A former colleague of mine, Bibiana Garcia-Bailo, has graciously provided her notes on this conference. I post them here. Later, these will appear in edited form on Jim Kaput's NutriAlerts.
Notes From the MGP Workshop. September 26th, 2010. INCON/ICMAA 2010 – Brazil.
Welcome and Introduction
Lucia Regina Ribeiro gave the first welcome address to the workshop. She stressed that the work of the MGP will be important to Latin America particularly in the following areas:
1. 1. Multiple micronutrient deficiencies.
2. 2. Interactions between nutritional status and infectious diseases. For example, attempts to increase iron status conflict with the fact that excess iron can precipitate infections. Supplementation is beneficial in promoting resistance against infection, but pathogens also require micronutrients for growthà so we must critically evaluate supplementation in developing countries to make sure it does not affect pathogens and diseases such as malaria, pneumonia, tuberculosis or HIV
3. 3. Micronutrient supplementation
4. 4. Individual and ancestral genetic variability
A hearty congratulations and thanks must be extended to Lucia for organizing this event, where so many scientists and students from around the world have gathered to learn from one another, exchange ideas and establish connections.
After Lucia’s welcome address, John Hesketh introduced the idea and history of the MGP and stressed its nature as a community effort. He presented the topic of variation in micronutrient requirements among individuals, and how one size does not fit all. The current, population-derived recommendations for micronutrient intakes may not target the health needs of individuals, who have different requirements based on their genetic background, lifestyle, physiologic status, etc.
In light of this idea, the MGP was formed. The MGP had three drivers to it:
- · Human genome and human variome information
- · Omic data, transcriptomic,s proteomics, metabolomics
- · Bioinformatics
The idea was, then, to combine expertise in these areas and to move the science forward as a group. The MGP has been conceived as a community bioinformatic resorce and an online portal for access to a comprehensive database of micronutrient ‘omics’ information. It links to existing tools and databases, and has several components:
- · A genetic variation portal
- · Knowledge of all micronutrient-relevant variations
- · A micronutrient pathway portal with gene micronutrient interactions;
- · A database of array data, other omics data, phenotypes and study design information
The goal is to put together a single-source portal of information on micronutrient pathways, metabolism, genetics, omics and microarrays data. The MGP’s overall aims are the following:
- · Creation of a public portal and bioinformatics toolbox for all ‘omics’ information on micronutrients and related pathways
- · New generation of micronutrient research to improve understanding of individual requirements and health outcomes
- · Individualized strategies based on micronutrients to improve health
To this end, expert groups have been assembled for each micronutrient and are taking charge of developing gene lists and pathway information to begin populating the MGP portal. John stressed the importance of collaboration and extended an invitation to researchers to participate in these expert groups. He then presented preliminary work on Selenium as an exemplar of how the MGP is working,
As a conclusion, John emphasized the MGP as a way to gather the various ‘omics’ information available in a central point of access, through the MGP website. The MGP is a community bioinformatic resource that links to existing tools and databases. In terms of future developments, John mentioned the EU-fnded MICROGENNET project for global collaboration in this area.
Micronutrient Genomics Pathway Portal – Comparing the Needles in Two Haystacks
During this talk, Chris Evelo introduced some of the bioinformatic aspects behind the MGP work.
He stressed that the MGP expert micronutrient teams must produce gene lists for two reasons:
- · To be able to start epidemiological studies, and get companies to develop technology to sequence those genes cheaply and effectively
- · To start looking at nutritional phenotypes associated with those genes so that we can begin filling the genetic portal.
He then introduced WikiPathways, to which there is a link from the MGP website, as well as its new features.
Beyond building pathways by hand, in the traditional ways, one can also use the tool in an automated way through Webservice. Webservice is intended for use by software, rather than users. For example, Webservice can automatically integrate data from ArrayExpress Atlas with WikiPathways pathways.
Wikipathways now also has a Cytoscape plugin, so one can upload a pathway from Wikipathways into Cytoskape. Based on the connections formed, one can access a great amount of information, search and open pathways from Wiki into Cytoscape.
With another tool, SSOAP (http://www.omegahat.org/SSOAP/), one can upload the pathways from Wikipathways into R, allowing for the ability to do things like GSEA in R on Wikipathways using webservices.
Chris also presented an example of the applications of this technology through a gene list for vitamin D. The list is actually pathway data, and can be used to show information directly on the pathway. A number of sources were used to get the list, such as Wikipathways, KEGG Pathway and Gene Ontology. The list can be used immediately to load data onto it. One can connect to the network, and can for example connect SNP data to candidate gene studies. Clicking on any of the genes gives the user all the information, which is in part contained in the pathway and has enough information to connect to the online network in order to obtain even more information. One can also open the data in a pathway editing program, and also see the literature references where the genes came from. There is also the possibility of extensive curation of the list.
A few good questions were asked after Chris’s presentation. One was whether biochemical pathways are included in the Wiki. Chris responded that Wikipathways accepts any definition of a pathway that the user has. So ‘pathway’ is a combination of biochemical pathways, network interactions, transport processes, protein-protein interactions, and gene lists. One suggestion was made that the pathways be able to show transcriptomics information and different versions of pathways.
Overall, the point was stressed that MICROGENNET will be able to facilitate exchanges between individuals and groups from all the expert teams, to share methods and educate on approaches among the different groups.
New Pathways: The EURRECA Network
Suzan Wopereis introduced the EURRECA Network to the audience.
EURRECA is a European-funded FW6 program consisting of scientists, universities, small companies, etc. Suzan highlighted the fact that every country in Europe has its own micronutrient recommendations, and the aim of EURRECA is to develop tools to align micronutrient recommendations across Europe, with a special focus on vulnerable groups and consumer understanding. Suzan is involved in Integrating Activity 3, whose goal is to go from dietary recommendations for one micronutrient for the whole population towards recommendations for multiple micronutrients for one person.
Suzan presented her group’s work by first elaborating on the idea that nutrition-related health is a balance of three overarching processes: oxidative stress, inflammatory stress, and metabolic stress. If these processes are imbalanced, nutrition-related diseases occur. Micronutrients contribute towards the balance of the three processes. Suzan and her collaborators are creating micronutrient-related networks by searching the literature for micronutrient status parameters. They then collect information on micronutrient function parameters such as enzymes and cofactors, to get an idea of micronutrient function. They then collected health parameters.
All this information is to be available in the EURRECA website, with all different micronutrients and links to specific wikis. For each micronutrient, one can see a table with the three overarching stresses (oxidative, inflammatory, metabolic). The table contains biomarkers for these processes, and for each specific micronutrient it shows the relationship with established biomarkers as extracted from the literature, plus a score for the ‘trustworthiness’ of the relationship.
This sort of information allows users to begin creating micronutrient networks, observing the interactions between various micronutrient networks, and observing how the micronutrients relate to inflammation, oxidation and metabolic stress. Suzan mentioned that they have created a standard way of drawing pathways using systems biology graphical notation. These pathways can be edited by everyone, using PathVisio. She highlighted some of the work done on selenium, folic acid and vitamin B12 pathways. All of these are available on the Wikipathways micronutrient portal.
Suzan’s talk elicited some good discussion. The point was brought forward that, from the point of view of the MGP, adding genetic information to these networks is key. In turn, the pathways can help identify genes associated with the micronutrient of interest. One interesting addition in the future would be to include information on kinetic dynamics and substrate dependence. Another useful addition would be the ability to see the overlap in pathways. This can already be done in a rudimentary fashion, but will be developed further in the future.
Micronutrient Genetic Variation Portal
Jim Kaput’s talk on the Micronutrient Genetic Variation Portal highlighted the importance of looking at genotype – environment interactions. Jim started by exposing the fact that current, powerful database repositories such as in the NCBI website contain a wealth of genetic and molecular databases and tools, but they fail to include important information on nutrition and lifestyle. This omission is fatal, since nutrients constitute the most long-lasting environmental influence on our biology, from the uterine environment to the end of life, interacting with the genome and leading to different responses in individuals depending on their genetic background and other environmental factors. Therefore, we as nutrition scientists MUST include nutrition and lifestyle data in databases such as NCBI, and put our tools on these websites. Jim mentioned the PhenX Project as one example that aims to analyze phenotype and prioritizes 20 research domains.
In the US, a recent meeting resulted in a paper in Journal of Nutrition, 2010 (in press), where experts got together to assess resources and see what we need as a group to go forward to make nutritional and lifestyle information available together with current genetic and molecular databases and tools.
The Nutritional Phenotype Database, dbNP, is one major effort to create a toolbox for nutrigenomics research in order to link environmental with genetic information. Pathway maps are merged with genetic information, in order to build a genetic module that covers both. Jim showed a screenshot from the database. The genetic module is divided into various tabs or sections. The tabs so far included are pathways, variant information, gene, genetic, epigenetic, and epistatic. One can download the information from the database for statistical analysis, so it’s not just a visualizing tool. Epistasis is very important because of gene-gene interactions.
Jim then illustrated a potential use for the database. One could select candidate genes for analysis by looking into nutrient pathway maps, to then test whether they are involved in the complex phenotype of interest. In such a way, we combine nutrient pathway and genetic information to answer the question of how to bring this information into the clinic and health outcomes.
Jim also spent some time discussing how genetic variation in individuals makes it difficult to design case-control studies or interventions. In designing studies, we must take into account different genetic backgrounds , cultures and diet around the world, which may lead to different results in different regions. This idea was summarized in the paper ‘Planning the Human Variome Project: The Spain Report,’ Hum Mutat Res 2009. This is of particular importance for nutrigenomics projects. Along these lines, Jim brought up the shortcomings of GWAS research, since this technique does not take into account epistasis or epigenetics. One must take into account the entire genome of the individual and the whole set of interactions that may be going on between genes, many of which may not be represented in the platform used depending on the population under study.
Finally, Jim suggested future steps for dbNP. These would include reaching a consensus on the public data elements that should be available, as well as the allocation of tasks to specific groups. Funding of the work to make dbNP take off should also be considered.
The talk provided a note of resonance on the idea that we should be assessing variation in different populations across the globe, as well as considering that gene silencing occurs with aging, and this will become an emerging issue with aging populations.
Genomic Perspective on Vitamin D Signaling
Carsten Carlberg provided a summary of the current genomics research on vitamin D signalling. The case of vitamin D is interesting because it is both a micronutrient that can be obtained from the skin or the diet, and also, in its bioactive form, a transcription factor that activates genes in numerous pathways, such as inflammation, cell growth and metabolism, directly through binding to the vitamin D receptor (VDR). Therefore, the bioactive form of vitamin D can have an impact for various diseases and should be a major target for nutrigenomics research.
However, there are major challenges. For example, the whole genome consists of 22,000 genes. A few percent are targets of vitamin D. Furthermore, there are 250 different tissues, all with the same genome but a different, tissue-specific transcriptome. In addition, there is time-dependent regulation, with dynamics ranging from minutes to days.
Carsten and his group have been conducting research on 1,25(OH)D –the bioactive form of vitamin D- in human cell lines such as monocytes, non-malignant human breast and prostate. For example, they examined genome-wide VDR binding in THP-1 cells. They exposed the cells to 1,25(OH)D for just 40 minutes of treatment, in order to understand regulation of pro- and anti-inflammatory genes by VDR. In untreated cells, they found 1,406 sites occupied by VDR. In treated cells, VDR was on 2,700 binding sites. In addition, many of these sites were found far away from the transcription binding site. VDR was found to bind to the proximity of genes participating in RNA-related functions, vitamin D metabolism, the inflammatory response and insulin signalling. And this happened very quickly, after just 40 minutes of exposure.
After illustrating the importance of vitamin D as a key player in important physiologic processes through his research on cell lines, Carsten highlighted the complexity of studying genomics associated with this micronutrient. There are millions of SNPs, both functional and regulatory, in coding regions, synonymous, non-synonymous, coding and non-coding. Regulatory SNPs can have a variety of effects, ranging from slightly reducing to completely preventing transcription factor binding, so that there is no gene expression whatsoever. There are a number of common traits that show associations with VDR binding. These include type 1 diabetes, Crohn’s Disease, lupus, colorectal cancer, chronic lymphocytic leukemia, tanning, hair colour, height, rheumatoid arthritis, multiple sclerosis, etc.
Carsten suggested two different approaches to genomics research. One is biological hypothesis generation a priori, with genome-wide screening for SNPs in regulatory elements, followed by selection of promising candidate SNPs for population-based studies. Another is an a posteriori approach, with refinement or generation of biological hypotheses of SNPs that are significantly associated with specific outcomes.
Carsten’s talk highlighted the importance of integrative bioinformatics in the genomics research process. In this case, this includes linking one’s information on transcription sites, regulatory SNPs, etc. with curated annotations and data sets, followed by the synergistic administration of public and in-house data.
General Discussion
One of the main points driven forth from the workshop was that expert teams are already making gene lists on various micronutrients, and any individuals who are interested in contributing to any of the micronutrients should contact the team leaders, whose information can be found in the MGP website. A number of representatives from the Human Variome Project (HVP), such as Richard Cotton, were present at the MGP workshop and also put out a call for interested individuals to participate in that endeavour.
The shortcomings of NCBI and EBI were discussed. While both databases have a wealth of information, the data are not readily available and contain little dietary information. The MGP and HVP are collecting material and wrapping it around the existing databases, but there needs to be an even greater concerted effort to contribute to these projects in order to cover the deficiencies of the existing databases.
Also raised was the point that the concepts and the expert teams are already in place, but not much face-to-face discussion has happened yet. At the next workshop, the teams should be brought together in one room to start tracking progress. However, moving the teams forward and having successful workshops in the future will require funding, and it is critical to identify funding sources, since this has been a ‘weekend’ effort for a lot of people so far. Recently obtained funding, such as MICROGENNET, has begun to allow people to exchange ideas, visit one another and collaborate on small bits of projects, but more needs to come.
In regards to this, it was discussed that breaking tasks into small pathway parts might make it easier to assign them so that they can be carried out with less money. Teams could meet in subgroups, and every team needs to chip in to bring their own bit of funding. To this effect, involving students is key since they bring great energy and passion to the project without requiring large salaries.
Another suggestion came through for selling the MGP under a genomic medicine umbrella, since the EU is currently offering funding for this. At this point, a discussion began about the potential Latin American, and in particular Brazilian, contribution to the MGP. Lucia Regina Ribeiro elaborated on this, stating that Brazil can contribute in two ways. One is through vitamin D, where a team is already being coordinated in cooperation with Carsten Carlberg. The other is through a lab with a study of variation in genes that work to transform beta-carotene into vitamin A. The work to screen for these gene variants is being carried out not just in Brazil, but also elsewhere in Latin America.
A discussion was also had on the major roadblocks to the success of the MGP. Focus is one. The group should not lose sight of the main point, which is about determining requirements for an individual for each micronutrient. The project’s goal should be to define, using ‘omics’, what is deficiency and what is excess. The teams should identify key practical outputs and tangible outcomes with respect to how to move the research forward. This is not to say that micronutrient recommendations should be made at this state of knowledge. The objective at this point should be to increase knowledge and collect more data – genomic, physiologic and environmental. Along the process, ways to properly store and filter these data should be identified so that we are not buried in abundance without being able to make sense of the available information.