This tutorial introduces typical utilizations of NFFinder and some strategies to take advantage of the tool.

We introduce the following applications which are detailed below:

What do I want? Input profile Matching preferences
Input type Databases Search mode
1 Drugs to treat a disease Patient'sa RNA DrugMatrix/CMap Inverse
2 Similar diseases to a particular disease Patient'sa RNA GEO Direct
3 Experts working with similar diseases Patient'sa RNA DrugMatrix/CMap Direct
4 Similar drugs to a tested drug Drug'sb RNA DrugMatrix/CMap Direct
5 Opposite drugs to a tested drug Drug'sb RNA DrugMatrix/CMap Inverse
6 Diseases to be treated for a known drug Drug'sb RNA GEO Inverse
7 Drugs to treat a disease with known miRNA Patient'sc miRNA DrugMatrix/CMap Inverse
8 Similar disease with known miRNAs Patient'sc miRNA GEO Direct

Patient's profile(a): The aim of this study is to identify the gene expression changes associated with an NF1-related tumorgenesis with the comparasion of ST88-14 and HSC using Agilent microarrays. The up and down regulated genes from the diferential analysis are used in NFFinder as the patient’s profile.

Drug's profile(b): In this case the aim is to observe the expression changes in different drug treatments. The data used as a profile are the up- and down-regulated genes of the comparison by microarrays of cells treated with PD901/JQ1 and with DMSO.

Patient's profile(c): The profile used as input are the upregulated miRNAs obtained from the comparison of bening menignomas and arachnoidal tissue controls with miRNA arrays.

(a)Sun,D., Haddad,R., Kraniak,J.M., Horne,S.D. and Tainsky,M. a (2013) RAS/MEK-independent gene expression reveals BMP2-related malignant phenotypes in the Nf1-deficient MPNST. Mol. Cancer Res., 11, 616–27. doi:10.1158/1541-7786.MCR-12-0593 http://www.ncbi.nlm.nih.gov/pubmed/23423222/

(b)De Raedt,T., Beert,E., Pasmant,E., Luscan,A., Brems,H., Ortonne,N., Helin,K., Hornick,J.L., Mautner,V., Kehrer-Sawatzki,H., et al. (2014) PRC2 loss amplifies Ras-driven transcription and confers sensitivity to BRD4-based therapies. Nature, 514, 247–51. doi:10.1038/nature13561 http://www.ncbi.nlm.nih.gov/pubmed/25119042/

(c)Saydam O, Shen Y, Wurdinger T, Senol O, Boke E, James MF,Tannous BA, Stemmer-Rachamimov AO, Yi M, Stephens RM,Fraefel C, Gusella JF, Krichevsky AM, Breakefield XO (2009) Downregulated microRNA-200a in meningiomas promotes tumor growth by reducing E-cadherin and activating the Wnt/betacatenin signaling pathway. Mol Cell Biol 29, 5923–5940. doi: 10.1128/MCB.00332-09 http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2772747/

How to start interpreting your results depending on the databases selected for the analysis.

DrugMatrix/CMap

When we use DrugMatrix and CMap all the experiments in the results are reasonable. This is because a control and a tested drug are compared to obtain the profile. In this case there are 775 drug experiments entries with 391 compounds. A good way to start is selecting only some of the experiments to be further visualized. The criterion to select a smaller list depends on the purpose of the analysis (depending on the score, the number of times that a drug appears in the results, the condition of the experiment, etc.). The table is divided in different columns to an easier visualization and has options to make the filtering possible.

GEO

In this case the interpretation needs some manual curation due to NFFinder compares all possible profiles available in GEO datasets and sometimes these have no sense. The marked experiments in the figure below must be filtered, because the profiles are made from time series (300 min vs 17 min and 6h vs 3h) and dose responses that are not logic for this investigation. NFFinder allows you to filter by condition, but sometimes there are interesting experiments as the one in the red box that you might lost if you filter the results directly.

From these tables some visualizations are generated to represent the results in the different scenarios.

APPLICATION 1: Searching for drugs to treat a particular disease.

1. Fill the input boxes with the up- and down-regulated genes of the patient’s profile.

2. Specify the parameters:

3. Click Search.

Results: http://nffinder.cnb.csic.es/?job=http://nffinder.cnb.csic.es/api/jobs/42517a800833417d32ccc9870323fb035fa565b4/&title=GSE39764_DS4_CMap_DrugMatrix

4. Go to DRUGS related to experiments page.

Trichostatin A appears in more than 75 experiments with scores between 60 and 80.

The DRUGS related to experiments page displays the different drugs that have an opposite profile than the introduced one. As the figure shows, Trichostatin A appears in more than 75 experiments with a medium score, which means that we can consider it a hypothetical drug to revert the phenotype. On the right side of the page there is information to study the conditions of the experiment.

APPLICATION 2: Looking for diseases similar to the input.

Drugs already known to treat a disease might be candidate to treat diseases with similar gene expression profiles.

1. Fill the input boxes with the up- and down-regulated genes of the patient’s profile.

2. Specify the parameters:

3. Click Search.

Results: http://nffinder.cnb.csic.es/?job=http://nffinder.cnb.csic.es/api/jobs/75c93a4ce8283d36a20c3cc3c885ed28995aab0f/&title=GSE39764_DS4_GEO

4. Go to DISEASES related to experiments page

The marked diseases appear several times in the results (10-18 times). You should discern which are interesting for the study. Remark: Given that disease names were obtained from Metamap, we recommend a careful examination.

The DISEASES related to experiments page displays the different diseases that have a similar profile than the introduced one. We select few diseases in which we are interested because of its score and the several times that appear on the results. Next we will examine the data through the table of the right of the page.

APPLICATION 3: Searching for expert authors in similar diseases or biological processes related to our experiment.

Go to the Experts page with the same input as the one used in the previous section.

The Experts page shows up the different authors appearing in the results. The size of the boxes ranks the expert’s list according to the articles in the results. An FQ, Campitello N. and Renne R. appear in two studies despite the fact that Domany E. gets 4 profiles in one study. The table of References links to the research pubmed article. Remark: Several studies have more than one experiment, so different profiles are compared.

APPLICATION 4: Looking for drugs similar to the input.

1. Fill the input boxes with the up- and down-regulated genes of the drug’s profile.

2. Specify the parameters:

3. Click Search.

Results: http://nffinder.cnb.csic.es/?job=http://nffinder.cnb.csic.es/api/jobs/fbe0d99458ce9df5a7683f93ec472bea97e9ce5b/&title=GSE52777

4. Go to Drugs related to experiments page.

Trichostatin A: appears in more than 65 experiments with a score between 60 and 80. This suggests that this drug has a similar profile.

The DRUGS related to experiments page displays the different drugs that have a similar profile. As the figure shows, Trichostatin A appears several times with a score between 60 and 80. In this case it is important to look at the score to analyze how similar they are.

APPLICATION 5: Searching for opposite drugs to a known drug

Another useful application is to search for opposite drugs profiles to the tested drug. By changing the profile matching to inverse and following the previous steps you will be able to visualize this information.

APPLICATION 6: Looking for diseases to be treated with a tested drug.

1. Fill the input boxes with up- and down-regulated genes of the drug’s profile.

2. Specify the parameters:

3. Click Search.

Results: http://nffinder.cnb.csic.es/?job=http://nffinder.cnb.csic.es/api/jobs/774938b458f85b75826efc314b5bede510f028bc/&title=GSE52777%20JQ1_PD901%20GEO

4. Go to Diseases related to experiments page.

These are the diseases with a transcriptional genetic profile opposite from the drug’s profile. It helps us to generate hypothesis about what diseases can be treated with this drug. Here we obtain some diseases with high score which means that are very similar to the introduced profile.

The DISEASES related to experiments page displays the different diseases that have an opposite profile than the drug’s one. The next step is to study if those diseases could be treated with the drug. The table on the right side of the page provides you information about how was carried out the experiment and if it fits for the purpose.

APPLICATION 7: Drugs to treat a disease were known miRNAs are involved

1. Fill the input box with the miRNAs in the patient's profile.

2. Specify the parameters:

3. Click Search.

Results: http://nffinder.cnb.csic.es/?job=http://nffinder.cnb.csic.es/api/jobs/5756ed2b83d4de402f7e36eec132891d4cc4f82d/&title=inverse_drugs_mirna

4. Go to Drugs related to experiments page.

The data is from menignomas, so we are going to start focusing on the selected drugs.

The DRUGS related to experiments page displays the different drugs that have an opposite profile to the targets of the miRNAs introduced. These targets are inferred as a list of down-regulated genes to create the profile. As commented in the previous examples you must study the different results to ensure they match with your expectations.

APPLICATION 8: Searching for diseases with known miRNAs

1. Fill the input box with the miRNAs in the patient's profile.

2. Specify the parameters:

3. Click Search.

Results:http://nffinder.cnb.csic.es/?job=http://nffinder.cnb.csic.es/api/jobs/b3bcdbeee71e8065c5b09052d762c9746e0fe538/&title=

4. Go to Diseases related to experiments page.

As the figure below details, several types of cancer with high score are similar to the profile introduced, so the targets of the miRNAs are upregulated in these diseases.

The Diseases related to experiments page displays those experiments with the same down-regulated genes as the ones that should appear in those experiments that miRNAs are involved. As in other cases, you can find information about the experiment in the table on the right side.