I am new to bioinformatics and natural language processing. In linguistics, they have Treebank - derivation/parse tree for each sentence. For example, a sentence "Sara sleeps" can be visualized by a tree: Sentence -> Subject Verb, Subject -> Sara, Verb -> sleeps.

I've heard about Genbank, a database for nucleotide sequences. Does a DNA has its own derivation tree, or is any pattern in a DNA sequence already known with current technology? I only heard about repeats.

  • 3
    $\begingroup$ Hello kate, I don't understand your question. The "derivation tree" you are describing is a structure which is specific to linguistics. How could a DNA have such a thing? It does not represent human language. $\endgroup$ – rumtscho Nov 15 '15 at 13:58
  • $\begingroup$ Thank you for the comment. I think DNA has its own "language" which is probably unknown yet. My question is,(1) is the "language" of DNA already known? (2) is there any hint already known for understanding the "language" of DNA? Hint might be anything, such as patterns.. because I don't really understand much about bioinformatics. $\endgroup$ – kate Nov 15 '15 at 14:03
  • 1
    $\begingroup$ DNA does not have its own language, it's a simple storage medium. It encodes information, but not in any way similar to a human language. Maybe this encoding fulfils the definition of a formal language, but if you have never gone into that branch of maths, any comparison will be more confusing than enlightening. $\endgroup$ – rumtscho Nov 15 '15 at 14:31
  • $\begingroup$ So better to think it as a storage medium.. Thank you very much! $\endgroup$ – kate Nov 15 '15 at 16:23
  • $\begingroup$ The Gene Ontology Consortium might be of interest to you. $\endgroup$ – SANBI samples Nov 15 '15 at 17:12

From your example I guess that by "deriving" you mean something like decomposing into elements with diverse functions. Maybe we could apply this concept to DNA sequences: Some parts code proteins, which are decomposed into triplets, each of which codes an amino-acid. Other parts have regulatory functions: They can be motifs not coding a protein, but instead allowing a protein or something else to bind and modify the expression of nearby protein-coding parts. Other parts may contribute to the physical structure of the chromosome, or may have no function at all.

| improve this answer | |
  • $\begingroup$ Yes, this is what I mean. Thank you very much for your answer! Do all these - e.g. where is the part in a DNA sequence that involve in function x - already known? $\endgroup$ – kate Nov 16 '15 at 16:42
  • 1
    $\begingroup$ There are regulatory motif databases, and prediction algorithms for functions in a sequence, but all this is not fully known. Indeed, things are rarely fully regular in biology, so every bioinformatics predictions ideally should be completed with experimental confirmation. Also, the same sequence can sometimes be parsed in various way: a single DNA sequence may encode several proteins, depending on which portions are actually translated into amino-acids. Some keywords you may be interested in: start codon, stop codon, open reading frame, 3'UTR, 5'UTR, binding box. $\endgroup$ – bli Nov 17 '15 at 9:34

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.