Retiring and replacing the BLink protein similarity service


NCBI is discontinuing the BLink protein similarity service effective immediately. BLink provided graphical access to related proteins from protein records in the Entrez system. Because of the increasing volume of data in the protein database, BLink has become less useful as a tool for finding related sequences and is no longer maintainable.

Temporary replacement for BLink

The BLink service will redirect to a live protein-protein BLAST search against the Landmark database used by SmartBLAST. The Landmark database, described in the SmartBLAST documentation , provides matches from 27 selected cellular organisms with well-annotated complete genomes representing a broad taxonomic range. The results from the redirected BLink search will be shown as a Tax BLAST report  as shown in the figure below. The Tax BLAST report emphasizes the taxonomic source of the protein matches as did the BLink output.  From this new starting point, you can explore additional protein similarities through the BLAST service by re-submitting the search against other blast databases including the non-redundant (nr) database.

A Tax BLAST report

Figure 1. The Tax BLAST report for proton ATPase A. (click to enlarge)

Ultimate replacement service for BLink

We would like to provide a more sophisticated replacement for BLink. Please provide comments on this post and let us know what features would be useful to you in a service that would allow you to quickly identify and sort related proteins. For instance:

  • Are you interested in similar proteins in closely related or distantly related species or both?
  • Are BLAST statistics important on the output?
  • Is the ability to sort hits by taxonomy useful to you?
  • Are there other sorting options that would be useful to you?
  • What NCBI resources you would like to see linked to the protein similarity output (Gene, a genome browser, expression data, etc.)?
  • Is the initial selection of species in the Landmark database a good starting set? Are there representative species you think should be added?

And if you are a previous user of BLink please let us know:

  • What features of the BLink output did you find most useful?
  • What features of BLink output that you found useful are missing in the Tax BLAST report?
  • What advantages did BLink have for you over running a live blastp search with a protein of interest?

And, of course, provide any other suggestions here or via email to help us make a new protein similarity service that would be most useful to you.

3 thoughts on “Retiring and replacing the BLink protein similarity service

  1. It is pity that Blink was retired, but i can understand. it was slow and it took several months for new genomic sequences to enter the Blink server.

    My laboratory is constantly using Blink since 2002. I must admit that many of my colleague had no idea what was Blink, but once you walk them with Blink—they were captured -just like me.

    I found Blink to be one of the best and most valuable program ever created for both researcher like me with pure enzymology and structural protein in mind (i.e. ability to easily determine between distance species the most important aa involved in catalyses). and to people (like me) who wish to study evolution and metabolic pathways. Blink was an easy platform to just click and see the pre-calculated BLaST alignment results (of all species) and directly with a click on the bottom you linked in to the genome of the Sp and look at flanking sequences …very ..very valuable

    regarding your specific question:

    Are you interested in similar proteins in closely related or distantly related species or both?
    YES, very much. Observing sequence alignment of closely related or distantly related species was the strength of Blink.
    Are BLAST statistics important on the output?
    Very much. And a linked to the pre-calculated Blast alignment was also very useful.
    Is the ability to sort hits by taxonomy useful to you?
    Very much . in the Past 10 years my lab published numerous articles, where w/o the ease of Blink -the work and pre-research was more time consuming.

    Are there other sorting options that would be useful to you?
    The filters that blink had were good (i truly admire the people who formed Blink they had good vision).

    In addition, the dropdown capability to select view of say ONLY fungi for example, was good feature.
    Also i found ability to look at more than 100 per pg.- good

    – If other (NEW) features added to the new program: I suggest (in the filter) ability to linked the data to search Protein Data Bank proteins(pdb) for structural work.

    – I also found the ability to click (on the Blink) and view the PROTEIN seq very valuable (i don’t know the name of the view pg in GenPept report
    see for example with all the added links in the left, https://www.ncbi.nlm.nih.gov/protein/XP_015401470.1
    What NCBI resources you would like to see linked to the protein similarity output (Gene, a genome browser, expression data, etc.)?
    these are interesting features ;
    but it may complicate the simple function of Blink
    Is the initial selection of species in the Landmark database a good starting set? Are there representative species you think should be added?
    the original landmark :arachea, bacteria, animal, plant, virus, and other were PERFECT . I will not try to break the Bacteria to different subclasses
    And if you are a previous user of BLink please let us know:

    What features of the BLink output did you find most useful?
    See all species, the pre-calculated Blast and ability to immediately look at the alignment,
    The scoring , and ability to click and view the protein.
    What features of BLink output that you found useful are missing in the Tax BLAST report?
    I found Tax Blast report VERY confusing ….
    What advantages did BLink have for you over running a live blastp search with a protein of interest?
    Time.
    Ability to see alignment of a virus for the same protein as with human or plant,

    The thing i didn’t like in the OLD Blink was the time of genome seq updates. Often it would take 3-6 months from the time a sequence was published and it was in the Blink.

    Keep the good work , and i hope you can keep the Blink with new features.

    There are of course some features that you can show a person…and its hard to convey by words..especially for people who are not used to Blink,

    Maor

  2. I really loved Blink, it was so much faster than a live blast search, and very easy to work on multiple proteins at once, or hop from one blink hit to a new record and more blinks….

    By far the best change for me would be to make a semi-redundant Blast database, a bit like Uniref50, except maybe filtered by genomes. If I blink a human protein, the first 20 will be primate orthologs, and the first few hundred will be vertebrate orthologs, which is usually not very helpful. Trimming down the genomes used would allow search for close and distant homologs much more quickly. I’d also love to see such an option for Blast databases. It looks like the Landmark Database does that, but with just 27 genomes, it’s terribly thin…

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s