GExplore - Frequently Asked Questions

Frequently asked questions

General questions

Should something happen when I click on one of these arrows ?

Should something happen when I click on one of these arrows javascript is enabled !

Indeed, clicking on such an arrow should reveal some hidden text. In fact, just clicking on the text next to the arrow should have the same effect. This requires that javascript is enabled in your browser. To test click on the arrow above. If nothing happens, javascript is not enabled in our browser. Check your browser manual in this case to activate javascript.

The results page says: 'click on table headers to sort' - but nothing happens when I click?

There are three possible reasons: 1. only certain columns can be used for sorting (e.g. the domain column does not sort). 2. The output has more than 1000 genes, in which case this functionality is not available (it would be so slow, you wouldn't want to wait) 3. This requires that javascripts are enabled in your browser (see previous question for a quick test).

Can I upload my favourite set of genes for analysis ?

No, but you can paste even a large list of genes (I tested up to 5000) into the Gene Name search field. Simply generate a comma-delimited list of your favourite genes and paste them into the Gene Name search field.

Can I save the search results from one search (and use them as a starting point for additional searches) ?

You cannot directly save search results, but ....
The result of a search essentially is a list of genes fullfilling the search condition(s). This list of genes is always on the output page (right at the top). You can copy it from there and save it on your computer - or paste it into the 'gene name' search field for combinatorial searches.

How can I sort output with more than 1000 genes?

You cannot sort by simply clicking into the table header, since this would be very slow. However, you can use the 'limit output to the first xxx genes after sorting by ....'. While you can use this to reduce the number of genes shown, you don't have to. If the number you enter is larger than the number of genes in the output, you will get the complete set of genes and you can quickly sort them.

Can I turn off this row-drag function? (it prevents me from copying text out of text columns)

Turning off the domain display will turn off the row-drag functionality. It's main purpose is to allow you to drag proteins around to maybe group those with similar domain organisation.

These checkboxes next to genes names on the output page: what exactly can I do with them ?

If you uncheck a gene and click 'Update Display' the unchecked gene will be removed from the display. You an use this to manually fine-tune the output and maybe eliminate false positives from a domain search or certain genes you are not interested in before continuing with the analysis. Do can use the 'Back' button of your browser to 'undo' this quickly.

Domain Search and Display

The table of abbreviations doesn't cover my favourite domain. How can I search for proteins containing this domain?

Find the SMART or Pfam identifier of your domain of interest and use it in the domain search field.

The table of abbreviations collapses all the zinc-fingers (kinases, GTPases, etc). How can I retrieve only some zinc-finger proteins (kinases, GTPases) ?

If you care about certain subgroups, you can always use the corresponding SMART or Pfam identifiers in your search. The domain will still be displayed as generic 'Znf' ('KIN', 'GTPase'). There are two reasons for 'collapsing' sometimes large numbers of interpro domains into a single abbreviation. First, this interface is aimed at the 'big picture' (actively ignoring subtle subdivisions) and second, SMART or Pfam domains sometimes do not have the discriminatory power to distinguish subgroups (EGF and IG are typical examples). However, currently some groups of domains might be collapsed too generously. If you are an expert in the field and can suggest a subdivision, contact me...

I'm trying to get all transmembrane proteins, but a lot of proteins I got seem to have only a signal peptide (SP), but no transmembrane domain ?!

Try 'TM' in the Domain Pattern field. Explanation: Signal peptides have a hydrophobic core, which triggers a hit with algorithms looking for hydrophobic stretches of amino acids to predict transmembrane domains. While searches in the Protein Domain field operate on the 'raw' domain annotations, searches in the Domain Pattern field operate on the domain pattern as you see it in the ouput, i.e. after the 'display rules' have been applied - and one of them suppresses display of transmembrane domains overlapping signal peptides.

Can I search for proteins that do not have any known domain?

Yes, all proteins not having any SMART or Pfam annotation are tagged with 'no_domain'. You can enter this in the Protein Domain search field to retrieve them. Note that signal peptides and transmembrane domains are not considered 'domains' here. If you want proteins that have neither signal sequene, nor transmembrane domain nor any known domain, use 'no_domain not (SP or TM)' in the Protein Domain search field (there are surprisignly few). Note also that you cannot use 'no_domain' as part of a search pattern in the Domain Pattern field. However, since entries in the Protein Domain search field and the Domain Pattern field and effectively combined as logical 'AND' search, you can get all single pass transmembrane domain proteins with no known protein domains by searching for 'no_domain' in the Protein Domain field and 'TM and no more TM' in the Domain Pattern field.

Text searches and boolean logic in searches

I tried something like 'body wall muscle AND neuron' in the expression search field and got an error message. Why ?

The short answer is, you cannot use multi-word search terms, because SPACE is used to separate search terms. You will not get an error message from just using 'body wall muscle', because itis translated into 'body OR wall OR muscle' - probably not what you wanted. As a work-around for the example above, you could use the following search strategy: search for 'body AND wall AND muscle', copy the list of genes from the output page (Control-C), paste it into the Gene Name field of a new search page and search for 'neuron'.