iRefScape Test Cases 1.0

Last edited: 2011-05-11

Search cases

All tests have been performed against iRefIndex 8.0.

Query	Search Type	Options	Expected Result	Pass/Fail
Q39009 Q9ZNV8	UniProt_Ac	Taxon id: Any Iterations: 1 Create new view: no Use canonical expansion: no Add edges between neighbours: no	33 nodes, 94 edges. 2 queried nodes are blue (DMC_ARATH and AHP2_ARATH)	Pass
			Node AHP2_ARATH linkout from i.RefSeq_Ac_TOP attribute (NP_189581) to Entrez Protein gives http://www.ncbi.nlm.nih.gov/protein/NP_189581?report=GenPept	Pass
			i.taxid is 3702 and i.geneID is 822593	Pass
			Node AHP2_ARATH linkout from i.UniProt_Ac_TOP attribute (Q9ZNV8) to UniProt/KB beta gives http://www.uniprot.org/uniprot/Q9ZNV8	Pass
			UniProt record agrees with iRefScape on iRefSeq_Ac_TOP (see "Sequence databases"), i.taxid, i.geneID (see "Genome annotation databases")	Pass
			Two edges between query nodes are EBI-1555390, EBI-1555417	Pass
			Linkouts for query node edges (i.src_intxn_id) to Intact ("Interaction databases") provide PubMed #17937504 which should match i.PMID, and an interaction detection method of "anti tag coip" which should match i.method_name	Pass
			The molecule names are DMC1 and ATHP1 in IntAct and these names should be available under the i.interactor_alias node attribute	Pass
			Both interactions should have http://wodaklab.org/iRefWeb/interaction/show/102203 as i.iRefWEB	Pass
Q39009-1 Q9ZNV8-2			Return the same results as for UniProtKB: the isoform information is ignored when searching
Q39009.1 Q9ZNV8.2			Returns no results. Version information is not a valid annotation for UniProtKB.

Test case name: RefSeq_Ac 1

Operation:

Query box:

NP_188928.2

Search type: RefSeq_Ac

Taxon id: Any

Iterations: 1

Create new view: yes

Use canonical expansion: no

Expected Observation:

11 nodes, 29 edges. i.RefSeq_Ac = NP_188928 was returned. Edges returned include two edges described in test UniProt_Ac 1.

The same results are returned if the query box is changed to NP_188928.

The same results are returned if the query box is changed to NP_188928.567.

No results if the query box is changed to 188928. (Expected behaviour-OK)

No results are returned if the query box is changed to NP 188928. (Expected behaviour-OK)

Change the search type to UniProt_Ac. The interface says no results and when user presses ok.(Expected behaviour-OK)

The search type can also be changed to UniProt_ID, geneID, geneSymbol, ipi,, mass or rog with no results returned.

No results are returned if the query box is changed to 0. No results are returned if the query box is changed to abcde.

Last tested: April 20, 2011

Plugin Version: 1.00

Data Version: 8.0

Result: PASS

Tester: Sabry

Confirmed by: NA

Test case name: UniProt_ID 1

Operation:

Query box:

AHP2_ARATH

Search type: UniProt_ID

Taxon id: Any

Iterations: 1

Create new view: yes

Use canonical expansion: no

Expected Observation:

26 nodes, 67 edges.

No results are returned if the query box is changed to AHP2. (Since UniProt_ID is now an exact-only search. The production of results should not be allowed here since the user might have entered gene names, selected UniProt_ID by mistake and would not be aware of their mistake because their search returned results. Also, since the first few characters of UniProt_IDs may be shared by multiple proteins from different organisms, an inexact match would need to trigger the query helper.)

AHP2 ARATH returns no results.

ARATH returns no results. (Since UniProt_ID is now an exact-only search.)

AH2_ARATH returns no results. (Previously it returned CAH2_ARATH. Imagine if the user accidentally had such a query term embedded in a long list. They would never detect this search error! This should be is an exact match search by default - which it now is - or require use of the query helper to allow the user to decide.)

AHP2_ARATH.2 returns no results.

23 returns no results.

Searching for AHP2_ARATH but with search type UniProt_Ac, geneID, geneSymbol, ipi or mass all return 0 results. The interface says no results and when user presses ok, shows a blank panel. (The geneID search gives a warning about any organism being specified in the criteria.)

No results are returned if the query box is changed to 0.

No results are returned if the query box is changed to abcde.

Last tested: April 28, 2011

Plugin Version: 1.0

Data Version: 8.1

Result: PASS

Tester: Paul

Notes: (SABRY) UniProt_ID is now EXACT match only. If the user failed to enter exactly matching query no results will be returned. Query helper will never be invoked.

Test case name: geneID 1

Operation:

Query box:

821860
822593

Search type: geneID

Taxon id: Any

Iterations: 1

Create new view: yes

Use canonical expansion: no

Expected Observation:

Should return the same result as a UniProt_Ac query for Q39009 and Q9ZNV8 (33 nodes and 94 edges).

0 returned no results.

00001 returned no results.

abcdes returned no results.

1234 returned 47 nodes and 158 edges.

Note	Keep the following example for further reference.

Retrieved all geneIds from above query (using "File" -> "Export" -> "Node Attributes" and choosing "i.geneID", then saving and processing the saved file to get a list, producing 73 values which reduced to 71 unique gene identifiers) and queried (iterations = 0). This should return at least as many nodes as above (47 protein nodes), and in fact returns 110 nodes (31 complex nodes and 79 protein nodes).

The gene in question is CXCR4 and has two isoforms (P61073-2 and P61073).

A geneSymbol search for CXCR4 (iteration=0) returns two nodes as expected.

A UniProt_Ac search for P61073-2 returns CXCR4 and CXCR4_HUMAN, whereas P61073 returns only CXCR4_HUMAN when canonical expansion is turned off. Both searches return two nodes when canonical expansion is turned on. POSSIBLE FAIL Notes: (Paul) This is recognised (if somewhat counterintuitive) behaviour: the isoform-based search for P61073-2 actually results in the removal of the "-2" and a search for all isoforms, under the assumption that the user is unsure which isoform should be retrieved; as a result, all isoforms are returned, even though a specific isoform was requested. In contrast, the search without a "-" character results in just one protein with that exact name being returned.

This can be used as a test case for the Export function "on Target network Node ID type"= i.Canonical_ROGID_TOP

Searched 821860 using geneID (iterations = 0) which returned one node. Searching using other search types (geneSymbol, UniProt_Ac, UniProt_ID) returned no results.

Last tested: April 29, 2011

Plugin Version: 1.0

Data Version: 8.1

Result: SEE NOTE ABOVE

Tester: Paul

Notes: (SABRY) The search results of a protein accession and a gene identifier (which could refer to more than one protein) will not be the same and it is OK as I see it (correct me if I am wrong). But as requested if canonical expansion is turned on the results will be the same. Although this is a very good suggestion for this search, I am sorry that I am against the idea of default canonical expansion and hiding the stop expansion in advanced (which is now features under development), but this does not mean I will not implement it.

1. When turned on canonical expansion affects many of the default behavior (abstract implementations). I use the same network construction for ROG, molecular mass and geneID and so on.
2. The expansion uses a second phase index to index mapping and this slows down the query.
3. Batch queries will take long time to complete and will get more interactions. This seems to be a good thing, but the current network is already to much information, and the user has the option to expand if wanted.

Let me know the final decision.

Test case name: geneSymbol 1

Operation:

Query box:

CXCR4

Search type: geneSymbol

Taxon id: Any

Iterations: 0

Create new view: yes

Use canonical expansion: no

Expected Observation:

Returns two nodes with i.UniProt_Ac_TOP of P61073 and P61073-2.

Change search to PTK2. The user should be warned that they have not selected a taxon. "You have not limited your search to a specific organism. Proceed anyway? Yes No" This warning is given. With no taxon specifier present, the search (iteration=0) returns 6 nodes and 5 taxons. (This would have created havoc if the user had searched using a list of identifiers and no warning had been given.)

0 returns no results. Ok.

0000 returns no results. Ok.

12345 returns no results. Ok.

abcde returns no results. Ok.

cxcR4 returns identical results to CXCR4. Ok.

CXCR returns no results. Ok.

A search for "CXC CXCR CXCR4" (using newlines, not spaces, to separate the terms) returns results identical for CXCR4. Ok. However, no indication is given that CXC and CXCR failed to provide matches. Notes (PAUL): maybe some feedback could be given about this.

Search CXCR5 as a UniProt_ID, returned no results.

Search CXCR5 against other query types returned no results. Ok.

Last tested: May 9, 2011

Plugin Version: 1.0

Data Version: 8.1

Result: TESTED

Tester: Paul

Notes (SABRY):

1. hyperlink removed
2. Do you want to check the taxonid only for geneSymbol ?
3. The default batch file mode the user has to specify a taxonid, this was overridden as requested in an earlier release should this be reactivated?

Test case name: geneSymbol 2

Operation:

Query box:

RPB1

Search type: geneSymbol

Taxon id: Any

Iterations: 0

Create new view: yes

Use canonical expansion: no

Expected Observation:

RPB1 for any taxon returns 1 node (RPB1_SCHPO). Ok.

RPB1 for taxon 4932 returns 0 nodes. Ok. Even though an alias for yeast RPO21 (Gene Id 851415) is RPB1, this search only searches on official gene symbols from Entrez.

RPB1 for taxon 9606 returns 3 nodes one of which is SIRPB1. FAIL. This is an exact match search.

Last tested: June 7, 2010

Plugin Version: 0.83

Data Version: 7.0 public

Result: NOT TESTED

Tester: Ian ' Notes: ':(SABRY) :Changed to exact search.

1. There is a conflict between Test case name: geneID 1 and Test case name: geneID 2. When the search is now made a exact search, the query helper will never pop-up.

Test case name: rog search 1

Operation:

Query box:

5366033

Search type: rog

Taxon id: Any

Iterations: 1

Create new view: no

Use canonical expansion: no

' Notes: ':(SABRY) The bug coursing freeze when performing ROG searches is fixed in 0.85

Test case name: Iterations 1

Operation:

Perform search described in test UniProt_Ac 1 with Iterations=1. Then again with Iterations = 0 and confirm that the number of edges between the query nodes is identical.

Expected Observation:

Same number of edges (2 edges, 1 loop) in both cases.

Last tested: June 1, 2010

Plugin Version: 0.83

Data Version: 7.0 public

Result: NOT TESTED

Tester: Ian

Notes:

Test case name: ipi 1 #

Operation:

Query box:

IPI00543858
IPI00517160

Search type: ipi

Taxon id: Any

Iterations: 1

Create new view: no

Use canonical expansion: no

Expected Observation:

Returns results identical to test case UniProt_Ac 1. Ok

Changing the query to "IPI00543858.1 IPI00517160.1" produces identical results. Ok.

Changing the query to "00543858 00517160" produces identical results. Ok.

Changing the query to "00543858 0051716" produces identical results for first query term and initiates query helper to resolve second term. Ok.

Changing the query to "abcd" produces no results. Ok.

Changing the query to "IPI" initiates query helper. Ok.

Searching for IPI00543858 as any other search type returns no results. Ok

Last tested: June 1, 2010

Plugin Version: 0.83

Data Version: 7.0 public

Result: NOT TESTED

Tester: Ian

Notes:

Test case name: mass search #

Operation:

Query box:

72854<-->72866

Search type: ipi

Taxon id: Any

Iterations: 0

Create new view: yes

Use canonical expansion: no

Expected Observation:

The user is presented with the query helper to refine the query. A prompting message reads:

"One or more of your mass query terms corresponds to a range. You may select one or more of the masses on the right (shift-click and cntrl-click are supported). Press return to add the selected terms to a new query. Clear the "Create new view" tick-box if you want to add the results to the current view."

FAIL. The message given should be replaced with the above text. Also, there are inappropriate and unusable links to NCBI given in the query helper. Also, interface changes my selected iterations from 0 to 1 without asking. Annoying.

The query returns 4 nodes with mass in the expected range. Ok.

0 returns no results. Ok.

-1 returns no results. Ok.

MW returns no results. Ok.

72854 kda returns no results but hangs the interface. FAIL.

Last tested: June 1, 2010

Plugin Version: 0.83

Data Version: 7.0 public

Result: NOT TESTED

Tester: Ian

Notes:

Test case name: PMIDsearch 1

Operation:

Query box:

11401546

Search type: pmid

Taxon id: Any

Iterations: 0

Create new view: yes

Use canonical expansion: no

Expected Observation:

4 nodes and 7 edges are returned. Three edges have PMID 11401546 and these involve the 4 nodes shown. All other edges (from different PMIDs) involve these proteins. Ok.

11401546.1 returns 0 results. Ok.

1140154 returns 0 results and starts query helper. FAIL. This should be an exact match search.

SPTAN1 returns 0 results. Ok.

11401546 SPTAN1 returns 0 results. Ok.

0 finds 21493 matches and then says there are no results and open the query helper. FAIL. This is an exact match search.

-1 returns 0 results. Ok.

10551855 [return character] 11401546 returns 6 nodes and 27 edges. Ok.

10551855|11401546 returns 6 nodes and 27 edges. Ok.

10551855| 11401546 returns 6 nodes and 27 edges. Ok.

10551855 | 11401546 returns 6 nodes and 27 edges. Ok.

10551855 [tab] 11401546 returns 6 nodes and 27 edges. Ok.

10551855, 11401546 returns 6 nodes and 27 edges. Ok.

10551855 11401546 returns 0 results. Ok. We dont day we support space delimited queries.

Last tested: June 7, 2010

Plugin Version: 0.83

Data Version: 7.0 public

Result: NOT TESTED

Tester: Ian

Notes:

Test case name: example search 1

Operation:

Test the example searches listed at http://irefindex.uio.no/wiki/README_Cytoscape_plugin_0.8x#Using_the_Search_Panel

Query box: see below

Search type: see below

Taxon id: Any

Iterations: 1

Create new view: yes

Use canonical expansion: yes

Expected Observation:

Search Type	Example	Results
UniProt_Ac	Q7KSF4	67 nodes, 146 edges. 4 nodes are returned directly by the query, but only one is marked with i.query = Q7KSF4. FAIL nodes returned by a query (even with canonical expansion) should have an i.query value. i.order is 0 for Q7KSF4 but 10 for the other three nodes from the same canonical group. CAUTION. This is not what the documentation says for these attributes. With canonical expansion off, this search returns 43 nodes, 43 edges all from BioGRID.
RefSeq_Ac	NP_996224	67 nodes, 146 edges. See above. i.query feature fails to update to reflect the most recent search. FAIL.
UniProt_ID	Q7KSF4_DROME	67 nodes, 146 edges.
geneID	42066	67 nodes, 146 edges. See above. Now two nodes are marked as Q7KSF4 under i.query. FAIL. 42066 should be listed for the recently returned node.
geneSymbol	cher	67 nodes, 146 edges. Now i.query says Q7KS4 for all four nodes returned by the search. FAIL.
mass	72854<-->72866	40 nodes, 94 edges. Four molecular mass values are returned by the query helper. Selecting all of these and doing the search returns 8 hi-lited nodes and their interactors. The 72855 mass node returns 1 node with that mass and two other members of the same canonical group. Two of the nodes are returned with i.query of Q7KSF4. FAIL.
rog	10121899	FAIL. This should return the Q7KSF4 node. Instead got "Sorry, please wait until current operations finish or stop the proces and restart Cytoscape." Subsequent searches do not work. Have to restart Cytoscape
PMID	14605208	929 nodes and 1605 edges returned all with PMID of 14605208. SAme results if iteration is set to 0. Ok. Same results if canonical expansion is de-selected. Ok.
src_intxn_id	47513	2 nodes and 1 edge returned.
src_intxn_id	EBI-212627	2 nodes and 4 edges returned (2 from intact, 1 from BIND and 1 from MINT). Ok
src_intxn_id	147805	Returns 4 nodes and three interactions because BIND and BioGRID interaction id space overlap.
omim	227650	74 nodes and 224 edges returned. Ok.
digid	449	58 nodes and 60 edges returned. FAIL. This should return the same results as omim 227650 plus more nodes that have the same digid. digid search is like searching for multiple omims. Problem with file name _EXT__ROG_digid.irft
dig_title	fanconi	FAIL - do not fix until next release.

Last tested: June 7, 2010

Plugin Version: 0.83

Data Version: 7.0 public

Result: NOT TESTED

Tester: Ian

Notes:

Preferences

Test case name: what is being tested #

Operation:

Expected Observation:

Last tested: July 8, 2010

Plugin Version: 0.83

Data Version: 7.0

Result: NOT TESTED

Tester: Ian

Notes:

Test case template

Test case name: what is being tested #

Operation:

Expected Observation:

Last tested:

Plugin Version:

Data Version:

Result: NOT TESTED

Tester:

Notes:

To be corrected

• Remove the non-proprietary flag/check for current data
• Handle the neighbourhood completion when expanding network (do not use all the nodes)
• show_inxc dynamic index behaviour change not working
• Scaling GUI at low resolution, maximise button may get hidden
• Path finding cancelling time
• focus progress when path finding
• Ending with collapsed node error
• unselect all nodes before edge filtering

List of GeneIDs to test the new caonical expansion

• Available in data version 8.3