In my previous article, Crash Landing into Zulip, I shared my experience joining the BugSigDB community and finding my feet in an established open-source project.
Now that you've joined the community and familiarized yourself with the project, it's time to understand what BugSigDB actually does and, more importantly, what you'll be contributing to.
Before attempting your first curation, there are a few concepts you need to understand. Let's start with the project itself.
What Exactly Is BugSigDB?
BugSigDB is an open-source, community-editable database of microbial signatures reported in scientific literature.
But what does that mean?
Microbiome research investigates communities of microorganisms, including bacteria, fungi, viruses, and other microbes, that inhabit different environments, including the human body.
Researchers conduct studies to understand how these microbial communities differ across conditions.
For example:
- Do people with a particular disease have a different gut microbiome from healthy individuals?
- Does the vaginal microbiome change across different reproductive stages?
- Does a particular treatment influence the abundance of certain microorganisms?
- Are there differences in microbial communities between people with different dietary habits?
These studies generate findings about microorganisms that are more or less abundant under particular conditions.
However, these findings are scattered across thousands of scientific publications, often reported using different methods, terminology, and formats.
This is where BugSigDB comes in.
BugSigDB brings these findings together in a structured database. It standardizes information about the studies, their experimental comparisons, and the microbial signatures reported in them.
This makes it possible to search, compare, and analyze findings across different microbiome studies.
In other words, BugSigDB helps turn findings scattered across scientific publications into structured, reusable research data.
First, What Is a Microbial Signature?
A microbial signature is essentially a collection of microbial taxa that share a common property or response to a particular study condition.
Let's use a hypothetical example.
Imagine researchers conduct a study comparing the gut microbiome of two groups:
- Group 0: Healthy individuals.
- Group 1: Individuals diagnosed with a particular disease.
After analyzing their samples, the researchers identify three bacterial genera that are more abundant in the disease group and two that are less abundant.
These findings represent microbial signatures associated with the comparison.
Increased in the disease group:
- Bacteroides
- Prevotella
- Escherichia
Decreased in the disease group:
- Faecalibacterium
- Roseburia
These are hypothetical findings, but they illustrate the kind of information BugSigDB captures.
Notice that a signature is not necessarily one microorganism. It can contain several taxa reported in the same direction for a particular comparison.
Also, a taxon is a classification unit used to identify microorganisms. Depending on the level of identification in a study, it could be a phylum, class, order, family, genus, species, or strain.
One important thing to remember: BugSigDB records what the researchers reported. As a curator, your job is to accurately extract and standardize those findings, not to independently decide which microorganisms should be associated with a disease.
Understanding How BugSigDB Organizes Research
This is probably one of the most important concepts to understand before starting your first curation.
BugSigDB organizes information into four interconnected levels:
Study → Experiment → Signature → Taxa
Let's break them down.
1. Study
A study represents a published microbiome differential-abundance research report.
It contains general information about the scientific paper, including its title, authors, publication details, and study design.
For example, if researchers publish a paper investigating changes in the gut microbiome among individuals with HIV, that paper represents one study in BugSigDB.
2. Experiment
An experiment represents a comparison between two or more groups for which differential-abundance analysis was performed.
A single scientific paper can contain several experiments because researchers may investigate multiple comparisons.
For example, a paper might investigate:
- HIV-positive individuals versus HIV-negative individuals.
- Male participants versus female participants.
- Participants receiving treatment versus those not receiving treatment.
Each distinct comparison may be represented as a separate experiment in BugSigDB.
This distinction is important. Do not assume that one scientific paper automatically corresponds to one experiment.
3. Signature
A signature represents the collection of microbial taxa reported as having increased or decreased abundance in a particular comparison.
An experiment can have one or two signatures, depending on the findings.
For example:
- One signature for taxa increased in Group 1.
- Another signature for taxa decreased in Group 1.
4. Taxa
These are the individual microorganisms or taxonomic groups reported in a signature.
BugSigDB uses the NCBI Taxonomy to standardize microbial names and identifiers.
This is important because scientific papers may report microorganisms using different naming conventions, abbreviations, or taxonomic classifications.
The goal is to make sure that the same organism can be consistently identified across different studies.
Why Does BugSigDB Need a Curation Policy?
Now that you understand how information is organized, you might wonder why there are so many guidelines to follow.
Why can't we simply read a scientific paper, identify the microorganisms, and enter them into the database?
The problem is ✨consistency✨.
Different researchers may use different study designs, sequencing methods, statistical tests, terminology, and reporting formats.
Even when two studies investigate the same condition, the information may be presented differently.
Without a common set of guidelines, two curators could interpret the same scientific paper differently and enter inconsistent information.
The BugSigDB Curation Policy provides a standardized approach to extracting and representing this information.
It helps ensure that curators follow consistent rules when identifying studies, defining experiments, recording metadata, and entering microbial signatures.
Think of it as your reference manual. You will return to it frequently, even after completing your first few curations.
What Should You Pay Attention to in the Curation Policy?
You don't need to memorize the entire policy before starting. However, there are some important areas you should become familiar with.
1. Study Design
Study design describes the general framework researchers use to conduct their investigation.
You will encounter different designs, including:
- Case-control studies.
- Cross-sectional studies.
- Prospective cohort studies.
- Randomized controlled trials.
- Longitudinal studies.
Understanding the study design requires more than identifying the terminology used in the paper. You need to understand how participants were selected, how groups were formed, and how the research was conducted.
One important lesson from the policy is to select the study design that best describes the study rather than selecting several options simply because they appear applicable.
2. Experimental Information
When curating an experiment, you will encounter several pieces of information about the study participants and their samples.
These include:
- Host species: The organism from which the microbiome was sampled.
- Location of subjects: The country where participants were recruited, or environmental samples were collected.
- Body site: The anatomical location from which the sample was obtained.
- Condition: The disease, treatment, health status, exposure, or other variable being investigated.
- Contrast: The actual comparison between the groups.
Pay particular attention to the distinction between condition and contrast. They are related, but they are not interchangeable.
3. Laboratory and Sequencing Information
Microbiome studies use different laboratory techniques to identify and quantify microorganisms.
The Curation Policy covers information such as:
- Sequencing type, such as 16S or whole metagenome sequencing (WMS).
- The 16S variable region, where applicable.
- Sequencing platform, such as Illumina.
- Data transformation methods.
Much of this information is usually found in the Materials and Methods section of a scientific paper.
4. Statistical Analysis
This is another important area, particularly if you are coming from a Data Science or Statistics background.
BugSigDB records information about the statistical methods used to identify differentially abundant taxa.
Some of the information you will encounter includes:
- Statistical test.
- Significance threshold.
- Multiple hypothesis testing correction.
- LDA score, where applicable.
- Matching information.
Do not confuse the significance threshold with the actual p-value reported for an individual taxon. The threshold is the criterion used by the researchers to determine statistical significance.
Always check how the paper reports its results and follow the Curation Policy when recording them.
5. Microbial Taxonomy and Signatures
This is where you begin working directly with the microbial findings.
You will need to identify the taxa reported in the paper, determine whether they increased or decreased in the relevant group, and enter them using the appropriate taxonomic identifiers.
BugSigDB uses the NCBI Taxonomy to standardize microbial names.
You will also encounter ontologies and controlled vocabularies when recording other information. For example:
- EFO (Experimental Factor Ontology) for conditions.
- UBERON for anatomical structures and body sites.
- NCBI Taxonomy for microbial organisms.
You don't have to become an expert in these resources immediately. However, understanding why they are used will make your curation work much easier.
A Few Things I Learned Along the Way
Reading the documentation is one thing. Applying it to an actual scientific paper is another.
During my first mini-curation, I worked with a study investigating differences in the gut microbiome between people living with HIV and those without HIV.
One of the things I had to work through was identifying the appropriate condition and its corresponding ontology term.
I also had to pay close attention to the comparison groups and the direction of microbial abundance.
These details may seem small, but they matter.
For example, if a paper reports that a particular bacterium is less abundant in the HIV group, recording it as increased would misrepresent the study's findings.
I also learned that scientific terminology and ontology terms do not always correspond perfectly. Sometimes, a term you encounter in a paper or an existing record may be outdated or may have an alternative identifier in an ontology.
This is why you should never assume that finding a familiar-looking term is enough. Check the appropriate ontology and consult the community when you are uncertain.
The biggest lesson for me was that curation is not simply data entry. It requires reading carefully, interpreting scientific findings, understanding the methodology, and making sure that the information is represented accurately.
Before You Move to Your First Curation
Here is a little checklist to help you prepare:
- [ ] Read the BugSigDB Curation Policy.
- [ ] Understand the difference between a study, experiment, signature, and taxon.
- [ ] Familiarize yourself with the different study designs.
- [ ] Understand how comparison groups and conditions are defined.
- [ ] Learn where to find sequencing and statistical information in a paper.
- [ ] Familiarize yourself with EFO, UBERON, and NCBI Taxonomy.
- [ ] Request your BugSigDB account if you haven't already.
- [ ] Keep the documentation handy. You will need it!
You don't need to know everything before beginning. The important thing is to understand the basics and know where to find answers when you encounter something unfamiliar.
What's Next?
Now that you have a better understanding of BugSigDB, its structure, and the purpose of its Curation Policy, it's time to get practical.
In the next article, we'll explore how to get started with your first contribution, including navigating the project, finding a study to curate, and preparing for your mini-curation.
Remember: take your time, read carefully, and don't be afraid to ask questions.
Happy curating!
