PubGraph


PubGraph Query
  Basic Query
  Advanced Query
  Choosing Colors
Gallery -- Sample Output
Phenomics
PubAtlas
PubBrain
PhenoWiki
UCLA CNP



PubMed hits for your vocabulary/set of queries


PubMed Results
PubMed Hits Vocabulary Term/PubMed Query Query answered by PubMed (if different)
1048     DRD4 OR "DRD-4" OR "dopamine receptor D4"    
15134     "personality" [tiab] AND ("disorder" [tiab] OR "disorders" [tiab])    
42684     "sleep disorders" [MH]    
59645     dopamine [MH]     "dopamine"[MeSH Terms]
66148     "stress psychological" [MH]    
Currently all terms with more than 100,000 hits are excluded from analysis. Terms with fewer than 3 hits are also excluded.

Term Association Heatmap


click for heatmap (heatmap.pdf)


Term Association Graph


The images below were produced by graphviz for Mac OS X, freely available from both pixelglow and graphviz.org. The graphviz documentation describes the entire family of tools.

Legend:
click for hierarchical graph

Trimmed Term Association Graphs


Simpler graphs are obtained by trimming away all but the highest-magnitude or lowest-magnitude edges:
Trimmed graphs containing only the edges of highest magnitude
trimmed to about 6 n edges [trimmed graph]
trimmed to about 5 n edges [trimmed graph]
trimmed to about 4 n edges [trimmed graph]
trimmed to about 3 n edges [trimmed graph]
trimmed to about 2 n edges [trimmed graph]
trimmed to about 1 n edges [trimmed graph]

Trimmed graphs containing only the edges of lowest magnitude
trimmed to about 6 n edges [trimmed graph]
trimmed to about 5 n edges [trimmed graph]
trimmed to about 4 n edges [trimmed graph]
trimmed to about 3 n edges [trimmed graph]
trimmed to about 2 n edges [trimmed graph]
trimmed to about 1 n edges [trimmed graph]


Legend:
  • node size: number of occurrences
  • edge width: chi-square significance (measure of association nonrandomness)
  • cyan = co-occurrence; cyan edge = the terms co-occur (Jaccard co-occurrence).
  • yellow = insignificance; yellow/green edge = the terms' co-occurrence appears random.
  • magenta = subsumption; violet/blue edge = the less frequent term usually occurs with the other.


You can obtain another trimming of the graph:
keep the highest-weighted edges (with highest absolute value of association)
keep the lowest-weighted edges (with lowest absolute value of association)
percentage of edges to keep (0-100):



Interactive Association Graph Exploration (with GUESS)


click to launch a GUESS applet window (interactive exploration and editing of the association graph)


Co-occurrence statistics


The table of PubMed co-occurrence statistics obtained for this vocabulary is as follows:
  • red: number of abstracts mentioning the row term AND the column term (= intersection).
  • green: number of abstracts mentioning the row term OR the column term (= union).
  • blue: Co-occurrence measure, where e(x) is an abbreviation for exp(x). (This log format is useful for comparing association values that follow a power law.) A popular co-occurrence measure is the Jaccard coefficient = red/green = intersection/union.
Thus the red and green values for diagonal entries are identical. Terms with no abstracts in common are left blank.

By clicking on the numbers in the table you can query PubMed for the corresponding documents.

Click here to browse the table in another window.


Execution log



Sat Jul 25 16:50:50 PDT 2009 6560291/input.vocab

time initiated:
Sat Jul 25 16:50:50 PDT 2009 6560291/pubmed_stats.tab
Sat Jul 25 16:50:50 PDT 2009 6560291/pubmed_stats.csv
Sat Jul 25 16:50:50 PDT 2009 6560291/choices.vocab
Sat Jul 25 16:50:50 PDT 2009 6560291/pubmed_hits.info.gz
Sat Jul 25 16:50:51 PDT 2009 6560291/table.html
Sat Jul 25 16:50:52 PDT 2009 6560291/heatmap.R
Sat Jul 25 16:50:53 PDT 2009 6560291/heatmap.pdf
Sat Jul 25 16:50:54 PDT 2009 6560291/heatmap.gif
Sat Jul 25 16:50:54 PDT 2009 6560291/guess.gdf
Sat Jul 25 16:50:55 PDT 2009 6560291/guess.html
Sat Jul 25 16:50:56 PDT 2009 6560291/graph.T6.dot
Sat Jul 25 16:50:56 PDT 2009 6560291/graph.T6.pdf
Sat Jul 25 16:50:56 PDT 2009 6560291/graph.T6.gif
Sat Jul 25 16:50:57 PDT 2009 6560291/graph.T5.dot
Sat Jul 25 16:50:57 PDT 2009 6560291/graph.T5.pdf
Sat Jul 25 16:50:57 PDT 2009 6560291/graph.T5.gif
Sat Jul 25 16:50:57 PDT 2009 6560291/graph.T4.dot
Sat Jul 25 16:50:57 PDT 2009 6560291/graph.T4.pdf
Sat Jul 25 16:50:57 PDT 2009 6560291/graph.T4.gif
Sat Jul 25 16:50:58 PDT 2009 6560291/graph.T3.dot
Sat Jul 25 16:50:58 PDT 2009 6560291/graph.T3.pdf
Sat Jul 25 16:50:58 PDT 2009 6560291/graph.T3.gif
Sat Jul 25 16:50:58 PDT 2009 6560291/graph.T2.dot
Sat Jul 25 16:50:58 PDT 2009 6560291/graph.T2.pdf
Sat Jul 25 16:50:58 PDT 2009 6560291/graph.T2.gif
Sat Jul 25 16:50:59 PDT 2009 6560291/graph.T1.dot
Sat Jul 25 16:50:59 PDT 2009 6560291/graph.T1.pdf
Sat Jul 25 16:50:59 PDT 2009 6560291/graph.T1.gif
Sat Jul 25 16:50:59 PDT 2009 6560291/graph.B6.dot
Sat Jul 25 16:50:59 PDT 2009 6560291/graph.B6.pdf
Sat Jul 25 16:50:59 PDT 2009 6560291/graph.B6.gif
Sat Jul 25 16:51:00 PDT 2009 6560291/graph.B5.dot
Sat Jul 25 16:51:00 PDT 2009 6560291/graph.B5.pdf
Sat Jul 25 16:51:00 PDT 2009 6560291/graph.B5.gif
Sat Jul 25 16:51:00 PDT 2009 6560291/graph.B4.dot
Sat Jul 25 16:51:00 PDT 2009 6560291/graph.B4.pdf
Sat Jul 25 16:51:01 PDT 2009 6560291/graph.B4.gif
Sat Jul 25 16:51:01 PDT 2009 6560291/graph.B3.dot
Sat Jul 25 16:51:01 PDT 2009 6560291/graph.B3.pdf
Sat Jul 25 16:51:01 PDT 2009 6560291/graph.B3.gif
Sat Jul 25 16:51:01 PDT 2009 6560291/graph.B2.dot
Sat Jul 25 16:51:01 PDT 2009 6560291/graph.B2.pdf
Sat Jul 25 16:51:02 PDT 2009 6560291/graph.B2.gif
Sat Jul 25 16:51:02 PDT 2009 6560291/graph.B1.dot
Sat Jul 25 16:51:02 PDT 2009 6560291/graph.B1.pdf
Sat Jul 25 16:51:02 PDT 2009 6560291/graph.B1.gif
Sat Jul 25 16:51:02 PDT 2009 6560291/graph.dot
Sat Jul 25 16:51:02 PDT 2009 6560291/graph.pdf
Sat Jul 25 16:51:03 PDT 2009 6560291/graph.gif
Sat Jul 25 16:51:03 PDT 2009 done



If you fill in the form below with vocabulary terms and click submit, PubGraph will produce a a graphical view and analysis of associations among these terms in the PubMed/MEDLINE literature database. PubMed is vast, covering millions of publications and thousands of journals.

By specifying a context that limits the analysis done by PubGraph to specific subsets of publications and/or journals, you can control the meaning and strength of associations determined by PubGraph.

PubGraph also provides a number of options for controlling the display of association graphs.

For examples of use of PubGraph, see the Gallery.

Vocabulary Entry

Enter your vocabulary terms (PubMed queries) in the form below.

Note: if you want terms to be treated literally, double-quote them; otherwise PubMed will interpret them with its peculiar semantic methods.     (Explanation: how PubMed interprets queries)

Important PubMed quirk: use UPPERCASE for boolean operators (AND, OR, NOT).

Put each Vocabulary Term on a separate line.

You can also optionally start each line with a Group ID, a single digit 0, 1, ..., 9. (If no initial digit is present, Group ID 0 is used.)

Below you must specify Display of Terms in each Group, letting you control the PubGraph output for each vocabulary term by choosing its Group ID.



If you want additional terms to be loaded into the text box from a text file (with a separate term on each line), choose a filename and click submit below. This form will be redisplayed after including the file contents, giving you the option of making further modifications or file loads before submission for PubGraph analysis.





Display of Terms in each Group

Group ID Color Style Shape
              0              
              1              
              2              
              3              
              4              
              5              
              6              
              7              
              8              
              9              
    (how the colors look)

    Clustering for groups:

    Ordering of groups:

    Layout direction used, if groups are ordered:






PubGraph will generate a graph showing how your Terms co-occur in the PubMed/MEDLINE literature if you click submit (provided you have not selected a file for loading).




Thank you for visiting us.

PubGraph is a project of the UCLA Center for Cognitive Phenomics.
Copyright © 2006-2007 D.S. Parker.