IS-Dom: a dataset of Independent Structural Domains

Release 1.0, Feb 2013

IS-Dom is a dataset of independent structural domains (ISDs) that is highly likely to
fold in isolation. The structural independence of the domains is assessed by calculating
inter-domain hydrogen bonds and hydrophobic clusters from the atomic coordinates of the fulllength protein.

A domain is defined as an ISD when it forms less inter-domain hydrophobic clusters
and H-bond than a given set of threshold values that the user can set freely.

ISDs and adjustable parameters

IS-Dom release 1.0 includes 3,500,658 ISD datasets that are retrievable using the user’s adjusted parameters.
The adjustable parameters are the dataset (SCOP, CATH and DomainParser) from which ISDs are derived,
threshold numbers for inter-domain H-bonds and hydrophobic clusters, as well as a distance threshold
for detecting hydrophobic clusters.

Small threshold numbers will yield a strict ISD definition and provide small sets of ISDs,which will be highlyisolated from the rest of the protein.
On the other hand, large threshold numbers generate larger datasets containing slightly less isolated ISDs.

A small distance threshold will yield a strict definition for hydrophobic clusters (atoms need to be in close contact to fulfill the cluster’s definition).
A large distance will yield a looser definition: Atoms far apart will be accepted as a cluster’s member.

Users can download full/representative ISDs in the result page.
In the ISD list, first column indicates PDB IDs and chain IDs, and second to third column represents ISD regions in the PDB chain.

Standard IS-Dom

We defined default parameter values to calculate a standard set of ISDs: Standard IS-Dom.
The default parameters are determined by maximizing the fraction of ISDs that were >30% similar
and whose termini overlapped within a 10-residue-margin to sequences cataloged in the PDB (These domains were named
Autonomously Folded Domains and used as control in our study. See details in Ebina et al, Submitted).

ISD-Checker

The IS-Dom server can also test the structural independency of putative structural domains defined by the user.
Residue-residue interactions for all of the 213,010 PDB chains are stored, and ISD-checker uses them to assess the independence
of the user's defined domain and classify them either as ISD or non-ISDs.
The classification, number of inter-domain interactions and the location of the domain in the PDB structure will be shown in the result page.

Reference

Teppei Ebina, Yuki Umezawa, Yutaka Kuroda:
IS-Dom: a dataset of independent structural domains automatically delineated from protein structures.
J Comput Aided Mol Des. 2013 May;27(5):419-26.



BACK