Archive for the ChemSpider Searches Category

A small but very useful feature has been added for our MS user base. When Mass Spectrometrists acquire data the masses may carry a positive charge or a negative charge. They may actually be up two proton masses or down a proton mass. For certain species the masses may be up an electron mass or down an electron mass. Since so many measurements today are made as high resolution monoisotopic masses then the addition or subtraction of an electron mass can certainly have an impact on the results of the search.

Previously we have expected our users to add or subtract the masses of the protons or electrons prior to performing their searches. HOW UNFRIENDLY! So, we’ve introduced the ability to simply declare the appropriate charges during the search. Checkout the screen below…it should be self-explanatory. The search is available under the Advanced tab on the search page and by selecting the “Search by Properties” tab.

Charges

There are almost 17 million chemical structures indexed on the ChemSpider database now. The database is built up from data supplied by a long list of contributors (it does take a while to generate this page from the structures so give it time). As a user of ChemSpider you might be interested in searching one or more of the data sources from the overall database. This is available via an Advanced Search as shown below.


Advanced Search screen

Simply click on the data sources check box and select the list of databases that you would like to search. For example, if you are interested in Metabolomics you might want to search only the HMDB and Drugbank databases. So, simply select those databases to perform the search.

Data Sources Search




























Shortly we will be segregating the data sources so that users can select searches of PubChem, available chemicals, natural products, screening databases, spectroscopy databases, physicochemical parameter databases and other groupings. We will also be providing a deeper annotation of each of the databases so that it is clear what each database is providing. Watch this space.

At the request of a user to allow the searching of multiple strings, either names, numbers or both, we have enhanced the standard search box to allow searching of multiple terms. This applies to the webpage search box or that available as a browser add-in.

For example, imagine someone commented that drinking double espressos is good for asthmatics. You might question why? A quick search on google would tell you that caffeine and theophylline are in the same chemical group. However, typing in “caffeine theophylline” will give you a series of hits allowing you to compare on screen as shown below. Some other example searches that will work are combinations such as : 2425-79-8 NaCl viagra

This is work in progress and exposed for testing and feedback. Please stress test – constructive feedback preferred.

Caffeine Theophylline

At beta release we were still experimenting with indexing of the database amd visualization of the hit list. As a result we limited the number of hits that would be returned for review to only 50. Clearly there are MANY occasions where there are more than 50 hits. We’ve received requests from people using mass searching specifically to not constrain the number of hits to 50. For example, an advanced search on a monisotopic mass of 700+/-1 gives over 5000 hits. A search on the substring “furan” provides over 19000 hits. There are many other variants where 50 hits is an inappropriate limitation. Now the thousands of hits can be viewed albeit in multiple pages with scrolling.

What we are still lacking is a Progress Bar for such complex searches. The time bound on a complex search has been set to 180 seconds. Some of the searches may take this long so just wait after executing. If we cannot succeed within 3 minutes we will abort. Otherwise, expect your answers to show up. Please be sensible with your searches. Text searches of 10 million synonyms can be time consuming. For example, while you COULD do a search on “methyl” as a substring why would you? Do a real search…one you care about….it’s a real test of the system. Structure searches are generally very quick. They will get quicker – we have ordered a new server and will be putting it online in the coming week. Let us know your comments, failures and expectations…remember, we are still beta.