I have found no interactions for my protein rat Zap70. How do I find if a homolog/ortholog/paralog with a high degree of sequence similarity has any interactions in the IntAct database?

Go to www.ebi.ac.uk/tools/sss
Select your Search Engine of choice e.g. NCBI-BLAST
Click on protein then paste your sequence into the box, make any required changes to the parameters, add your e-mail address and Submit.

You will receive the results in tabular form, with additional hyperlinks through to other databases containing information about this protein.
You can see below that, although InAct contains no molecular interactions for rat or mouse it does for human

Press on the link to access the data in IntAct

How to link to the IntAct website?

There are currently 5 supported ways to link to the IntAct website:

1. Interaction tab using a free text query (MIQL):
   Example: http://www.ebi.ac.uk/intact/query/bbc1
   
2. Details tab by experiment AC:
   Given an experiment AC, one can open the details view.
   Note: this will not show any interaction details but only experiment and respective publication.
   Example: http://www.ebi.ac.uk/intact/pages/details/details.xhtml?experimentAc=EBI-768904
   
3. Details tab by interaction AC:
   Given an interaction AC, one can open the detail view, for instance:
   Example: http://www.ebi.ac.uk/intact/interaction/EBI-2307627
4. Interaction given a pair of interactor:
   Given two interactors' AC, one can open the interaction detail view, for instance:
   Example: http://www.ebi.ac.uk/intact/pages/details/details.xhtml?binary=EBI-3437,EBI-2750
5. Molecule view by interactor AC:
   Given an interactor AC, one can open the Molecule view (Dasty2), for instance:
   Example: http://www.ebi.ac.uk/intact/molecule/EBI-349905

Why does the number of interactions given in a particular high-throughput paper, differ from the number I can download from IntAct from the same publication?

Authors state the number of interactions they have identified, based on protein sequence identifications made at the time of going to press. When we upload this data into the IntAct molecular interaction database, we map these identifications to the UniProtKB protein sequence database, so that our users can benefit from the high-quality manual annotation of these sequences performed by the UniProtKB/Swiss-Prot curators. The number of interactions can vary both as a result of this process and also as a result of time for the following reasons.

1. The author may have used a redundant database to identify their proteins. We often find separate identifiers map to a common non-redundant UniProtKB entry and an interaction turns out to have been identified twice, but originally assigned to different protein IDs.
2. Proteins may have been withdrawn from the protein sequence database over time - this is particularity true for predicted proteins based on gene models. If we cannot remap the protein to an appropriate alternative, we may have to withdraw the interaction.
3. Gene models may merge resulting in only one interacting pair being identified, where previously there was thought to be two.
4. Gene models may be split over time - in this case any interaction based on a protein product of the original gene model will have to be withdrawn unless we have enough data to tell us which of the new proteins is involved in the interaction.
5. Authors may submit additional interactions which they did not describe in the original paper but were generated by the same method. This will be maintained separately and labelled as such.
6. Some interactions pairs included by the author may break the annotation rules of both the IntAct molecular interaction database and the IMEx consortium If so, these interactions are excluded and a note added as a "data-processing" comment to explain why.

How are binary interactions selected for export to UniProtKB/GOA records?

All binary interactions evidences in the IntAct database, including those generated by Spoke expansion of co-complex data, are clustered to produce a non- redundant set of protein pairs. Each binary pair is then scored, using a simple addition of the cumulated value of a weighted score for the interaction detection method and the interaction type for each interaction evidence associated with that binary pair. Only experimental data is scored, inferred interactions, for example, would not be scored, and any low confidence data, or data manually tagged by a curator for exclusion from the process, are also not scored. Isoforms and post-processed protein chains are regarded as individual proteins for scoring purposes. We will, in future, export negative information, particularly for isoforms when one isoform of a protein behaves differently from another.

Score weightings were determined using the PSI-MI CV (available here) hierarchy i.e

Interaction Detection Method	Weighting
Biochemical	3
Biophysical	3
Protein complementation Assay (PCA)	2
Imaging Techniques	0.6

Interaction Type	Weighting
Association	1
Physical Association	2
Direct Association (and child terms)	5
Colocalization	0.2

Example 1 :

Protein A - Protein B – the interaction has been shown by one yeast two-hybrid experiment and one coimmunoprecipitation as part of an affinity complex from a cellular environment.
1xY2H (PCA) + Physical interaction = 2+2 =4
1x coimmunopreciptiation + Association = 3+1 =4
Total score = 8

Example 2 :

Protein C - Protein D – the interaction has been shown by two yeast two-hybrid experiment, by one GST pulldown in vitro and by one X-ray crystal.
2xY2H (PCA) + Physical interaction = 2(2+2) =8
1x pulldown + direct interaction = 3+5 =8
1x x-ray crystal + direct interaction = 8
Total score = 24

Once the interactions have been scored, a cut-off filter of 9 has been established, below which the interaction is not exported to UniProtKB and to the Gene Ontology annotation files. Additional rules ensure that any protein pair scoring above 9 must also include at least one evidence of a binary pair, excluding spoke expanded data, before export to UniProtKB/GOA.

These criteria ensure that

1. Only experimental data is used for making the decision to export the protein pair to UniProtKB/GOA as a true binary interacting pair
2. The export decision is always based on at least two pieces of experimental data. A single evidence cannot score highly enough to trigger an export
3. An export cannot be triggered if the protein pair only ever co-occurs in larger complexes, there must be at least one evidence that the proteins are probably in physical contact.

Although these are stringent criteria, it is believed this will make the highest possible quality data available to UniProtKB and GO users. It is our intention to add the binary pairs annotated by other IMEx consortium members as part of the IMEx project to this process in the near future, to increase the data exported out to UniProtKB and GOA. The IntAct group are also happy to make this data available to any other databases who wish to import and display it. Please contact intact-help@ebi.ac.uk to discuss the most appropriate format for your resource.

How is the intact-miscore calculated?

The IntAct MI score is based on the manual annotation of every instance of a binary interaction (A-B) within the IntAct database. First all instances of the A-B interacting pair are clustered by accession number. Each entry has been annotated using the PSI-CVs and we use this information to score by the interaction detection method and by the interaction type. Additionally we count the number of publications the interaction has appeared in, up to a maximum of 8. Each of these variables is normalised between 0-1. The cumulative score is also normalised between 0-1 across the entire IntAct database, with 1 representing an interaction in which we have the highest confidence.