10. Frequently Asked Questions

10.1. rPredictorDB - general

10.1.1. What is rRpredictor?

rPredictorDB is a bioinformatical infrastructure designed to make life easier for people who work with ribosomal RNA. These webpages are the public front-end of the infrastructure. For a more thorough description, see rPredictorDB: the Why & the What.

10.1.2. What can rPredictorDB do?

We offer two general functionalities: searching in our large database of ribosomal RNA and predicting secondary structure of (mostly ribosomal) RNA. More specifically, we can for example:

  • find and retrieve rRNA molecules similar to molecules you give us,
  • compute detailed statistics about the secondary structures predicted for a sequence,
  • predict secondary structure for your rRNA sequences with an algorithm developed specifically with respect to ribosomal RNA

and much more. See the rPredictorDB record detail for a detailed description of information available from our database and CP-predict: a two-phase algorithm for rRNA structure prediction for details on the custom prediction algorithm.

10.1.3. That’s nice, but I don’t understand what it means!

The Biological background page will help you understand what’s going on here.

For unknown terminology, make sure you check out the rPredictorDB glossary.

10.1.4. What kind of information can I find on rPredictorDB?

We integrate data and information from the SILVA, Rfam and Taxonomy (NCBI) databases and the European Nucleotide Archive and add structural data. The records in our database have several sections. For each record, we provide:

  • Information about sequence origin: unique ID of source molecule, organism name, taxonomy, etc.
  • Information about sequence itself: length, the actual sequence, various quality indicators
  • Predicted secondary structures, an analysis of structural features and a visualisation of the predicted structure
  • References to relevant scientific literature
  • Cross-references to other databases containing the same sequence
  • Additional annotations from the source databases

A complete description of the individual fields for search results is available here: rPredictorDB record detail.

10.1.5. How can I run your prediction algorithm on my data?

Have a look at the rPredictorDB Predict page. rPredictorDB’s custom algorithm is Cp-predict v.2. A detailed description of the input fields is available here: CP-predict2

10.1.6. There are many similar sites. Why should I use rPredictorDB?

While there are several directly to ribosomal RNA, we (the Bioinformatics laboratory of the Microbiological Institute of the Czech Academy of Sciences and our team at the Faculty of Mathematics and Physics at Charles University) feel that a better job could have been done.

The main drawbacks of similar bioinformatical websites are a sharp learning curve and missing information or unclear purpose of information. In order to use a website such as the Protein Data bank or the Comparative RNA website, a user first has to have - or obtain - a good general idea of what the site does, why would anyone want such a site and what a lot of terminology means.

After this background knowledge is obtained, a user finds out that certain information is not well-curated or made explicit: no one (at least publicly) keeps track of whether a RNA sequence was obtained directly of from a DNA transcription site, information about the type of rRNA is sparse, sometimes the phylogeny of a molecule is missing, a resolved secondary structure is not sufficiently labeled, it is unclear what constitutes a truly unique identifier of an RNA sequence or structure, etc. For instance, the STRAND database of resolved RNA secondary structures contains a number of structures inconsistently labeled for RNA-protein complexes and duplicate sequences.

Another deficiency of such sites is often little or missing support for mass data retrieval. When present, it is usually in the form of pre-packaged archives and the user has little choice over what subset of the data to download. (A notable exception here is the SILVA database of rRNA molecules.)

While perhap such drawbacks are of less concern to biologists, the fast-growing field of bioinformatics is sensitive to this kind of volatility in data sources. Nobody wants to spend a lot of time by finding out what uniquely identifies an RNA sequence or structure, filling in missing fields, etc., let alone downloading a set of several hundred sequences of interest one by one.

rPredictorDB was designed with overcoming these drawbacks in mind. Our goal is first and foremost clarity: answers to questions such as “Why would I ever want to do that?” should be easy to find (and relatively easy to read).

Also, we may randomly decide to give you a cookie.

10.1.7. I have a great program you should use!

You can go right ahead and write a rPredictorDB tool for your program. It’s not hard! rPredictorDB was designed with extensibility in mind. How to write a tool is described here: Creating your own tool.

If you don’t feel like coding but still want to contribute your program, contact us and if you convince us your program is worth it - which might not be so hard -, we’ll implement the tool ourselves.

10.2. Site usage

10.2.1. It’s confusing: too many buttons!

Try the rWeb Tutorial and see if it’s still confusing afterwards. If it is, then ask us!

10.2.2. X isn’t working!

Tell us! Make sure you describe your problem so that we exactly follow your steps to reconstruct it. If a screen with a large red strip at the top comes up and shouts about some kind of error, send us a screenshot! (Actually, you can send us a screenshot of the error anyway, it will nearly always help.) We’ll get in touch.

10.2.3. Doing Y should be possible.

Tell us! We’ll get in touch and discuss the functionality with you.

If Y is a tool, you can also become an rPredictorDB developer and implement it yourself, according to your dreams and wishes - extending rPredictorDB with new tools is quite simple. (One of the key design concerns of rPredictorDB was extensibility.) Here’s how to do it: Creating your own tool

You still need to tell us, though, to get developer access.

10.2.4. Can I help?

Of course! You can best start by telling us about it, because nearly all ways of helping require access to the rPredictorDB development repository.

You can contribute in many ways. Some ideas, from less to more technical:

  • Feedback. We appreciate any and all feedback. By telling us what you like or don’t like about rPredictorDB, you help us make it better and thus help yourself to a better service.
  • Biological expertise. If you stumble across anything that’s incomplete, imprecise or unclear, we’ll be more than happy if you can help us clarify it.
  • The Glossary always needs more terms, better formulations, etc.
  • Testing. This is a step up from feedback: you can participate in the bug hunt. Weird corner cases for input parameters, pointless tool combinations - the gloves are off. (Just please inform us before you bring the server down.) Testing rPredictorDB means detailed feedback: what exactly you did, whether it worked or not. If you wish to test, contact us and we’ll point you to parts of rPredictorDB that need tester attention.
  • Tool development - the more, the merrier (within reason). Extensibility is one of the key design concern in rPredictorDB. Here’s how to do it: Creating your own tool