Skip to content

Commit

Permalink
Merge pull request #5891 from pughlab/multiple-genome-build-5652
Browse files Browse the repository at this point in the history 
support multiple reference genomes 2nd try
  • Loading branch information
@inodb
inodb committed Jul 17, 2019
2 parents de44419 + a46a54a commit 484a7b9
Show file tree
Hide file tree
Showing 113 changed files with 2,886 additions and 514 deletions.
13 changes: 4 additions & 9 deletions business/src/main/resources/org/mskcc/cbio/portal/persistence/GenePanelMapperLegacy.xml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -69,7 +69,6 @@
<result property="entrezGeneId" column="ENTREZ_GENE_ID"/>
<result property="hugoGeneSymbol" column="HUGO_GENE_SYMBOL"/>
<result property="type" column="TYPE"/>
<result property="cytoband" column="CYTOBAND"/>
</resultMap>

<resultMap id="geneticProfileResultMap" type="org.cbioportal.model.MolecularProfile">
Expand Down Expand Up @@ -128,8 +127,7 @@
gene_panel.DESCRIPTION as description,
gene.ENTREZ_GENE_ID as entrezGeneId,
gene.HUGO_GENE_SYMBOL as hugoGeneSymbol,
gene.TYPE as type,
gene.CYTOBAND as cytoband
gene.TYPE as type
from gene_panel
left join gene_panel_list on gene_panel.INTERNAL_ID = gene_panel_list.INTERNAL_ID
left join gene on gene_panel_list.GENE_ID = gene.ENTREZ_GENE_ID
Expand All @@ -149,8 +147,7 @@
select
gene.ENTREZ_GENE_ID,
gene.HUGO_GENE_SYMBOL,
gene.TYPE,
gene.CYTOBAND
gene.TYPE
from gene
<where>
gene.HUGO_GENE_SYMBOL = #{symbol}
Expand All @@ -162,8 +159,7 @@
select
gene.ENTREZ_GENE_ID,
gene.HUGO_GENE_SYMBOL,
gene.TYPE,
gene.CYTOBAND
gene.TYPE
from gene
<where>
gene.ENTREZ_GENE_ID = #{geneId}
Expand All @@ -174,8 +170,7 @@
select
gene.ENTREZ_GENE_ID,
gene.HUGO_GENE_SYMBOL,
gene.TYPE,
gene.CYTOBAND
gene.TYPE
from gene_alias
inner join gene on gene_alias.ENTREZ_GENE_ID = gene.ENTREZ_GENE_ID
<where>
Expand Down
2 changes: 1 addition & 1 deletion business/src/main/resources/org/mskcc/cbio/portal/persistence/StudyMapperLegacy.xml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -14,7 +14,7 @@
DESCRIPTION as description,
PMID as pmid,
CITATION as citation,
GROUPS as groups,
`GROUPS` as `groups`,
CANCER_STUDY_ID as internal_id
from cancer_study
where CANCER_STUDY_IDENTIFIER in <foreach item="item" collection="study_ids" open="(" separator="," close=")">#{item}</foreach>
Expand Down
47 changes: 30 additions & 17 deletions core/src/main/java/org/mskcc/cbio/portal/dao/DaoCancerStudy.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -95,7 +95,7 @@ private static synchronized void reCache() {
CancerStudy cancerStudy = extractCancerStudy(rs);
cacheCancerStudy(cancerStudy, new java.util.Date());
}
} catch (SQLException e) {
} catch (SQLException | DaoException e) {
e.printStackTrace();
} finally {
JdbcUtil.closeAll(DaoCancerStudy.class, con, pstmt, rs);
Expand Down Expand Up @@ -298,7 +298,7 @@ public static void addCancerStudy(CancerStudy cancerStudy, boolean overwrite) th
pstmt = con.prepareStatement("INSERT INTO cancer_study " +
"( `CANCER_STUDY_IDENTIFIER`, `NAME`, "
+ "`DESCRIPTION`, `PUBLIC`, `TYPE_OF_CANCER_ID`, "
+ "`PMID`, `CITATION`, `GROUPS`, `SHORT_NAME`, `STATUS`, `IMPORT_DATE` ) VALUES (?,?,?,?,?,?,?,?,?,?,?)",
+ "`PMID`, `CITATION`, `GROUPS`, `SHORT_NAME`, `STATUS`, `IMPORT_DATE`,`REFERENCE_GENOME_ID` ) VALUES (?,?,?,?,?,?,?,?,?,?,?,?)",
Statement.RETURN_GENERATED_KEYS);
pstmt.setString(1, stableId);
pstmt.setString(2, cancerStudy.getName());
Expand All @@ -319,6 +319,13 @@ public static void addCancerStudy(CancerStudy cancerStudy, boolean overwrite) th
//TODO - use this field in parts of the system that build up the list of studies to display in home page:
pstmt.setInt(10, Status.UNAVAILABLE.ordinal());
pstmt.setDate(11, java.sql.Date.valueOf(LocalDate.now()));
try {
ReferenceGenome referenceGenome = DaoReferenceGenome.getReferenceGenomeByGenomeName(cancerStudy.getReferenceGenome());
pstmt.setInt(12, referenceGenome.getReferenceGenomeId());
}
catch (NullPointerException e) {
throw new DaoException("Unsupported reference genome");
}
pstmt.executeUpdate();
rs = pstmt.getGeneratedKeys();
if (rs.next()) {
Expand Down Expand Up @@ -356,7 +363,7 @@ public static void addCancerStudyTags(CancerStudyTags cancerStudyTags) throws Da
/**
* Return the cancerStudy identified by the internal cancer study ID, if it exists.
*
* @param cancerStudyID Internal (int) Cancer Study ID.
* @param internalId Internal (int) Cancer Study ID.
* @return Cancer Study Object, or null if there's no such study.
*/
public static CancerStudy getCancerStudyByInternalId(int internalId) throws DaoException {
Expand All @@ -366,7 +373,7 @@ public static CancerStudy getCancerStudyByInternalId(int internalId) throws DaoE
/**
* Returns the cancerStudy identified by the stable identifier, if it exists.
*
* @param cancerStudyStableId Cancer Study Stable ID.
* @param stableId Cancer Study Stable ID.
* @return the CancerStudy, or null if there's no such study.
*/
public static CancerStudy getCancerStudyByStableId(String stableId) throws DaoException {
Expand Down Expand Up @@ -612,19 +619,25 @@ public static void purgeUnreferencedRecordsAfterDeletionOfStudy() throws DaoExce
/**
* Extracts Cancer Study JDBC Results.
*/
private static CancerStudy extractCancerStudy(ResultSet rs) throws SQLException {
CancerStudy cancerStudy = new CancerStudy(rs.getString("NAME"),
rs.getString("DESCRIPTION"),
rs.getString("CANCER_STUDY_IDENTIFIER"),
rs.getString("TYPE_OF_CANCER_ID"),
rs.getBoolean("PUBLIC"));
cancerStudy.setPmid(rs.getString("PMID"));
cancerStudy.setCitation(rs.getString("CITATION"));
cancerStudy.setGroupsInUpperCase(rs.getString("GROUPS"));
cancerStudy.setShortName(rs.getString("SHORT_NAME"));
cancerStudy.setInternalId(rs.getInt("CANCER_STUDY_ID"));
cancerStudy.setImportDate(rs.getDate("IMPORT_DATE"));
return cancerStudy;
private static CancerStudy extractCancerStudy(ResultSet rs) throws DaoException {
try {
CancerStudy cancerStudy = new CancerStudy(rs.getString("NAME"),
rs.getString("DESCRIPTION"),
rs.getString("CANCER_STUDY_IDENTIFIER"),
rs.getString("TYPE_OF_CANCER_ID"),
rs.getBoolean("PUBLIC"));
cancerStudy.setPmid(rs.getString("PMID"));
cancerStudy.setCitation(rs.getString("CITATION"));
cancerStudy.setGroupsInUpperCase(rs.getString("GROUPS"));
cancerStudy.setShortName(rs.getString("SHORT_NAME"));
cancerStudy.setInternalId(rs.getInt("CANCER_STUDY_ID"));
cancerStudy.setImportDate(rs.getDate("IMPORT_DATE"));
cancerStudy.setReferenceGenome(DaoReferenceGenome.getReferenceGenomeByInternalId(
rs.getInt("REFERENCE_GENOME_ID")).getGenomeName());
return cancerStudy;
} catch (SQLException e) {
throw new DaoException(e);
}
}

private static boolean studyNeedsRecaching(String stableId, Integer ... internalId) {
Expand Down
12 changes: 4 additions & 8 deletions core/src/main/java/org/mskcc/cbio/portal/dao/DaoGene.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -91,11 +91,10 @@ public static int updateGene(CanonicalGene gene) throws DaoException {
int rows = 0;
con = JdbcUtil.getDbConnection(DaoGene.class);
pstmt = con.prepareStatement
("UPDATE gene SET `HUGO_GENE_SYMBOL`=?, `TYPE`=?,`CYTOBAND`=? WHERE `ENTREZ_GENE_ID`=?");
("UPDATE gene SET `HUGO_GENE_SYMBOL`=?, `TYPE`=? WHERE `ENTREZ_GENE_ID`=?");
pstmt.setString(1, gene.getHugoGeneSymbolAllCaps());
pstmt.setString(2, gene.getType());
pstmt.setString(3, gene.getCytoband());
pstmt.setLong(4, gene.getEntrezGeneId());
pstmt.setLong(3, gene.getEntrezGeneId());
rows += pstmt.executeUpdate();
if (rows != 1) {
ProgressMonitor.logWarning("No change for " + gene.getEntrezGeneId() + " " + gene.getHugoGeneSymbolAllCaps() + "? Code " + rows);
Expand Down Expand Up @@ -143,13 +142,12 @@ public static int addOrUpdateGene(CanonicalGene gene) throws DaoException {
//add gene, referring to this genetic entity
con = JdbcUtil.getDbConnection(DaoGene.class);
pstmt = con.prepareStatement
("INSERT INTO gene (`GENETIC_ENTITY_ID`, `ENTREZ_GENE_ID`,`HUGO_GENE_SYMBOL`,`TYPE`,`CYTOBAND`) "
+ "VALUES (?,?,?,?,?)");
("INSERT INTO gene (`GENETIC_ENTITY_ID`, `ENTREZ_GENE_ID`,`HUGO_GENE_SYMBOL`,`TYPE`) "
+ "VALUES (?,?,?,?)");
pstmt.setInt(1, geneticEntityId);
pstmt.setLong(2, gene.getEntrezGeneId());
pstmt.setString(3, gene.getHugoGeneSymbolAllCaps());
pstmt.setString(4, gene.getType());
pstmt.setString(5, gene.getCytoband());
rows += pstmt.executeUpdate();

} else {
Expand Down Expand Up @@ -326,7 +324,6 @@ public static ArrayList<CanonicalGene> getAllGenes() throws DaoException {
Set<String> aliases = mapAliases.get(entrezGeneId);
CanonicalGene gene = new CanonicalGene(geneticEntityId, entrezGeneId,
rs.getString("HUGO_GENE_SYMBOL"), aliases);
gene.setCytoband(rs.getString("CYTOBAND"));
gene.setType(rs.getString("TYPE"));
geneList.add(gene);
}
Expand Down Expand Up @@ -374,7 +371,6 @@ private static CanonicalGene extractGene(ResultSet rs) throws SQLException, DaoE
Set<String> aliases = getAliases(entrezGeneId);
CanonicalGene gene = new CanonicalGene(geneticEntityId, entrezGeneId,
rs.getString("HUGO_GENE_SYMBOL"), aliases);
gene.setCytoband(rs.getString("CYTOBAND"));
gene.setType(rs.getString("TYPE"));

return gene;
Expand Down
30 changes: 7 additions & 23 deletions core/src/main/java/org/mskcc/cbio/portal/dao/DaoGeneOptimized.java
View file
Original file line number Diff line number Diff line change
Expand Up @@ -285,7 +285,7 @@ public CanonicalGene getGene(long entrezId) {
/**
* Gets Gene By Entrez Gene ID.
*
* @param entrezId Entrez Gene ID.
* @param geneticEntityId Gene Entity ID.
* @return Canonical Gene Object.
*/
public CanonicalGene getGeneByEntityId(int geneticEntityId) {
Expand Down Expand Up @@ -335,23 +335,8 @@ public List<CanonicalGene> guessGene(String geneId, String chr) {

List<CanonicalGene> genes = geneAliasMap.get(geneId.toUpperCase());
if (genes!=null) {
if (chr==null) {
return Collections.unmodifiableList(genes);
}

String nchr = normalizeChr(chr);

List<CanonicalGene> ret = new ArrayList<CanonicalGene>();
for (CanonicalGene cg : genes) {
String gchr = getChrFromCytoband(cg.getCytoband());
if (gchr==null // TODO: should we exlude this?
|| gchr.equals(nchr)) {
ret.add(cg);
}
}
return ret;
return Collections.unmodifiableList(genes);
}

return Collections.emptyList();
}

Expand Down Expand Up @@ -407,7 +392,7 @@ private static String getChrFromCytoband(String cytoband) {
* @return a gene that can be non-ambiguously determined, or null if cannot.
*/
public CanonicalGene getNonAmbiguousGene(String geneId) {
return getNonAmbiguousGene(geneId, null);
return getNonAmbiguousGene(geneId, true);
}

/**
Expand All @@ -417,19 +402,18 @@ public CanonicalGene getNonAmbiguousGene(String geneId) {
* @return a gene that can be non-ambiguously determined, or null if cannot.
*/
public CanonicalGene getNonAmbiguousGene(String geneId, String chr) {
return getNonAmbiguousGene(geneId, chr, true);
return getNonAmbiguousGene(geneId, true);
}

/**
* Look for gene that can be non-ambiguously determined
* @param geneId an Entrez Gene ID or HUGO symbol or gene alias
* @param chr chromosome
* @param issueWarning if true and gene is not ambiguous,
* @param issueWarning if true and gene is not ambiguous,
* print all the Entrez Ids corresponding to the geneId provided
* @return a gene that can be non-ambiguously determined, or null if cannot.
*/
public CanonicalGene getNonAmbiguousGene(String geneId, String chr, boolean issueWarning) {
List<CanonicalGene> genes = guessGene(geneId, chr);
public CanonicalGene getNonAmbiguousGene(String geneId, boolean issueWarning) {
List<CanonicalGene> genes = guessGene(geneId);
if (genes.isEmpty()) {
return null;
}
Expand Down

0 comments on commit 484a7b9

Please sign in to comment.