JAL-3017 parse multiple variant elements for uniprot feature
[jalview.git] / test / jalview / ws / dbsources / UniprotTest.java
1 /*
2  * Jalview - A Sequence Alignment Editor and Viewer ($$Version-Rel$$)
3  * Copyright (C) $$Year-Rel$$ The Jalview Authors
4  * 
5  * This file is part of Jalview.
6  * 
7  * Jalview is free software: you can redistribute it and/or
8  * modify it under the terms of the GNU General Public License 
9  * as published by the Free Software Foundation, either version 3
10  * of the License, or (at your option) any later version.
11  *  
12  * Jalview is distributed in the hope that it will be useful, but 
13  * WITHOUT ANY WARRANTY; without even the implied warranty 
14  * of MERCHANTABILITY or FITNESS FOR A PARTICULAR 
15  * PURPOSE.  See the GNU General Public License for more details.
16  * 
17  * You should have received a copy of the GNU General Public License
18  * along with Jalview.  If not, see <http://www.gnu.org/licenses/>.
19  * The Jalview Authors are detailed in the 'AUTHORS' file.
20  */
21 package jalview.ws.dbsources;
22
23 import static org.testng.AssertJUnit.assertEquals;
24 import static org.testng.AssertJUnit.assertFalse;
25 import static org.testng.AssertJUnit.assertNotNull;
26 import static org.testng.AssertJUnit.assertNull;
27
28 import jalview.datamodel.PDBEntry;
29 import jalview.datamodel.SequenceI;
30 import jalview.datamodel.xdb.uniprot.UniprotEntry;
31 import jalview.datamodel.xdb.uniprot.UniprotFeature;
32 import jalview.gui.JvOptionPane;
33
34 import java.io.Reader;
35 import java.io.StringReader;
36 import java.util.Vector;
37
38 import org.testng.Assert;
39 import org.testng.annotations.BeforeClass;
40 import org.testng.annotations.Test;
41
42 public class UniprotTest
43 {
44
45   @BeforeClass(alwaysRun = true)
46   public void setUpJvOptionPane()
47   {
48     JvOptionPane.setInteractiveMode(false);
49     JvOptionPane.setMockResponse(JvOptionPane.CANCEL_OPTION);
50   }
51
52   // adapted from http://www.uniprot.org/uniprot/A9CKP4.xml
53   private static final String UNIPROT_XML = "<?xml version='1.0' encoding='UTF-8'?>"
54           + "<uniprot>"
55           + "<entry dataset=\"TrEMBL\" created=\"2008-01-15\" modified=\"2015-03-04\" version=\"38\">"
56           + "<accession>A9CKP4</accession>"
57           + "<accession>A9CKP5</accession>"
58           + "<name>A9CKP4_AGRT5</name>"
59           + "<name>A9CKP4_AGRT6</name>"
60           + "<protein><recommendedName><fullName>Mitogen-activated protein kinase 13</fullName><fullName>Henry</fullName></recommendedName></protein>"
61           + "<dbReference type=\"PDB\" id=\"2FSQ\"><property type=\"method\" value=\"X-ray\"/><property type=\"resolution\" value=\"1.40\"/></dbReference>"
62           + "<dbReference type=\"PDBsum\" id=\"2FSR\"/>"
63           + "<dbReference type=\"EMBL\" id=\"AE007869\"><property type=\"protein sequence ID\" value=\"AAK85932.1\"/><property type=\"molecule type\" value=\"Genomic_DNA\"/></dbReference>"
64           + "<feature type=\"signal peptide\" evidence=\"7\"><location><begin position=\"1\"/><end position=\"18\"/></location></feature>"
65           + "<feature type=\"propeptide\" description=\"Activation peptide\" id=\"PRO_0000027399\" evidence=\"9 16 17 18\"><location><begin position=\"19\"/><end position=\"20\"/></location></feature>"
66           + "<feature type=\"chain\" description=\"Granzyme B\" id=\"PRO_0000027400\"><location><begin position=\"21\"/><end position=\"247\"/></location></feature>"
67           + "<feature type=\"sequence variant\"><original>M</original><variation>L</variation><location><position position=\"41\"/></location></feature>"
68           + "<feature type=\"sequence variant\" description=\"Pathogenic\"><original>M</original><variation>L</variation><location><position position=\"41\"/></location></feature>"
69           + "<feature type=\"sequence variant\" description=\"Pathogenic\"><original>M</original><location><position position=\"41\"/></location></feature>"
70           + "<feature type=\"sequence variant\" description=\"Foo\"><variation>L</variation><variation>LMV</variation><original>M</original><location><position position=\"42\"/></location></feature>"
71           + "<sequence length=\"10\" mass=\"27410\" checksum=\"8CB760AACF88FE6C\" modified=\"2008-01-15\" version=\"1\">MHAPL VSKDL</sequence></entry>"
72           + "</uniprot>";
73
74   /**
75    * Test the method that unmarshals XML to a Uniprot model
76    */
77   @Test(groups = { "Functional" })
78   public void testGetUniprotEntries()
79   {
80     Uniprot u = new Uniprot();
81     Reader reader = new StringReader(UNIPROT_XML);
82     Vector<UniprotEntry> entries = u.getUniprotEntries(reader);
83     assertEquals(1, entries.size());
84     UniprotEntry entry = entries.get(0);
85     assertEquals(2, entry.getName().size());
86     assertEquals("A9CKP4_AGRT5", entry.getName().get(0));
87     assertEquals("A9CKP4_AGRT6", entry.getName().get(1));
88     assertEquals(2, entry.getAccession().size());
89     assertEquals("A9CKP4", entry.getAccession().get(0));
90     assertEquals("A9CKP5", entry.getAccession().get(1));
91
92     /*
93      * UniprotSequence drops any space characters
94      */
95     assertEquals("MHAPLVSKDL", entry.getUniprotSequence().getContent());
96
97     assertEquals(2, entry.getProtein().getName().size());
98     assertEquals("Mitogen-activated protein kinase 13", entry.getProtein()
99             .getName().get(0));
100     assertEquals("Henry", entry.getProtein().getName().get(1));
101
102     /*
103      * Check sequence features
104      */
105     Vector<UniprotFeature> features = entry.getFeature();
106     assertEquals(7, features.size());
107     UniprotFeature sf = features.get(0);
108     assertEquals("signal peptide", sf.getType());
109     assertNull(sf.getDescription());
110     assertNull(sf.getStatus());
111     assertEquals(1, sf.getBegin());
112     assertEquals(18, sf.getEnd());
113     sf = features.get(1);
114     assertEquals("propeptide", sf.getType());
115     assertEquals("Activation peptide", sf.getDescription());
116     assertEquals(19, sf.getPosition());
117     assertEquals(19, sf.getBegin());
118     assertEquals(20, sf.getEnd());
119     sf = features.get(2);
120     assertEquals("chain", sf.getType());
121     assertEquals("Granzyme B", sf.getDescription());
122     assertEquals(21, sf.getPosition());
123     assertEquals(21, sf.getBegin());
124     assertEquals(247, sf.getEnd());
125
126     sf = features.get(3);
127     assertEquals("sequence variant", sf.getType());
128     assertNull(sf.getDescription());
129     assertEquals(41, sf.getPosition());
130     assertEquals(41, sf.getBegin());
131     assertEquals(41, sf.getEnd());
132
133     sf = features.get(4);
134     assertEquals("sequence variant", sf.getType());
135     assertEquals("Pathogenic", sf.getDescription());
136     assertEquals(41, sf.getPosition());
137     assertEquals(41, sf.getBegin());
138     assertEquals(41, sf.getEnd());
139
140     sf = features.get(5);
141     assertEquals("sequence variant", sf.getType());
142     assertEquals("Pathogenic", sf.getDescription());
143     assertEquals(41, sf.getPosition());
144     assertEquals(41, sf.getBegin());
145     assertEquals(41, sf.getEnd());
146
147     sf = features.get(6);
148     assertEquals("sequence variant", sf.getType());
149     assertEquals("Foo",
150             sf.getDescription());
151     assertEquals(42, sf.getPosition());
152     assertEquals(42, sf.getBegin());
153     assertEquals(42, sf.getEnd());
154     Assert.assertEquals(Uniprot.getDescription(sf),
155             "p.Met42Leu<br/>p.Met42LeuMetVal Foo");
156     /*
157      * Check cross-references
158      */
159     Vector<PDBEntry> xrefs = entry.getDbReference();
160     assertEquals(3, xrefs.size());
161
162     PDBEntry xref = xrefs.get(0);
163     assertEquals("2FSQ", xref.getId());
164     assertEquals("PDB", xref.getType());
165     assertEquals("X-ray", xref.getProperty("method"));
166     assertEquals("1.40", xref.getProperty("resolution"));
167
168     xref = xrefs.get(1);
169     assertEquals("2FSR", xref.getId());
170     assertEquals("PDBsum", xref.getType());
171     assertFalse(xref.getProperties().hasMoreElements());
172
173     xref = xrefs.get(2);
174     assertEquals("AE007869", xref.getId());
175     assertEquals("EMBL", xref.getType());
176     assertEquals("AAK85932.1", xref.getProperty("protein sequence ID"));
177     assertEquals("Genomic_DNA", xref.getProperty("molecule type"));
178   }
179
180   @Test(groups = { "Functional" })
181   public void testGetUniprotSequence()
182   {
183     UniprotEntry entry = new Uniprot().getUniprotEntries(
184             new StringReader(UNIPROT_XML)).get(0);
185     SequenceI seq = new Uniprot().uniprotEntryToSequenceI(entry);
186     assertNotNull(seq);
187     assertEquals(6, seq.getDBRefs().length); // 2*Uniprot, PDB, PDBsum, 2*EMBL
188
189   }
190
191   /**
192    * Test the method that formats the sequence id
193    */
194   @Test(groups = { "Functional" })
195   public void testGetUniprotEntryId()
196   {
197     UniprotEntry entry = new Uniprot().getUniprotEntries(
198             new StringReader(UNIPROT_XML)).get(0);
199
200     /*
201      * name formatted with Uniprot Entry name
202      */
203     String expectedName = "A9CKP4_AGRT5|A9CKP4_AGRT6";
204     assertEquals(expectedName,
205             Uniprot.getUniprotEntryId(entry));
206   }
207
208   /**
209    * Test the method that formats the sequence description
210    */
211   @Test(groups = { "Functional" })
212   public void testGetUniprotEntryDescription()
213   {
214     UniprotEntry entry = new Uniprot().getUniprotEntries(
215             new StringReader(UNIPROT_XML)).get(0);
216
217     /*
218      * recommended names concatenated with space separator
219      */
220     String expectedDescription = "Mitogen-activated protein kinase 13 Henry";
221     assertEquals(expectedDescription,
222             Uniprot.getUniprotEntryDescription(entry));
223   }
224
225   @Test(groups = { "Functional" })
226   public void testGetDescription()
227   {
228     UniprotFeature uf = new UniprotFeature();
229     assertEquals("", Uniprot.getDescription(uf));
230
231     uf.setDescription("Hello");
232     assertEquals("Hello", Uniprot.getDescription(uf));
233
234     uf.setPosition(23);
235     uf.setOriginal("K");
236     Vector<String> vars = new Vector();
237     vars.add("y");
238     uf.setVariation(vars);
239     assertEquals("p.Lys23Tyr Hello", Uniprot.getDescription(uf));
240
241     vars.clear();
242     vars.add("z"); // unknown variant - fails gracefully
243     uf.setVariation(vars);
244     assertEquals("p.Lys23z Hello", Uniprot.getDescription(uf));
245
246     uf.setVariation(null); // variant missing - is ignored
247     assertEquals("Hello", Uniprot.getDescription(uf));
248   }
249 }