PC204 – Challenge #2

Six Degrees of Separation

You may have heard of the Six Degrees of Kevin Bacon. To better understand the level of collaborative research that occurs at UCSF, Challenge 2 is about the Six Degrees of faculty in some of our Graduate Programs. In this challenge, we define degree of separation as co-authorship on a PubMed publication, and limit the list of people to faculty from just three programs. We also make the (sometimes incorrect) simplifying assumption that a publication author and a faculty member match if they have the same last name and initials.

1. Who has the most faculty members with one degree of separation? (For a faculty member, it does not matter how many papers s/he co-authored, only the number of faculty members with whom s/he has at least one co-authored paper.)
2. Who has the most faculty members with three or fewer degree of separation? (For a faculty member, it does not matter how many different ways s/he is connected to another faculty member, only that they are connected.)

To satisfy the challenge, you will need to turn in a program that uses the faculty list from the links below, queries PubMed for some information, and computes the answers to the questions. There may be several faculty members with the same number of connections; all of them should be listed in your answers.

Useful Links

• Faculty Data Files (Note that some faculty are members of more than one graduate program.)
  • Biophysics
  • Bioinformatics (BMI)
  • Chemistry and Chemical Biology


• PubMed EUtils API
  • ESearch and ESummary are the ones you want
  • ESearch and ESummary results are, by default, in XML format. If you are not familiar with XML processing, you may want to look at our example code
  • ESummary returns at most 10,000 results per request and may be somewhat slow
  • PubMed may get mad at you if you send too many queries too fast
  • You may want to construct smaller test input files while developing your code

Connection Graph

Generating a graph is not part of the challenge, but you may want use the graph below (created using Cytoscape.js) for reference when developing your solution. The vertices represent faculty members and the edges co-authored publications. Hovering your mouse over vertices and edges should bring up tool tips with faculty names and PubMed ids, respectively. Your answers may differ somewhat from the graph, depending on how you collect publications from faculty names, but the connectivity of the graph should be similar. (An alternative rendering of the same connection graph may be made using D3.)

Click on circles to display the name of the faculty member. Click on an edge to display the list of PubMed IDs of papers that two faculty members co-authored. Zoom in with the scroll wheel to display all faculty names. Click-and-drag does different things depending on what is under the cursor at the time; experiment to find out what.