SLIMS Online Help -

Double click on the NINDS CLUC Screen to view the primary screening results of the NINDS CLUC screening data. NINDS is a collection of bioactive compounds obtained from the National Institute of Neurological Disorders and Stroke. The initial screen shows related comments, in this case it is the cell culture and compound seeding methods used to prepare the screen.

Notice the six tabs at the top of the screen.

Tab	Description
comment	This tab contains comments about the experiment. This is a good place to place cell culture descriptions and various assay related information.
plates	This tab shows a view of all the plates used in the dataset. These plates also show the detected systematic error. Any value below 0.05 is considered suspect and should be analyzed by hand. If the dataset has been corrected, this view also shows the plates that have had systematic error removed.
raw	This tab shows the raw data collected from each plate as well as the scored data. This is a global view of the experiment and does not show any replicate statistic data. The control wells are also present in this view.
normalized	If the data set has been loaded with the option to correct systematic error, this tab shows the raw data from each plate after the systematic error has been normalized. Similar to the raw data, this tab does not show replicate data. The control wells are also present in this view.
results	This tab shows the results organized by replicates. For each batch (compound) in the dataset, this view sumarizes all the replicate data including the average and standard deviation of both the raw data and the scored data. For instance, if the protocol was computing percent enhancement, this view shows the statistics for the percent enhancement for all batches (compounds) in the screen.
corrected	This is the same as the results tab except that it computes the score (i.e. percent enhancement) and raw data from the systematic error corrected plates.

Select the results data tab. At this point you may want to make the scren larger to view more data. Note that to the left and right of the data view there are sashes that you can drag to the left and right to increase the size of the viewed data. See the image below. Clicking on any header in the grid will sort the data in that column. The first click will sort descending, the second will sort ascending and the third will revert to the original view. Click on the mean_percent_enhancement tab.

After sorting the data, the view will look like this:

Now the data is sorted by the top compounds. You might want to plot the data as well. Click on the plot button in the commands menu. Select "mean_percent_enhancement" for the Y axis and "batch_id" for the X.

Click ok. You will see the following plot:

Click on a point in the plot will highlight the structure in the view window. The box shown highlighted above is our friend Indoprofen which is one of the top hits from the NINDS dataset. To see if there are any similar compounds to indoprofen, click on Find Similar in the commands menu. This will compute a similarity score based on the compounds fingerprint and rank all compounds relative to the selected structure. A Similarity score of 0 indicates that there is no difference in fingerprint while a score of 1 means that the fingerprints have nothing in common. note: this is normally called a diference score.

Now from the newly created Similarituy view, make a new plot, this time with Similarity Score as the Y axis and mean percent enhancement as the X axis. You will see the following plot.

The highlighted compound at the bottom (at Similarity score 0) is indoprofen. Note that this compound stands alone. There are no compounds clustered near this data point indicating that the structure of indoprofen is highly different than any other structure in the NINDS data set.

Making a SOM
Final, we will make a self-organizing map in order to browse through the dataset. A self organizing map is an analysis technique that is used to place similar compounds together in space. The SOM is a collection of connecting nodes where nodes closer together are generally more similar than nodes farther apart.

To create a SOM, click on the analyze button in the commands pane. You have two options at this point, View Self Organizing map or Create self organizing map. Viewing an existing map will allow you to plate compounds located in a quicklist. This is useful if you have already created a SOM and wish to see where new compounds fall. Click on Create Self Organizing Map. You will see a wizard at this point.

Clicking on Next will show you the current compound set that you are going to use to create the SOM.

Note that the SOM only uses the compounds fingerprint to generate the data. Click on Next to set the creation options.

In general, the default values are appropriate to use. Remember that the size of the map (Map size rows and Map size cols) above, the number of iterations and the number of compounds will influence the time it takes to generate a self organizing map. Once the map has been started, you will not be able to stop the operatation until the generation of the map has completed. You may click on Quick SOM to make the map faster at the expense of quality, although this might be appropriate for browsing.

Enter a name for the SOM such as NINDS CLUC Dataset and click on next. This map will take about five to ten minutes to build on a current computer (1.5gHz to 2.0gHz PC).

When the SOM is complete, click on finish and the SOM will appear (Note, due to the random nature of SOM generation, the map you see will look not look exactly as follows) You can resize the window if it is too small. The map is initially colored by the number of compounds in a node. Red has the most and dark blue has the least. Holding the cursor over a node will indicate how many compounds are in a well.

Clicking on a well will show the compounds in the well, for example this node has compounds with a Dihydroxyflavone motif:

The som can also be recolored based on various data. Click on the SOM menu and then click on recolor. Choose mean_percent_enhancement

Then choose your favorite coloring scheme:

The map will then be recolored using the selected criteria:

Clicking on the light blue hits (purple, indigo...) wil reveal clusters of higher scoring compounds. We immediately find indoprofen using this technique:

Note that, in general, this is not a good way to find singlular hits, but is intended to find clusters of hits with similar structures.

Perform A Medline Analysis

Now, go back to the results tab and click on search and find all compounds that have greater than 60% percent enhancement:

Set the column to "Mean_percent_enhancement" and the expression to ">" then set the value to 60. Click on Search for All. Inspect the newly created Search tab and then click on Analyze and select Search Medline Database.

You will be presented with a list of drug names. These only appear if your selected drugs actually do have names, if they do not, then you cannot search medline. Obviously, if the drug name is novel and does not appear in Medline, you will not find any results either.

The next step is to remove all salts/base and acids identifiers from the drug names as well as any extraneous information. Simply edit the text in the window and click on Search.

You will be presented with the Medline Search result seen below. Each mechanism link takes you to the relevant papers in PubMed. Notice that there is at least on Cyclooxegenase result from the Indoprofen compound. This table is a good way to organize mechanisms with hit compounds. Unfortunately, you need to screen compounds such as the Active Compound Library for maximum effectiveness when using the Medline Searching Capabillity. The index Medline papers only go through 2002, we are in the process of indexing 2003 and will release the index when it is available.

Medline Analysis

**Medline Annotation v 1.0**
mechanism	literature percentile	compound percentile	score	%hits	%drugs	num selected drugs	num Library drugs	num selected papers	num Library papers	selected mean	selected median	selected std	Library mean	Library median	Library std	pos	neg	epos	eneg	drug	drug
ANTIBIOTICS	3.00E-001	3.00E-001	9.00E-002	0.1111	0.0602	1	177	1	194	1.0000	1.0000	NAN	1.5763	1.5763	NAN	1	176	0.5418	176.4582	ACLARUBICIN
ANTIOXIDANTS	3.00E-001	3.00E-001	9.00E-002	0.1111	0.0231	1	73	1	66	1.0000	1.0000	NAN	1.4795	1.4795	NAN	1	67	0.2082	67.7918	ACLARUBICIN
PROTEIN SYNTHESIS	4.00E-001	4.00E-001	1.60E-001	0.1111	0.0143	1	42	1	39	1.0000	1.0000	NAN	1.3095	1.3095	NAN	1	41	0.1286	41.8714	ACLARUBICIN
DNA	5.00E-001	5.00E-001	2.50E-001	0.1111	0.1167	1	359	1	659	1.0000	1.0000	NAN	2.6657	2.6657	NAN	1	342	1.0500	341.9500	ACLARUBICIN
GLYCOSYLATION	5.00E-001	5.00E-001	2.50E-001	0.1111	0.0082	1	25	1	25	1.0000	1.0000	NAN	1.3200	1.3200	NAN	1	23	0.0735	23.9265	ACLARUBICIN
UP	5.00E-001	5.00E-001	2.50E-001	0.1111	0.0898	1	281	1	267	1.0000	1.0000	NAN	1.4270	1.4270	NAN	1	263	0.8082	263.1918	RHAPONTIN
ANTI-INFLAMMATORY AGENTS	5.00E-001	5.00E-001	2.50E-001	0.1111	0.0412	1	129	1	135	1.0000	1.0000	NAN	1.6744	1.6744	NAN	1	120	0.3704	120.6296	INDOPROFEN
PHOSPHATASE	5.00E-001	5.00E-001	2.50E-001	0.1111	0.0231	1	74	1	56	1.0000	1.0000	NAN	1.4324	1.4324	NAN	1	67	0.2082	67.7918	RHAPONTIN
ANTIBIOTIC	5.00E-001	5.00E-001	2.50E-001	0.1111	0.0316	1	93	1	91	1.0000	1.0000	NAN	1.4839	1.4839	NAN	1	92	0.2847	92.7153	ACLARUBICIN
CDK2	5.00E-001	5.00E-001	2.50E-001	0.1111	0.0041	1	12	1	9	1.0000	1.0000	NAN	1.3333	1.3333	NAN	1	11	0.0367	11.9633	RHAPONTIN
JUN	7.00E-001	7.00E-001	4.90E-001	0.1111	0.0932	1	287	1	410	1.0000	1.0000	NAN	1.6481	1.6481	NAN	1	273	0.8388	273.1612	ANISINDIONE
CDK6	1	1	1	0.1111	0.0007	1	2	1	2	1.0000	1.0000	NAN	1.0000	1.0000	NAN	1	1	0.0061	1.9939	RHAPONTIN
ANTICOAGULANTS	1	1	1	0.1111	0.0082	1	26	1	24	1.0000	1.0000	NAN	1.4231	1.4231	NAN	1	23	0.0735	23.9265	ANISINDIONE
CYCLOOXYGENASE	1	1	1	0.2222	0.0221	2	69	2	59	1.0000	1.0000	0	1.4638	1.4638	1.5008	2	63	0.1990	64.8010	INDOPROFEN	RHAPONTIN
CELL CYCLE	1	1	1	0.1111	0.0214	1	63	1	67	1.0000	1.0000	NAN	2.0952	2.0952	NAN	1	62	0.1929	62.8071	RHAPONTIN
TF	1	1	1	0.1111	0.0058	1	18	1	16	1.0000	1.0000	NAN	1.1111	1.1111	NAN	1	16	0.0520	16.9480	ANISINDIONE
ANTICARCINOGENIC AGENTS	1	1	1	0.1111	0.0037	1	11	1	11	1.0000	1.0000	NAN	1.0909	1.0909	NAN	1	10	0.0337	10.9663	RHAPONTIN

Missing Drugs

COLEOFORSIN

CEFOXITIN SODIUM

6,7-DIHYDROXYFLAVONE

LEFUNAMIDE

BIOCHANIN A

That is it for the introduction. Feel free to play with different analyses and colorings.