1 |
duarte |
123 |
package proteinstructure; |
2 |
duarte |
134 |
import java.io.BufferedReader; |
3 |
|
|
import java.io.File; |
4 |
|
|
import java.io.FileNotFoundException; |
5 |
duarte |
123 |
import java.io.FileOutputStream; |
6 |
duarte |
134 |
import java.io.FileReader; |
7 |
duarte |
123 |
import java.io.PrintStream; |
8 |
|
|
import java.io.IOException; |
9 |
duarte |
135 |
import java.sql.ResultSet; |
10 |
|
|
import java.sql.SQLException; |
11 |
|
|
import java.sql.Statement; |
12 |
duarte |
123 |
import java.util.ArrayList; |
13 |
duarte |
129 |
import java.util.TreeMap; |
14 |
duarte |
144 |
import java.util.regex.Matcher; |
15 |
|
|
import java.util.regex.Pattern; |
16 |
duarte |
135 |
import tools.MySQLConnection; |
17 |
duarte |
123 |
|
18 |
|
|
|
19 |
|
|
public class Graph { |
20 |
|
|
|
21 |
duarte |
135 |
public final static String MYSQLSERVER="white"; |
22 |
|
|
public final static String MYSQLUSER=getUserName(); |
23 |
|
|
public final static String MYSQLPWD="nieve"; |
24 |
duarte |
144 |
|
25 |
|
|
public final static String GRAPHFILEFORMATVERSION = "1.0"; |
26 |
duarte |
135 |
|
27 |
duarte |
123 |
ArrayList<Contact> contacts; |
28 |
duarte |
135 |
// nodes is a TreeMap of residue serials to residue types (3 letter code) |
29 |
duarte |
129 |
TreeMap<Integer,String> nodes; |
30 |
duarte |
145 |
public String sequence; |
31 |
duarte |
143 |
public String accode; |
32 |
|
|
public String chain; |
33 |
duarte |
144 |
public String chaincode=""; // when reading graph from file the field will be filled, otherwise no |
34 |
duarte |
146 |
public double cutoff; |
35 |
|
|
public String ct; |
36 |
duarte |
135 |
boolean directed=false; |
37 |
duarte |
123 |
|
38 |
duarte |
135 |
// these 2 fields only used when reading from db |
39 |
|
|
int graphid=0; |
40 |
|
|
int sm_id=0; |
41 |
|
|
|
42 |
duarte |
134 |
/** |
43 |
|
|
* Constructs Graph object by passing ArrayList with contacts and TreeMap with nodes (res serials and types) |
44 |
|
|
* Must also pass contact type, cutoff, accession code and chain |
45 |
|
|
* @param contacts |
46 |
|
|
* @param nodes |
47 |
|
|
* @param sequence |
48 |
|
|
* @param cutoff |
49 |
|
|
* @param ct |
50 |
|
|
* @param accode |
51 |
|
|
* @param chain |
52 |
|
|
*/ |
53 |
duarte |
129 |
public Graph (ArrayList<Contact> contacts, TreeMap<Integer,String> nodes, String sequence, double cutoff,String ct, String accode, String chain) { |
54 |
duarte |
123 |
this.contacts=contacts; |
55 |
|
|
this.cutoff=cutoff; |
56 |
duarte |
129 |
this.nodes=nodes; |
57 |
|
|
this.sequence=sequence; |
58 |
|
|
this.accode=accode; |
59 |
|
|
this.chain=chain; |
60 |
duarte |
123 |
this.ct=ct; |
61 |
duarte |
129 |
if (ct.contains("/")){ |
62 |
|
|
directed=true; |
63 |
|
|
} |
64 |
duarte |
123 |
} |
65 |
duarte |
135 |
|
66 |
|
|
/** |
67 |
duarte |
141 |
* Constructs Graph object from graph db, given the dbname, accode, chaincode (classic pdb chain code), ct and cutoff |
68 |
duarte |
135 |
* @param dbname |
69 |
|
|
* @param accode |
70 |
duarte |
141 |
* @param chaincode |
71 |
duarte |
135 |
* @param cutoff |
72 |
|
|
* @param ct |
73 |
|
|
*/ |
74 |
duarte |
141 |
public Graph(String dbname, String accode, String chaincode, double cutoff, String ct) throws GraphIdNotFoundError{ |
75 |
duarte |
135 |
this.cutoff=cutoff; |
76 |
|
|
this.accode=accode; |
77 |
|
|
this.ct=ct; |
78 |
duarte |
142 |
// we set the sequence to empty when we read from graph db. We don't have the full sequence in graph db |
79 |
|
|
// when we pass the sequence in getCM to the ContactMap constructor we want to have either a full sequence (with unobserveds) or a blank in case we don't have the info |
80 |
|
|
this.sequence=""; |
81 |
duarte |
135 |
//TODO graphs in db are never directed, so this doesn't really apply here. Must solve all this! |
82 |
|
|
if (ct.contains("/")){ |
83 |
|
|
directed=true; |
84 |
|
|
} |
85 |
|
|
MySQLConnection conn = new MySQLConnection(MYSQLSERVER,MYSQLUSER,MYSQLPWD,dbname); |
86 |
duarte |
141 |
getgraphid(conn, chaincode); // initialises graphid, sm_id and chain |
87 |
|
|
read_graph_from_db(conn); // gets contacts, nodes and sequence |
88 |
duarte |
135 |
conn.close(); |
89 |
|
|
} |
90 |
duarte |
129 |
|
91 |
duarte |
134 |
/** |
92 |
|
|
* Constructs Graph object by reading a file with contacts |
93 |
duarte |
144 |
* If the contacts file doesn't have the sequence then the graph object won't have sequence or nodes |
94 |
|
|
* That means it won't be possible to get a ContactMap from it using getCM because CM needs both sequence and nodes |
95 |
duarte |
134 |
* @param contactsfile |
96 |
|
|
* @throws IOException |
97 |
|
|
* @throws FileNotFoundException |
98 |
|
|
*/ |
99 |
duarte |
144 |
public Graph (String contactsfile) throws IOException, FileNotFoundException{ |
100 |
duarte |
142 |
// we set the sequence to blank when we read from file as we don't have the full sequence |
101 |
duarte |
144 |
// if sequence is present in contactsfile then is read from there |
102 |
duarte |
142 |
this.sequence=""; |
103 |
duarte |
144 |
this.ct=""; |
104 |
|
|
this.cutoff=0.0; |
105 |
duarte |
134 |
if (ct.contains("/")){ |
106 |
|
|
directed=true; |
107 |
|
|
} |
108 |
duarte |
144 |
read_graph_from_file(contactsfile); |
109 |
duarte |
134 |
} |
110 |
|
|
|
111 |
duarte |
135 |
//TODO implement (from python) write_graph_to_db, do we really need it here?? |
112 |
|
|
|
113 |
|
|
/** get user name from operating system (for use as database username) */ |
114 |
|
|
private static String getUserName() { |
115 |
|
|
String user = null; |
116 |
|
|
user = System.getProperty("user.name"); |
117 |
|
|
if(user == null) { |
118 |
|
|
System.err.println("Could not get user name from operating system. Exiting"); |
119 |
|
|
System.exit(1); |
120 |
|
|
} |
121 |
|
|
return user; |
122 |
|
|
} |
123 |
duarte |
144 |
|
124 |
|
|
public void read_graph_from_file (String contactsfile) throws FileNotFoundException, IOException { |
125 |
duarte |
134 |
contacts = new ArrayList<Contact>(); |
126 |
|
|
System.out.println("Reading contacts from file "+contactsfile); |
127 |
|
|
BufferedReader fcont = new BufferedReader(new FileReader(new File(contactsfile))); |
128 |
|
|
String line; |
129 |
|
|
while ((line = fcont.readLine() ) != null ) { |
130 |
duarte |
144 |
Pattern p = Pattern.compile("^#"); |
131 |
|
|
Matcher m = p.matcher(line); |
132 |
|
|
if (m.find()){ |
133 |
|
|
// Pattern ps = Pattern.compile("^#VER: (\\d\\.\\d)"); |
134 |
|
|
// Matcher ms = ps.matcher(line); |
135 |
|
|
// if (ms.find()){ |
136 |
|
|
// if (!ms.group(1).equals(GRAPHFILEFORMATVERSION)){ |
137 |
|
|
// throw new GraphFileFormatError("The graph file "+contactsfile+" can't be read, wrong file format version"); |
138 |
|
|
// } |
139 |
|
|
// } |
140 |
|
|
Pattern ps = Pattern.compile("^#SEQUENCE:\\s*(\\w+)$"); |
141 |
|
|
Matcher ms = ps.matcher(line); |
142 |
|
|
if (ms.find()){ |
143 |
|
|
sequence=ms.group(1); |
144 |
|
|
} |
145 |
|
|
ps = Pattern.compile("^#PDB:\\s*(\\w+)"); |
146 |
|
|
ms = ps.matcher(line); |
147 |
|
|
if (ms.find()){ |
148 |
|
|
accode=ms.group(1); |
149 |
|
|
} |
150 |
|
|
ps = Pattern.compile("^#PDB CHAIN CODE:\\s*(\\w)"); |
151 |
|
|
ms = ps.matcher(line); |
152 |
|
|
if (ms.find()){ |
153 |
|
|
chaincode=ms.group(1); |
154 |
|
|
} |
155 |
|
|
ps = Pattern.compile("^#CHAIN:\\s*(\\w)"); |
156 |
|
|
ms = ps.matcher(line); |
157 |
|
|
if (ms.find()){ |
158 |
|
|
chain=ms.group(1); |
159 |
|
|
} |
160 |
|
|
ps = Pattern.compile("^#CT:\\s*([a-zA-Z/]+)"); |
161 |
|
|
ms = ps.matcher(line); |
162 |
|
|
if (ms.find()){ |
163 |
|
|
ct=ms.group(1); |
164 |
|
|
} |
165 |
|
|
ps = Pattern.compile("^#CUTOFF:\\s*(\\d+\\.\\d+)"); |
166 |
|
|
ms = ps.matcher(line); |
167 |
|
|
if (ms.find()){ |
168 |
|
|
cutoff=Double.parseDouble(ms.group(1)); |
169 |
|
|
} |
170 |
|
|
} |
171 |
|
|
else{ |
172 |
|
|
int i = Integer.parseInt(line.split("\\s+")[0]); |
173 |
|
|
int j = Integer.parseInt(line.split("\\s+")[1]); |
174 |
|
|
contacts.add(new Contact(i,j)); |
175 |
|
|
} |
176 |
duarte |
134 |
} |
177 |
|
|
fcont.close(); |
178 |
duarte |
144 |
nodes = new TreeMap<Integer, String>(); |
179 |
|
|
for (int i=0;i<sequence.length();i++){ |
180 |
|
|
String letter = String.valueOf(sequence.charAt(i)); |
181 |
|
|
nodes.put(i+1, letter); |
182 |
|
|
} |
183 |
|
|
|
184 |
duarte |
134 |
} |
185 |
|
|
|
186 |
duarte |
135 |
/** |
187 |
|
|
* Reads contacts and nodes from db. |
188 |
|
|
* The db must be a graph db following our standard format, i.e. must have tables: |
189 |
|
|
* chain_graph, single_model_graph, single_model_node, single_model_edge |
190 |
|
|
* We don't care here about the origin of the data (msdsd, pdbase, predicted) for the generation of the graph as long as it follows our data format |
191 |
|
|
* We read both edges and nodes from single_model_edge and single_model_node. |
192 |
|
|
* The sequence is taken from nodes, thus it won't have unobserved or non standard aas. |
193 |
|
|
* @param conn |
194 |
|
|
*/ |
195 |
|
|
public void read_graph_from_db(MySQLConnection conn){ |
196 |
|
|
contacts = new ArrayList<Contact>(); |
197 |
|
|
nodes = new TreeMap<Integer, String>(); |
198 |
|
|
try { |
199 |
duarte |
142 |
// we read only half of the matrix (contacts in one direction only) so that we have the same type of contacts as when creating Graph from Pdb object |
200 |
duarte |
135 |
String sql="SELECT i_num,j_num FROM single_model_edge WHERE graph_id="+graphid+" AND j_num>i_num ORDER BY i_num,j_num "; |
201 |
|
|
Statement stmt = conn.createStatement(); |
202 |
|
|
ResultSet rsst = stmt.executeQuery(sql); |
203 |
|
|
while (rsst.next()) { |
204 |
|
|
int i=rsst.getInt(1); |
205 |
|
|
int j=rsst.getInt(2); |
206 |
|
|
contacts.add(new Contact(i,j)); |
207 |
|
|
} |
208 |
|
|
rsst.close(); |
209 |
|
|
stmt.close(); |
210 |
|
|
sql="SELECT num,res FROM single_model_node WHERE graph_id="+graphid+" ORDER BY num "; |
211 |
|
|
stmt = conn.createStatement(); |
212 |
|
|
rsst = stmt.executeQuery(sql); |
213 |
|
|
while (rsst.next()){ |
214 |
|
|
int num=rsst.getInt(1); |
215 |
|
|
String res=rsst.getString(2); |
216 |
|
|
nodes.put(num, AA.oneletter2threeletter(res)); |
217 |
|
|
} |
218 |
|
|
rsst.close(); |
219 |
|
|
stmt.close(); |
220 |
|
|
} catch (SQLException e) { |
221 |
|
|
e.printStackTrace(); |
222 |
|
|
} |
223 |
|
|
|
224 |
|
|
} |
225 |
|
|
|
226 |
duarte |
141 |
public void getgraphid (MySQLConnection conn, String chaincode) throws GraphIdNotFoundError{ |
227 |
|
|
// input is chaincode i.e. pdb chain code |
228 |
|
|
// we take chain (internal chain identifier, pchain_code for msdsd and asym_id for pdbase) from pchain_code field in chain_graph |
229 |
|
|
// (in the chain_graph table the internal chain identifier is called 'pchain_code') |
230 |
duarte |
135 |
int pgraphid=0; |
231 |
duarte |
141 |
String chainstr="='"+chaincode+"' "; |
232 |
|
|
if (chaincode.equals("NULL")){ |
233 |
|
|
chainstr=" IS NULL "; |
234 |
|
|
} |
235 |
duarte |
135 |
try { |
236 |
duarte |
141 |
String sql="SELECT graph_id, pchain_code FROM chain_graph WHERE accession_code='"+accode+"' AND chain_pdb_code"+chainstr+" AND dist="+cutoff; |
237 |
duarte |
135 |
Statement stmt = conn.createStatement(); |
238 |
|
|
ResultSet rsst = stmt.executeQuery(sql); |
239 |
|
|
int check=0; |
240 |
|
|
while (rsst.next()) { |
241 |
|
|
check++; |
242 |
|
|
pgraphid=rsst.getInt(1); |
243 |
duarte |
141 |
chain=rsst.getString(2); |
244 |
duarte |
135 |
} |
245 |
|
|
if (check!=1){ |
246 |
duarte |
141 |
System.err.println("No pgraph_id match or more than 1 match for accession_code="+accode+", chain_pdb_code="+chaincode+", dist="+cutoff); |
247 |
duarte |
135 |
} |
248 |
|
|
rsst.close(); |
249 |
|
|
stmt.close(); |
250 |
|
|
// we set the ctstr to the same as ct except in ALL case, where it is BB+SC+BB/SC |
251 |
|
|
String ctstr=ct; |
252 |
|
|
if (ct.equals("ALL")){ |
253 |
|
|
ctstr="BB+SC+BB/SC"; |
254 |
|
|
} |
255 |
|
|
sql="SELECT graph_id,single_model_id FROM single_model_graph WHERE pgraph_id="+pgraphid+" AND CT='"+ctstr+"' AND dist="+cutoff+" AND CR='(true)' AND CW=1"; |
256 |
|
|
stmt = conn.createStatement(); |
257 |
|
|
rsst = stmt.executeQuery(sql); |
258 |
|
|
check=0; |
259 |
|
|
while (rsst.next()){ |
260 |
|
|
check++; |
261 |
|
|
graphid=rsst.getInt(1); |
262 |
|
|
sm_id=rsst.getInt(2); |
263 |
|
|
} |
264 |
|
|
if (check!=1){ |
265 |
|
|
System.err.println("No graph_id match or more than 1 match for pgraph_id="+pgraphid+", CT="+ctstr+" and cutoff="+cutoff); |
266 |
|
|
throw new GraphIdNotFoundError("No graph_id match or more than 1 match for pgraph_id="+pgraphid+", CT="+ctstr+" and cutoff="+cutoff); |
267 |
|
|
} |
268 |
|
|
} catch (SQLException e) { |
269 |
|
|
e.printStackTrace(); |
270 |
|
|
} |
271 |
|
|
|
272 |
|
|
} |
273 |
|
|
|
274 |
duarte |
123 |
public void write_contacts_to_file (String outfile) throws IOException { |
275 |
|
|
PrintStream Out = new PrintStream(new FileOutputStream(outfile)); |
276 |
|
|
for (Contact pair:contacts){ |
277 |
|
|
int i_resser=pair.i; |
278 |
|
|
int j_resser=pair.j; |
279 |
|
|
Out.println(i_resser+"\t"+j_resser); |
280 |
|
|
} |
281 |
|
|
Out.close(); |
282 |
|
|
} |
283 |
duarte |
144 |
|
284 |
|
|
public void write_graph_to_file (String outfile) throws IOException { |
285 |
|
|
PrintStream Out = new PrintStream(new FileOutputStream(outfile)); |
286 |
|
|
Out.println("#VER: "+GRAPHFILEFORMATVERSION); |
287 |
|
|
Out.println("#SEQUENCE: "+sequence); |
288 |
|
|
Out.println("#PDB: "+accode); |
289 |
|
|
Out.println("#PDB CHAIN CODE: "+chaincode); |
290 |
|
|
Out.println("#CHAIN: "+chain); |
291 |
|
|
Out.println("#CT: "+ct); |
292 |
|
|
Out.println("#CUTOFF: "+cutoff); |
293 |
|
|
for (Contact pair:contacts){ |
294 |
|
|
int i_resser=pair.i; |
295 |
|
|
int j_resser=pair.j; |
296 |
|
|
Out.println(i_resser+"\t"+j_resser); |
297 |
|
|
} |
298 |
|
|
Out.close(); |
299 |
|
|
} |
300 |
|
|
|
301 |
duarte |
129 |
public ContactMap getCM() { |
302 |
|
|
// residues is the map from residue nums to residue types used in ContactMap class, i.e. it is the same as Pdb.resser2restype or Graph.nodes |
303 |
|
|
TreeMap<Integer,String> residues = new TreeMap<Integer,String>(); |
304 |
|
|
// we copy residues from nodes (deep copy) |
305 |
|
|
for (int node:nodes.keySet()){ |
306 |
|
|
residues.put(node, nodes.get(node)); |
307 |
|
|
} |
308 |
|
|
// check if we are in directed or undirected case. If undirected we fill the opposite contacts to pass a full list of contacts to ContactMap (which contains full matrix) |
309 |
|
|
ArrayList<Contact> contacts2pass = new ArrayList<Contact>(); |
310 |
|
|
if (directed){ |
311 |
|
|
contacts2pass=contacts; |
312 |
|
|
} else { |
313 |
|
|
for (Contact cont:contacts){ |
314 |
|
|
int i_resser = cont.i; |
315 |
|
|
int j_resser = cont.j; |
316 |
|
|
contacts2pass.add(new Contact(i_resser,j_resser)); |
317 |
|
|
contacts2pass.add(new Contact(j_resser,i_resser)); |
318 |
|
|
} |
319 |
|
|
} |
320 |
|
|
// construct the ContactMap object and return it |
321 |
|
|
ContactMap cm = new ContactMap(contacts2pass,residues,sequence); |
322 |
|
|
return cm; |
323 |
|
|
|
324 |
|
|
} |
325 |
|
|
|
326 |
duarte |
123 |
} |