GOtcha Help

Help with GOtcha

GOtcha requires very few parameters in order to run. You need a sequence (DNA or Protein), choose genomes against which to compare your sequence, select evidence codes for the annotations you are prepared to accept and select a cutoff. GOtcha does all the rest.

Sequence input

GOtcha can take a sequence in any format accepted by EMBOSS in three ways. Please select the appropriate method.
As a text file you upload
You can select the text file option and upload a sequence from a file on your computer. This should be a text file (i.e. a file that can be read by notepad). Word documents are not text files and cannot be read by the GOtcha server.
As sequence text pasted in to the box
You can paste your sequence directly into the text box. Please make sure that you copy the whole of the sequence file properly. If in doubt, make sure it corresponds to fasta format.
As an EMBOSS Uniform Sequence Address (USA)
This is primarily aimed at local Dundee users who have a sequence reference they wish to search.A USA is an EMBOSS convention for referring to a particular database entry as database:id. So the sequence with swissprot ID TF_HUMAN would be sw:tf_human.
GOtcha will use EMBOSS to automatically determine the format and type of sequence and will then run the appropriate database searches. The sequence GOtcha uses can be found in the results archive as query.seq

Databases

GOtcha performs a function prediction by matching the query sequence against a number of well annotated genomic datasets. The final results are averaged across these genomes. If you wish, you can only select a subset of genomes against which to compare your sequence. If in doubt select all.

Evidence codes

Gene Ontology assign evidence codes to the gene product - function associations. These evidence codes reflect the degree of confidence in the robustness of a particular annotation. At the bottom of the scale are terms like IEA, Inferred from Electronic Annotation where a program like GOtcha or BLAST has been used to assign function and has not been checked by a real biologist. At the other end of the scale are peer reviewed experiments, Traceable Author Statement. GOtcha allows you to select which annotations you will accept.

The real decision is whether to include IEA annotations or not.
GOtcha is more accurate without IEA annotations but may miss some functions. If in doubt, run once with all terms and once with all terms except IEA.

Probability cutoff

For some searches there will be a large number of very low scoring predictions. You have teh opportunity to screen these out using the probability cutoff. This has no effect on the way GOtcha runs, only on the way the results are displayed graphically. A cutoff of 10 or 20% is reccommended for most searches.