Discussion | DivCHED CCCE: Cheminformatics OLCC

Project Discussions

Students (and potential mentors), If you need a project (or have an idea you are willing to work with students on), please go to this page, <a href="http://olcc.ccce.divched.org/content/potential-olcc-project-mentors">http://olcc.ccce.divched.org/content/potential-olcc-project-mentors</a> and subscribe to it. We need to submit our projects to the ACS symposium this week! (Monday at the latest). Cheers, Bob

Non-scripting Alternative Project Follow Up

I put the promised quick and dirty Googe/NIST version of the Model Kit Mini Up. If you've been to the page recently, clear browser cache: <a href="http://chemagic.com/molecules/mini.html">http://chemagic.com/molecules/mini.html</a> Have some fun with it: 1. Click the O atom button and click a methane H to make methanol. 2. Click the NIST Google button and the subsequent yellow link. That's Google filtering the NIST results for the model in the window - filtering for spectra. 3) Click the NIST Direct button and the subsequent yellow link. That's the direct link to the related NIST page for the model in the window. Cute (probably overly cute) title for a paper related to this type of Google use: "Google as Webpage Scraper" Google is the biggest Webpage scraper in the world! If you Google the above title, you'll get a ton of hits related to apps and extensions - not hits about using Google for what it is - a webpage scraper. Best, Otis

Non-scripting Alternative

I can suggest a variation of my original proposal that is definitely less technical, yet it could lead to a useful end product (an informative paper). Over the years there have been a number of articles, papers, and blogs on the subject of Google as a possible serious chemical data tool - Henry Rzepa, Rich Apodaca, Egon Willighagen, and Christopher Southan to name a few. Using the latter contributor’s paper "InChI in the wild: an assessment of InChIKey searching in Google” as a model, it might be interesting for a group of students to reinvestigate this 2013 paper with Google and possibly Wikipedia in mind: <a href="http://www.jcheminf.com/content/5/1/10">http://www.jcheminf.com/content/5/1/10</a> I will put up a quick and dirty version (later today) of the Mini Model Kit that illustrates the technical concept (original proposal). The version will have two buttons - one going directly to NIST and the other going to NIST via Google for specific NIST info. This would just be an example of the more general idea of Google as a possible serious tool, - i.e. the main theme of the suggested non-technical project variation. Things for students to explore: 1) Advanced Google searches in general - not just for NIST images 2) Persistence of Google search results and resulting provenance problems 3) Possible ways of dealing with item 2. 4) The problem of query truncation, particularly re InChI 5) InChI vs InChIKey 6) Other popular search engine responses to InChI and InChIKey queries 7) Advanced search capability of various popular engines I’m sure more points would come up in the study. I'll post a short note here when I have the Mini Model Kit modification running. Best, Otis

Projects with mentors

Hi All, We have created a page where mentors can propose projects students can join. It is the top link of the student projects. If students decide to work with any mentor, we will create a student project page for that project. Herman Bergwerf of MolView and Otis Rothenberger of CheMagic have posted to projects students could potentially work on. The mentor's emails are included when they describe their projects, but students need to remember to clear all projects with their facilitators. Cheers, Bob

Project Creation

To All, I have posted my project on the project list section of the form. You can find my specific project here. <a href="http://olcc.ccce.divched.org/2015OLCCProject1">http://olcc.ccce.divched.org/2015OLCCProject1</a> -Alex Williams

Sun, 10/11/2015 - 00:25

Alex Williams (not verified)

Mac Equivalent

Comment:

Hi Dr. Belford, I wanted to make sure people with macs knew that they already have a great screen recorder. Its actually in quicktime! I have made a few screen casts with it at Centre for explaining how certain chemistry web apps work. Hope the other students find it useful! Here is a link to a tutorial for it. <a href="http://www.makeuseof.com/tag/5-best-screen-recorders-capturing-mac-os-x/">http://www.makeuseof.com/tag/5-best-screen-recorders-capturing-mac-os-x/</a>

Re: Question 2.4.3

Oops - you're correct, part (b) should refer back to part (a). Thanks for catching that typo. And, yes, you've got it - showing two CAS numbers for one stereoisomer is pretty strange. We've discussed elsewhere in the Q&A how this can happen - going through parts (c) and (d) of this question will give you a concrete example of how this can arise in practice.

Chemical Structure Search vs. PubChem Search

Currently, PubChem has two search interfaces: (1) Chemical Structure Search (<a href="https://pubchem.ncbi.nlm.nih.gov/search/search.cgi">https://pubchem.ncbi.nlm.nih.gov/search/search.cgi</a>) (2) PubChem Search (<a href="https://pubchem.ncbi.nlm.nih.gov/search/">https://pubchem.ncbi.nlm.nih.gov/search/</a>) Chemical Structure Search allows users to search PubChem for chemical structures. It accepts as a query a chemical name or chemical structure identifier, which refers to a chemical structure. It has been used for the past 10 year. PubChem Search allows users to search PubChem for chemical structures as well as other things, including patents, assays, assay targets, assay outcomes (bioactivities). Because PubChem Search uses more modern technologies, it is much faster than Chemical Structure Search. It also provides some advanced features like faceted navigation, and hit highlighting. However, PubChem Search is a beta version (a test version). Therefore, this service is subject to change in the near future, and we can’t say this service will exist in its current form. For this reason, I would say that the “Chemical Structure Search”, which is more stable, is more appropriate for this course. Of course, if you want to try “PubChem Search” for test purposes, feel free to do so.

Question 2.4.3

For question 2.4.3, question b, is the question referring to part a)? And if so, all that I could possibly assume would be the (R) stereoisomer having two CAS numbers.

Interpreting chemical graphs

Let's take (c) and (d). Both are labeled graphs: that is, the computer looks at them and seems nodes and edges. We've told the computer that this particular graph is a molecule, and that in a molecule, nodes are atoms and edges are bonds. In (c), the computer is looking at each atom label expecting to see a single element symbol to identify the atom. It doesn't recognize F3C as a single element symbol, so this hypothetical interpreter will not recognize this graph as a molecular graph. In (d), we assume that we've told the molecule that each node has two labels. The first label is either an element symbol or a contracted label (for more on contracted labels, see 4a) for a group. If the first label is an element symbol, the computer is instructed to interpret the node as the corresponding atom. In this case, the second label indicates how many hydrogen atoms are bound to that atom. However, if the first label is a contracted label for a group (from the computer's perspective, anything that's not an element symbol from the periodic table), then the second label is *another graph* for the group represented in human-readable fashion by that contracted label. The point is that when a computer is programmed to recognize a regular structural formula, something of this sort has to be going on. A computer can "read" structural formulas only insofar as it is given instructions sufficient for translating whatever information it's given into a full molecular graph. Does that make sense?