline of the HoloFood data portal- HoloFood data follows the FAIR principles - a set of guidelines for making data Findable, Accessible, Interoperable and Reusable.
    Lorna at the front during a morning jog around Bilbao during the Applied HoloGenomics conference! Photo courtesy of Sandy Rogers

    Sandy – I’m excited that we are working to make the project’s data practically FAIR, not just theoretically. Imagining real users of our data archive (I think back to being a fresh-faced first year PhD student grappling with data archives) helps understand why the extra effort now makes the project’s legacy stronger.


    Lorna-  It’s great to see how HoloFood is setting the gold standard on how holo-omics data can be organised and stored to make it accessible and open in the long run. Pulling on all levels of information requires thorough planning from the start and the input of all data generators. Long-term usability of the HoloFood data, beyond the scope of the project, is very important to me, so having a data portal ready to go we are excited to see what other questions can be answered using these datasets!

    Stay tuned for the release of the HoloFood data portal later in this year!

    In the meantime take a look at these resources:
    ​MGnify resource: https://www.ebi.ac.uk/metagenomics/
    ​HoloFood data training materials:  https://holofood-course.readthedocs.io/en/latest/
    ​Follow us on Twitter: https://twitter.com/HoloFood_EU or sign up to our newsletter!

    HoloFood-KU