v2.2 Changes

HAShihab · May 10, 2013 · e148151 · e148151
1 parent dfb6b48
commit e148151
Show file tree

Hide file tree

Showing 9 changed files with 96 additions and 49 deletions.
diff --git a/.gitignore b/.gitignore
@@ -4,3 +4,6 @@
 # Server Submission(s)
 *.txt
 *.tab
+
+# Benchmarking
+benchmark/
diff --git a/CHANGES.md b/CHANGES.md
@@ -0,0 +1,14 @@
+fathmm
+======
+
+This page is a continuous listing of new features and bug fixes for the
+Functional Analysis through Hidden Markov Models software and server 
+(http://fathmm.biocompute.org.uk).
+
+## 10th May 2013 (v2.2)
+
+* We now use the Representative Proteome (RP) database when searching for homologous sequences.
+* We now incorporate Pfam 27.0 into our domain-based predictions.
+* Human: added support for ENSEMBL 71, UniProt 2013_05, and RefSeq 59.
+* Cow: added support for ENSEMBL v71
+* We have fixed a minor database bug
diff --git a/README.md b/README.md
@@ -87,9 +87,6 @@ AP : Arabidopsis Plant Ontology
 KW : UniProtKB KeyWords
 ```
 
-## Known Issues:
-
-* Using database v2.0, our predictions appear to be derived/written slower than previous version(s) of our 
-software - this has now been resolved in database v2.1 (missing index).
+## Contributing:
 
 We welcome any comments and/or suggestions that you may have regarding our software and server - please send an email directly to [email protected]
diff --git a/cancer.html b/cancer.html
@@ -83,7 +83,7 @@
                 <h2>Analyze Cancer-Associated Variants</h2>
                 <br />
                 <p>
-                    Obtain instant predictions capable of distinguishing between cancer-promoting/driver mutations, passsenger mutations and other germline polymorphisms.
+                    Obtain instant predictions capable of distinguishing between cancer-promoting/driver mutations, passenger mutations and other germline polymorphisms.
                 </p>
             </div>
 
@@ -242,8 +242,8 @@ <h1>Input Format:</h1>
                                 <h1>Prediction Threshold:</h1>
                                 <br />
                                 <p>
-                                As described in our paper (submitted), our server uses a default prediction threshold of -0.75; however, this can be adjusted and tuned to cater 
-                                for your individual needs.  If you are interested in minimising the number of false positives in your analysis, then you should opt 
+                                As described in our paper, our server uses a default prediction threshold of -0.75; however, this can be adjusted and tuned to cater 
+                                for your individual needs.  For example, if you are interested in minimising the number of false positives in your analysis, then you should opt 
                                 for a more conservative threshold, e.g. -3.0; however, if you are interested in capturing a large proportion of cancer-associated mutations (regardless of 
                                 the number of false positives), then a less stringent threshold should be selected, e.g. 0.0 or higher.  To inform you of this choice, the specificity and sensitivity 
                                 of our software at various prediction thresholds can be seen using the below interactive graph:

diff --git a/cgi-bin/fathmm.py b/cgi-bin/fathmm.py
@@ -19,16 +19,16 @@ def map_position(domain, substitution):
 
     # return if the amino acid substitution falls outside of the domain 
     # assignment 
-    if int(substitution[1:-1]) < int(domain['seq_start']) or \
+    if int(substitution[1:-1]) < int(domain['seq_begin']) or \
        int(substitution[1:-1]) > int(domain['seq_end']):
         return None
 
     # adjust the sequence/HMM position (Python has zero-based coordinates)
-    x = int(domain['seq_start']) - 1 
-    y = int(domain['hmm_start']) - 1 
+    x = int(domain['seq_begin']) - 1 
+    y = int(domain['hmm_begin']) - 1 
 
     # map amino acid substitution onto the HMM
-    for residue in list(domain['alignment']):
+    for residue in list(domain['align']):
         if residue.isupper() or residue.islower():
             # upper-case (match)/lower-case (insertion) characters map 
             # onto a sequence residue
@@ -67,21 +67,21 @@ def fetch_prediction(facade, substitution):
             else:
                 # derive a "Weighted" prediction depending on the selected
                 # pathogenicity weights ...
-                dbCursor.execute("select * from Weights where id='" + x['id'] + "' and type='" + options.weights + "'")
+                dbCursor.execute("select * from WEIGHTS where id='" + x['id'] + "' and type='" + options.weights + "'")
                 Weights   = dbCursor.fetchone()
 
                 if Weights:
                     p     = x[substitution[0]]
                     q     = x[substitution[-1]]
-                    r     = Weights['polymorphic']
-                    s     = Weights['pathogenic']
+                    r     = Weights['other']
+                    s     = Weights['disease']
 
                     Score = "%.2f" % math.log(((1.0 - p) * (r + 1.0)) / ((1.0 - q) * (s + 1.0)), 2)
 
         # derive domain-phenotype associations (via the most informative
         # SUPERFAMILY HMM)
         if not Phenotypes and x['accession']:
-            dbCursor.execute("select * from Phenotypes where accession='" + x['accession'] + "' and ontology='" + options.phenotypes + "' and origin=1 order by score")
+            dbCursor.execute("select * from PHENOTYPES where accession='" + x['accession'] + "' and type='" + options.phenotypes + "' and origin=1 order by score")
             Phenotypes = dbCursor.fetchall()
 
         if Score and Phenotypes:
@@ -104,7 +104,7 @@ def process_record(dbSNP, protein, substitution):
 
 
     # fetch pre-computed sequence record
-    dbCursor.execute("select a.* from Clusters a, Protein b where a.id=b.cluster and b.name='" + protein + "'")
+    dbCursor.execute("select a.* from SEQUENCE a, PROTEIN b where a.id=b.id and b.name='" + protein + "'")
     SeqRecord = dbCursor.fetchone()
 
     if not SeqRecord:
@@ -138,7 +138,7 @@ def process_record(dbSNP, protein, substitution):
     Facade    = []
 
     # fetch substitution-harbouring protein domains
-    dbCursor.execute("select * from Domains where id='" + SeqRecord['id'] + "' and " + substitution[1:-1] + " between seq_start and seq_end order by evalue")
+    dbCursor.execute("select * from DOMAINS where id='" + str(SeqRecord['id']) + "' and " + substitution[1:-1] + " between seq_begin and seq_end order by score")
     DomRecord = dbCursor.fetchall()
 
     for x in DomRecord:
@@ -147,7 +147,7 @@ def process_record(dbSNP, protein, substitution):
 
         if residue:
             # fetch description/probabilities for mapped position
-            dbCursor.execute("select a.*, b.accession, b.description from Probabilities a, Library b where a.id=b.id and a.id='" + str(x['hmm']) + "' and a.position='" + residue + "'")
+            dbCursor.execute("select a.*, b.accession, b.description from PROBABILITIES a, LIBRARY b where a.id=b.id and a.id='" + str(x['hmm']) + "' and a.position='" + residue + "'")
             dbRecord = dbCursor.fetchone()
 
             if dbRecord:
@@ -157,7 +157,7 @@ def process_record(dbSNP, protein, substitution):
     # derive a prediction based on the most informative HMM
     if options.weights.upper() == "UNWEIGHTED":
         # if "Unweighted" prediction", append JackHMMER probabilities/conservation ...
-        dbCursor.execute("select a.*, b.accession, b.description from Probabilities a, Library b where a.id=b.id and a.id='" + SeqRecord['id'] + "' and a.position='" + substitution[1:-1] + "'")
+        dbCursor.execute("select a.*, b.accession, b.description from PROBABILITIES a, LIBRARY b where a.id=b.id and a.id='" + str(SeqRecord['id']) + "' and a.position='" + substitution[1:-1] + "'")
         ConRecord = dbCursor.fetchone()
 
         if ConRecord:
@@ -174,7 +174,7 @@ def process_record(dbSNP, protein, substitution):
 
         if not Score:
             # append JackHMMER probabilities/conservation
-            dbCursor.execute("select a.*, b.accession, b.description from Probabilities a, Library b where a.id=b.id and a.id='" + SeqRecord['id'] + "' and a.position='" + substitution[1:-1] + "'")
+            dbCursor.execute("select a.*, b.accession, b.description from PROBABILITIES a, LIBRARY b where a.id=b.id and a.id='" + str(SeqRecord['id']) + "' and a.position='" + substitution[1:-1] + "'")
             ConRecord = dbCursor.fetchone()
 
             if ConRecord:
@@ -320,7 +320,7 @@ def process_record(dbSNP, protein, substitution):
                 # dbSNP record ...
                 if re.compile("^rs\d+$", re.IGNORECASE).match(record):
                     # ... fetch protein consequence(s)
-                    dbCursor.execute("select distinct * from Variants where id='" + record + "'")
+                    dbCursor.execute("select distinct * from VARIANTS where id='" + record + "'")
                     dbRecords = dbCursor.fetchall()
 
                     if not dbRecords:

diff --git a/cgi-bin/results.cgi b/cgi-bin/results.cgi
@@ -210,9 +210,14 @@ if __name__ == '__main__':
                 <hr>
                 <footer>
                 <p>
-                We welcome any comments and/or suggestions that you may have regarding our software and server - please send an email directly to [email protected]
+                    If you have found this resource useful, please cite the following work:
+                    <br />
+                    <a href="http://www.ncbi.nlm.nih.gov/pubmed/23033316">Shihab HA, Gough J, Cooper DN, Stenson PD, Barker GLA, Edwards KJ, Day INM, Gaunt, TR. (2013). Predicting the Functional, Molecular and Phenotypic Consequences of Amino Acid Substitutions using Hidden Markov Models. Hum. Mutat., <b>34</b>, 57-65 </a>
+                </p>
+                <p>
+                    We welcome any comments and/or suggestions that you may have regarding our software and server - please send an email directly to [email protected]
                 </p>
-                </footer>
+            </footer>
                 
             </div>
         </body>
@@ -323,9 +328,14 @@ if __name__ == '__main__':
               
             <hr>
             <footer>
-            <p>
-            We welcome any comments and/or suggestions that you may have regarding our software and server - please send an email directly to [email protected]
-            </p>
+                <p>
+                    If you have found this resource useful, please cite the following work:
+                    <br />
+                    <a href="http://www.ncbi.nlm.nih.gov/pubmed/23033316">Shihab HA, Gough J, Cooper DN, Stenson PD, Barker GLA, Edwards KJ, Day INM, Gaunt, TR. (2013). Predicting the Functional, Molecular and Phenotypic Consequences of Amino Acid Substitutions using Hidden Markov Models. Hum. Mutat., <b>34</b>, 57-65 </a>
+                </p>
+                <p>
+                    We welcome any comments and/or suggestions that you may have regarding our software and server - please send an email directly to [email protected]
+                </p>
             </footer>
             
         </div>
@@ -424,9 +434,14 @@ if __name__ == '__main__':
             
             <hr>
             <footer>
-            <p>
-            We welcome any comments and/or suggestions that you may have regarding our software and server - please send an email directly to [email protected]
-            </p>
+                <p>
+                    If you have found this resource useful, please cite the following work:
+                    <br />
+                    <a href="http://www.ncbi.nlm.nih.gov/pubmed/23033316">Shihab HA, Gough J, Cooper DN, Stenson PD, Barker GLA, Edwards KJ, Day INM, Gaunt, TR. (2013). Predicting the Functional, Molecular and Phenotypic Consequences of Amino Acid Substitutions using Hidden Markov Models. Hum. Mutat., <b>34</b>, 57-65 </a>
+                </p>
+                <p>
+                    We welcome any comments and/or suggestions that you may have regarding our software and server - please send an email directly to [email protected]
+                </p>
             </footer>
             
         </div>
@@ -549,9 +564,14 @@ if __name__ == '__main__':
             
             <hr>
             <footer>
-            <p>
-            We welcome any comments and/or suggestions that you may have regarding our software and server - please send an email directly to [email protected]
-            </p>
+                <p>
+                    If you have found this resource useful, please cite the following work:
+                    <br />
+                    <a href="http://www.ncbi.nlm.nih.gov/pubmed/23033316">Shihab HA, Gough J, Cooper DN, Stenson PD, Barker GLA, Edwards KJ, Day INM, Gaunt, TR. (2013). Predicting the Functional, Molecular and Phenotypic Consequences of Amino Acid Substitutions using Hidden Markov Models. Hum. Mutat., <b>34</b>, 57-65 </a>
+                </p>
+                <p>
+                    We welcome any comments and/or suggestions that you may have regarding our software and server - please send an email directly to [email protected]
+                </p>
             </footer>
         </div>
         

diff --git a/index.html b/index.html
@@ -76,8 +76,10 @@
                 </div>
             </div>
 
+
+
             <div class="hero-unit">
-                <h2>Functional Analysis through Hidden Markov Models</h2>
+                <h2>Functional Analysis through Hidden Markov Models (v2.2)</h2>
                 <br />
                 <p>
                     A high-throughput web-server capable of predicting the functional, molecular and phenotypic consequences of 
@@ -91,43 +93,49 @@ <h2>Functional Analysis through Hidden Markov Models</h2>
 
             <div id="info" class="modal hide fade" tabindex="-1" role="dialog" aria-labelledby="head" aria-hidden="true">
                 <div class="modal-header">
-                    <h3 id="head">About our Software and Server:</h3>
+                    <h3 id="head">Description:</h3>
                 </div>
                 <div class="modal-body">
                     <p>
                         <br />
                         The Functional Analysis through Hidden Markov Models (fathmm) software and server is a <b>sequence-based algorithm</b> capable of predicting the 
-                        <b>functional and phenotypic consequences of protein missense variants</b> via the amalgamation of sequence conservation within hidden Markov 
-                        models (HMMs), representing the alignment of homologous sequences and conserved protein domains, with <b>'pathogenicity weights'</b>, representing 
+                        <b>functional and phenotypic consequences of protein missense variants</b> by combining sequence conservation within hidden Markov 
+                        models (HMMs), representing the alignment of homologous sequences and conserved protein domains, with <b>"pathogenicity weights"</b>, representing 
                         the tolerance of the corresponding model to mutations.
                         <br /><br />
-                        
+
                         If you have found this resource useful, please cite the following work: 
                         <br /><br />
-                        Shihab HA <i>et al.</i> 2012. Predicting the Functional, Molecular, and Phenotypic Consequences of Amino Acid Substitutions 
-                        using Hidden Markov Models. Human Mutation <a href="http://www.ncbi.nlm.nih.gov/pubmed/23033316"><img src="./img/pubmed.png" /></a>
+
+                        Shihab HA, Gough J, Cooper DN, Stenson PD, Barker GLA, Edwards KJ, Day INM, Gaunt, TR. (2013). Predicting the Functional, Molecular and 
+                        Phenotypic Consequences of Amino Acid Substitutions using Hidden Markov Models. <i>Hum. Mutat.</i>, <b>34</b>, 57-65 
+                        <a href="http://www.ncbi.nlm.nih.gov/pubmed/23033316"><img src="./img/pubmed.png" /></a>
                         <br />
                     </p>
                 </div>
                 <div class="modal-footer">
                     <button class="btn" data-dismiss="modal" aria-hidden="true">Close</button>
                 </div>
             </div>
-
+
+
+
             <div class="row">
                 <div class="span6">
                     <div class="well well-large" style="height: 150px;">
                         <h4>Analyze dbSNP/Protein Missense Variants</h4>
                         <br />
-                        <p>Obtain instant predictions describing the <b>functional, molecular and phenotypic consequences</b> of protein missense variants.</p>
+                        <p>Obtain instant predictions capable of discriminating between disease-causing mutations and neutral polymorphisms - including 
+                        predictions on the molecular and phenotypic consequences of protein missense variants.</p>
                         <p><a href="./inherited.html" class="btn pull-right">go &raquo;</a></p>
                     </div>
                 </div>
                 <div class="span6">
-                    <div class="well well-large" style="height: 150px; background-image:url('./img/new.png'); background-repeat:no-repeat; background-position:top right; ">
+                    <!--<div class="well well-large" style="height: 150px; background-image:url('./img/new.png'); background-repeat:no-repeat; background-position:top right; ">-->
+                    <div class="well well-large" style="height: 150px;">
                         <h4>Analyze Cancer-Associated Variants</h4>
                         <br />
-                        <p>Obtain instant predictions capable of distinguishing between <b>cancer-promoting/driver mutations</b>, <b>passsenger mutations</b> and <b>other germline polymorphisms</b>.</p>
+                        <p>Obtain instant predictions capable of distinguishing between cancer-promoting/driver mutations, passenger mutations and other germline polymorphisms.</p>
                         <p><a href="./cancer.html" class="btn pull-right">go &raquo;</a></p>
                     </div>
                 </div>