*PCv* = *In Prot*/(amount of proteins submitted)

A high PCv value (e.g., 80%) indicates that a pattern containing a certain type of residues is found in many proteins (e.g., 80% of the proteins analyzed). It is worth noting that the sites composing a 3D pattern found do not necessarily exhibit the same structural topology in all the proteins in which such a pattern occurs.


The second level of results appears in the exploration of a particular 3D-pattern. Here, all clusters identified for the selected 3D-pattern are shown as a dynamic data table, where the following features are available:


*CCv* = *In Prot*/(amount of proteins into a particular cluster)

A high *CCv* value (e.g., 80%) indicates that a pattern with the same structural topology is present in most of the proteins (80%) of the corresponding cluster.


The last level of results is displayed selecting a particular cluster. Here, all the sites grouped into a specific cluster are shown as a dynamic data table and the following features are available:


#### **4. Conclusions**

In this work, we present 3D-PP, a new free access web server for discovering and recognition of all similar 3D amino acid patterns among a set of protein structures. Our software has three main features that confer competitive advantages as compared with other similar computational tools: **(a)** 3D-PP does not require previous structural knowledge about ligand(s), motif(s) or binding site(s); **(b)** 3D-PP utilizes a scalable, high-performance graph database; **(c)** 3D-PP can be used with protein structures from both experimental biophysics techniques (X-ray crystallography, NMR, etc.) and in silico homology modeling. Also, the results are shown as simple and intuitively dynamic lists of sequence/structural patterns that can be further analyzed within the web page.

We performed three representative types of uses of 3D-PP. **(I)** In the first case, using a set of protein structures containing the small 3D-pattern knows as Zinc finger, our software was able to detect almost all (98%) Zinc finger C3H1-type contained in the PROSITE database and described in crystal structures. Also, 3D-PP unveiled several new sites that have similar structural features to the known sites but which neither have a Zinc ion in the original structure nor a match between the sequence of these sites and the established sequence pattern for this type of motif. Thus, our results indicate that 3D-PP discovered new putative Zinc ion binding sites. As discussed in the Results, some of these new identified sites might serve to enhance the robustness of a crucial biological structure-derived function, by keeping the coordination of the Zinc ion in cases in which, for example, a punctual specific mutation might occur; **(II)** In the second case, we discovered some conserved 3D-patterns in the serotonin target proteins. This finding is significative considering that these proteins (5-HTRs, SERT and MAO-A) have distinct functions, different sequences and diverse structural folding; **(III)** In the third case, we found some conserved 3D-patterns in a set of protein structures coming from the in silico homology models methodologies. Considering that the X-ray structures solved until March 2019 reach a coverage of nearly 50% of the human proteome, the use of homology models substantially improves the scope of these kinds of structure-based methods. In this case for example, our criteria of selection was as ample as "Proteins overexpressed in different types of cancer',' which indicates the versatility of 3D-PP.

It is important to mention at least two limitations of 3D-PP. First, it should be noted that to identify two (or more) 3D-patterns as conserved, 3D-PP considers only sites that contain the same components (amino acid residues). It is known that, for instance, some promiscuous drugs/ligands can interact with more than one target even if the corresponding binding sites are not composed of identical amino acids but of residues with similar properties (e.g., hydrophobicity, acid or basic character, aromatic character, etc.). Therefore, 3D-patterns with "similar" structural and functional properties, but with a different composition, will not be detected by 3D-PP. The other limitation is that 3D-PP gives no information about the accessibility/drugability of the conserved 3D-patterns identified. Therefore, if a 3D-pattern is either embedded into the protein structure or in a relatively inaccessible location, it could be unproductive to try to develop compounds aimed to act at that site. Beyond these limitations, and considering as a basic idea that protein structure is more conserved than sequence, 3D-PP appears to be a flexible and user-friendly tool for identifying conserved structural motifs, which could be relevant to improve our knowledge of protein function or classification.

**Supplementary Materials:** Supplementary materials can be found at http://www.mdpi.com/1422-0067/20/13/ 3174/s1.

**Author Contributions:** Conceptualization, G.N.-V. and M.R.-P.; Methodology, J.-L.L.-P. and A.V.-J; Software, A.V.-J. and G.N.-V.; Validation, A.V.-J. and G.N.-V.; Resources, J.-L.L.-P. and M.R.-P.; Writing–original draft preparation, G.N.-V.; Writing–review and editing, G.N.-V. and M.R.-P.

**Funding:** This research was funded by the Fondo Nacional de Desarrollo Científico y Tecnológico (FONDECYT) grant number 1170662 (M.R.-P. and G.N.-V.).

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
