General     remarks
Post-Translational Modifications can be in charge of many biological modulation and therefore could be implicated in various pathologies. They induce significant changes in protein dynamics (Xin F. et Radivojac P., 2012). Many databases (see LINKS) exist to define residues associated with PTM, but none with structural information of the protein. PTM-SD curates these information and provides important structural details on all PTMs found in protein structures. PTMs are more complicated to analyze on protein structures as their annotation are quite versatile. PTM-SD provides an unified view on them. It gives a precise position and correspondance with true protein sequences as the different problematic cases observed It is so useful for scientists interested in one PTM associated to one protein or scientists implicated in the analysis of a particular kind of PTM. Moreover, PTM-SD gives information PTMs in terms of localisation, secondary structures, Protein Blocks and could be used also to define a non-redundant dataset. Publication Please cite: Pierrick Craveur, Joseph Rebehmed, and Alexandre G. de Brevern PTM-SD: a database of structurally resolved and annotated posttranslational modifications in proteins Database 2014: bau041 doi:10.1093/database/bau041 published online May 24, 2014 Data     source We have used different databases to characterize the different PTM resolved in the protein structures (see the FLOWCHART for details): Protein structures are found in Protein Data Bank ( [ref]. Note: Actually only the first model of NMR structures is kept. Sequences contained in the PDB were compared with sequences of the UniProt database ( [ref]. PTMs found in the protein structures were compared with PTMs annotated in dbPTM ( [ref] and PTMCuration ( [ref]. Flowchart
For the last step it was necessary to verify if the PTM annotations correspond to the chemical structure of the modified residue found at the same position. We use automatic and manual verification processes, which are based on the atom information extract from PDB files and a correspondence annotation table.
PTM-SD     Statistics On ... there are ... entries in PTM-SD. Pass the pointer over the pie graph below to get the corresponding number of entries.
Update There are 2 different processes of update: the structures and the annotations update.
The structures data are weekly updated, just after the weekly update of the PDB. Updated data in red. The annotations data are monthly updated. Updated data in red.

Browse     Database
To access records from PTM-SD, the database can be browsed using 2 search modes according to different filters; the Simple and the Advanced Mode. The filters in Simple mode are:
P D B - I D You can specified one or a list of PDB id code separated by semicolon. This filter is not case sensitive (Ex: 1fzc;1FZG;1Lt9).
U n i P r o t - A C You can specified one or a list of UniProt AC separated by semicolon. This filter is not case sensitive (Ex: P00396;P06008;P0C1V1).
O R G A N I S M Select among the list of organism (UniProt species identification codes followed by common name or, whether missing, by official scientific name).
P T M Select among the list of Post Translational Modification annotations.
Press Ctrl and Shift key to deal with individual and multiple selection. For more complex filters use the advanced mode (see below). Search By clicking in the Search button you request Database as follow: Each selected values in same filter category are treat with the "OR" operator. Each filter categorie are treat with the "AND" operator. So for example by filter records like this : It will request records corresponding to the ARAHY (Peanut) organism "AND" corresponding to PTMs N-linked Glycosylation "OR" Hydroxylation. Request traduction: "All records corresponding to N-linked Glycosylation and Hydroxylation in Peanut organism." Advanced     search To pass in Advanced Mode click on the Advanced search button. By using Advanced Mode you can add more detailed and specific filters for your request:
MODIFIED AMINO ACIDModified amino acid type ( = the residue type found at the PTM site position in the UniProt sequence).
SECONDARY STRUCTURE ON PTM LOCATIONSecondary structure assignment at the PTMs positions. The secondary structure is assign by DSSP, and the corresponding output "letters" are :
Bresidue in isolated beta-bridge
Cloop or irregular
Eextended strand, participates in beta ladder
G3-helix (3/10 helix)
I5 helix (pi helix)
Thydrogen bonded turn
SCOP CLASSIFICATION OF PROTEINThe Structural Classification Of Proteins (SCOP). By using this filter, you can restraint your search to specific structural classes of proteins.
The different classes available in PTM-SD are:
aAll alpha proteins
bAll beta proteins
cAlpha and beta proteins (a/b)
dAlpha and beta proteins (a+b)
eMulti-domain proteins (alpha and beta)
fMembrane and cell surface proteins and peptides
gSmall proteins
hCoiled coil proteins
iLow resolution protein structures
NoneNo yet classified by SCOP
Number of PTM by pdb_chainDefined the number of PTM sites found on the same PDB chain.
Length of pdb_chainDefined the length of the PDB chain in which the PTM sites are found. It is possible to take or don't take into account the number of missing residues on the chain.
ANNOTATION FOUND IN MODRES PDB RECORDSMODRES comment found in the PDB structural files selected to build PTM-SD.
The MODRES record provides descriptions of macromolecule modifications found in structure; its generated by the wwPDB, and could be related to chemical, engineering, or post-translationnal modification.
ANNOTATION FOUND IN dbPTMThe dbPTM PTM annotations. Its correspond to the detailed ones from the downloadable data for experimentally verified PTM sites.
By selecting value in this filter you request for specific PTM annotations.
However these annotations could not correspond to the PTM found in the cristal structure. For more clarification see the examples below.
To come back to Simple Mode click on the Hide advanced search button. Note: your selection in Advanced Mode will be deselected by returning to Simple Mode. Some examples to understand how to use the ANNOTATION FOUND IN dbPTM advanced filter:
First a simple example of search. By selecting Acetylation in the PTM filter, you can search for: "All records corresponding to Acetylation". There are (on January 2014) 93 corresponding modified residues both structurally resolved in PDB structures and experimentaly annotated in dbPTM.
By selecting "N6-acetyllysine" in the abvanced filter dbptm_annot, you can search for: "All records for which there is 'N6-acetyllysine' in the dbPTM annotation.". There are (on January 2014) 181 corresponding modified residues both structurally resolved in PDB structures and experimentaly annotated in dbPTM. As you can see there is more corresponding entries than in the previous example, in which the search was made for Acetylation entries. It is important to understand that the position of one PTMs site both structurally resolved and experimentaly annotated could also correspond to the position of other PTM annotations. See the N-TRIMETHYLLYSINE found in the Chain B of PDB 3N9L; This PTMs site are annotated in dbPTM as N6,N6,N6-trimethyllysine but also as N6-acetyllysine.
So by using both simple and advanced filters, it is possible to make some specific request. Thereby selecting Acetylation in the PTM filter, and "N6-acetyllysine" in the abvanced one dbptm_annot, you can search for: "All records corresponding to Acetylation, on which there is 'N6-acetyllysine' in the dbPTM annotation.". This could be traduct as "All N6-acetylation of lysine both structurally resolved and experimentaly annotated". There are (on January 2014) 83 corresponding entries in PTM-SD.
Result     table By clicking on the Filter button, the search results display below the filter table. The result table contains one line for each corresponding PTM site from PTM-SD. It contains several information:
Detailsa link on the PTM-SD details page of the PTM site,
organismthe protein organism,
uniprot_acthe UniProt AC with a link on the UniProt page of the protein,
pdb_idthe PDB id Code with a link on the PDB page of the structure,
pdb_chainthe PDB id Chain,
uniprot_posthe position of the PTM site in the UniProt sequence,
pdb_posthe position of the PTM site in the PDB structure,
pdb_modresthe PDB MODRES comment related to the PTM site,
dbptm_annotall the dbPTM annotations related to the PTM site position in the protein,
PTMthe general annotation of the PTM site,
scop_classthe structural classification of the protein from SCOP,
aa_uniprotthe amino acid type of the PTM site position in the UniProt sequence,
pbthe local structure assignment (Protein Blocks) of -10/+10 positions surrounding the PTM site (highlighted in purple),
ssthe secondary structure assignment (DSSP) of -10/+10 positions surrounding the PTM site (highlighted in purple).

PTM-SD provide 3 tools to compute statistics and analysis on selected entries from the result table.
Select entries by using the checkbox at the begin of each result table line. Use the top checkbox to Select/Deselect all entries in the result table.
Statistics Click on the Statistics button to compute pie chart of organism distribution, protein distribution and general PTM annotation distribution. Click on the Show details button to display the raw data of the related pie chart. The raw data are sortable by name, frequency or percentage. Neq Click on the Neq button to compute Neq graphic.
The equivalent number of PBs (Neq) is a statistical measurement similar to an entropy, and represents the average number of PBs at a given position. Neq is calculated as follows: Where px is the probability of PB x. Note: A Neq value of 1 indicates that only one type of PB is observed, while a value of 16 is equivalent to a random distribution.
Clustering The visitor has the possibility to compute dendogram and histogram of pairwise sequence identity between selected proteins. He could also reduce redundancy in his selected entries thereby creating a non-redundant selection. For the moment the clustering process is automatically made by the statistical sofware R (using hclust and cutree function). Follow the example below to know how to use the Clustering tools:
1) Make a request and select all the return PTM-SD entries in the result table. (see Browse Database section for help) In our example we searched for all Methylation entries. On January 2014 there was 861 related entries. 2) Then use Statistic tool to compute pie charts. There are some proteins that are more represent than the other, and maybe there are some protein that are homologs. These 861 Methylation entries are observed in 163 PDB files, which represent 40 different proteins (40 different UniProt AC).
3) Click on "Clustering" On the left is compute a dendrogram which represents the homology between these 40 proteins based on pairwise sequence identities provided by Clustal W. On the center is display a histogram which represents the distribution of the pairwise sequence identities used to compute the dendogram. On the rigth you have the possibility to reduce redundancy in your selected entries. 4) Choose a threshold value (default is 30%) and 5) Click on "Reduce redundancy"
In our case 21 proteins were remove from our selection, which represent the deselection of 485 lines (PTM-SD entries) in the results table.
The dendrogram and the histogram were automatically updated according to entries which remained selected in the table. The threshold value is indicated by a red line in both figures. Some entries which remain selected in the table could correspond to the same position in the same protein but found in different PDB files. 6) Click on "Remove duplicate positions"
By using again the Statistic tool, the reduction of the redundancy in the selected entries is observed.

Create     your     dataset
At any moment you have the possibility to donwload 2 kind of data corresponding to selected entries in the result table.
List of the P D B - I D in which the selected PTMs are found. A .csv style file containing the PTM-SD data of selected entries.
By using Tools you have the possibility to create a non-redundant selection, and then create your own dataset. To obtain it follow this 4 steps:
1 Request the database by using filters in the Browse Database section
2 Select entries in the result table. If needed, use the Clustering tool to reduce redundancy in your selection.
3 Go to CREATE YOUR DATASET section. Click on the HERE button :
4 And follow the instructions that have just appear.

The "Details" page of each PTM site provide an alignment table highlighting the sequence/structure relationships in the chain which contains the concerned PTM site. This last includes an alignment of the protein amino acids sequence, secondary structure and local structure (Protein Blocks) assignments. The PTM position corresponding to the current page is highlighting in purple and the other PTM sites present in the chain are colored in yellow. The same color code is used in the images gallery. The PTM positions found in other PDB structures are highlighting in green. Uniprot     Sequence The Uniprot sequence corresponds to protein sequence in which the PTM site is annotated. Clustal The sequences alignment were performed using Clustal W (2.0.12). The symbols for similarity/idendity of position, provide by Clustal W, were also aligned. They facilitate the observation of differences between the protein sequence from UniProt and the chain residues from the PDB. PDB     Sequence This sequence corresponds to the chain residues found in the PDB file. The amino acids are sorted according to the polypeptide bond order, from the N-termini to the C-termini. The amino acids are in one letter code, and the 'X' label correponds to unnatural residues. Note: This sequence could not corresponds to the fasta sequence which could be found in the related PDB page. Indeed, the fasta could not include residues insertions, mutations, ... PDB     Information The PDB information sequence is a linear representation of the reccord type of the PDB file (version 3.30). A one letter code is used for each PDB chain positions, as follow:
HThe residue is defined on the 'HETATM' PDB record. "Non-polymer or other “non-standard” chemical coordinates, such as water molecules or atoms presented in HET groups use the HETATM record type."
_The residue is defined on the 'ATOM ' PDB record. "The ATOM records present the atomic coordinates for standard amino acids and nucleotides."
IThe residue is numeroted as an inserted residue. "Alphabet letters are commonly used for insertion code. The insertion code is used when two residues have the same numbering. The combination of residue numbering and insertion code defines the unique residue."
MThe residue is defined in the 'REMARK 465 MISSING RESIDUES' PDB record. "REMARK 465 lists the residues that are present in the SEQRES records but are completely absent from the coordinates section."
SThe residue is defined on the 'SEQADV' PDB record. "The SEQADV record identifies differences between sequence information in the SEQRES records of the PDB entry and the sequence database entry given in DBREF."
As some positions could cumulate these information, the following hierarchical relation is used to choose the one letter representation: M > S > H > I > _ DSSP     Sequence This sequence correspond to the secondary structure assignment made by DSSP. The output "letters" are :
Bresidue in isolated beta-bridge
Cloop or irregular
Eextended strand, participates in beta ladder
G3-helix (3/10 helix)
I5 helix (pi helix)
Thydrogen bonded turn
Protein     Block     Sequence This sequence correspond to the Protein Blocks (PBs [ref]) assignment made by a slightly modified Python PBxplore tool.
This structural alphabet is composed of 16 local structure prototypes of 5 residues in length. They efficiently approximate every part of protein structures. PBs m and d can be roughly described as prototypes for the central region of α-helix and β-strand, respectively; PBs a through c primarily represent the N-cap of β-strand while e and f correspond to C-caps; PBs g through j are specific to coils; PBs k and l correspond to N cap of α-helix while PBs n through p to C-caps. They have been used in various approaches, e.g. protein superimposition [ref, ref], protein binding site analysis [ref, ref] or prediction [ref, ref].