Help/What's new with the SILVER?

• Is SILVER restricted to use with any particular software?

  SILVER requires only a list of candidate peptides. How you generate that list is up to you. If you're using SEQUEST, then the .OUT file will do nicely as your candidate-peptide list. If you're using Mascot, we've written this Python module to help generate a list from the output results.

• Can I use SILVER on results obtained from the Mascot website?

  No. We offer a Python script for parsing the results file produced by the command-line version of Mascot. The web pages produced by Mascot running in CGI mode are difficult to parse.

• How can I use SILVER with Mascot results?

  Assuming you have a complete results file produced by Mascot running in command-line mode (not the results obtained from the website), you can download this Python module to help parse it. The output produced consists of one line for each spectrum, and lists all peptides that matched that spectrum.

• What are the allowed formats for the candidate-peptide list?
  1. One option is to use a SEQUEST .OUT file. You can upload it as is. Since the format is well known, it can be interprted by SILVER. Only the most relevant data will be shown.
  2. If that is not available (for example, if you're not using SEQUEST), you can use the 'generic format', as specified in the following:
     • Each line consists of an arbitrary number of elements, separated by whitespace (e.g., tabs, blanks).
     • The first element on each line is assumed to be the peptide name.
     • There can be only one peptide per line.
     • Lines beginning with '#' will be ignored.
     • All elements will be displayed in a table, but elements other than the peptide name will not be interpreted in any way.

• February 2004
  As mentioned on SILVER's front page,
  • the main paper came out in Nature Biotechnology this month,
  • and the application note has been accepted for Journal of the American Society for Mass Spectroscopy.
• September 2003
  • Main LOD-scoring paper almost accepted, applications note for SILVER website ready to go too.
• August 2003
  • New tree representation makes LOD-scoring instantaneous. No need to make it optional.
  • Predicted spectra available too.
• July 2003
  • Moving to the Gygi Lab, where it will be work alongside GoFishing. It will remain available here too.
  • CSS redesigned for greater readability (dark blue too similar to black).
  • Final bugs ironed out, dealing with unobserved values, significant mostly on modified residues.
  • Suite of demonstration files augmented.
  • Clean up code.
• May 2003
  • Added horizontal bars to break up page into logical sections.
  • Switched SEQUEST-result and peptide-fragment tables.
  • Fixed parsing and display of modified residues.
  • LOD-threshold widget is gone - not useful.
  • Demo files now show how SEQUEST missed best match, but LOD gets it.
• April 2003
  • Updated demo files to show example where LOD has better resolving power than SEQUEST.
  • SEQUEST output shows Sp scores, and true SEQUEST rank.
  • Layout cleaner.
  • Fragment labels are on by default, and now located in two rows at the bottom of the figure (top row for B ions, lower row for Y).
  • Added 'demo' button, to load a sample spectrum and corresponding SEQUEST output.
  • Modified residues are handled more gracefully, though still not correctly.
  • More layout changes.
  • Check boxes are remembered between peptides.
  • LOD scores re-computed only when new .OUT file is loaded (and only when this feature is turned on).
  • Fixed parsing of .OUT files, when extra ORF information is added.
  • Peptides in OUT-table are URL-encoded requests to reload the page with that peptide. Right-clicking will open a new browser window.
  • LOD-scoring can be turned off, since it's still rather slow.
  • Layout's a little different.
  • When peak labeling is turned on, peaks show fragment names (e.g., 'b4', 'y13'), not m/z values.
  • When no peptide is specified, the top-ranked peptide in the .OUT file is used.
  • Page generation takes two seconds, on average.
  • Filenames are remembered between pages.
  • Filenames not truncated, wrapped by browser.
  • OUT file includes deltCn information, soon to have LOD-scores added too.
  • Various minor layout tweaks.
  • Next version will label LOD peaks with fragment indicators, not m/z-values.
  • Ability to upload .OUT files. Only SEQUEST rank, peptide name, and xcorr shown. No related functionality yet.
  • Performance restored - images generated in a second or two.
  • Labels in graph match those of fragments (sometimes off by 0.1).
  • All peaks shown correctly in graph now, even those which are not matched.
  • Horizontal axis trimmed to show region with peaks (to within nearest 100 amu).
  • Discrepancies between Josh's and Frank's parsing of the decision tree almost completely ironed out. Only two remaining questions.
  • To do: make labels on LOD indicators in graph show the fragment masses, rather than the observed masses.
  • Check that all LODs are truly indicated on the graph - it looks like some of them may not be...
• March 2003
  • Fragment table:
    • Fragment ends added to fragment table.
    • Corrected 'offset' problem, so that complementary B/Y fragments now share the same row. Grayed-out boxes indicate no fragment possible.
    • Extra columns added to indicate which B/Y ion is in each row.
  • In chart, horizontal axis tick-labels are now aligned with ticks for improved readability.
  • Default peptide ends in '.N', not '.K'
  • Tree figures updated to show trees actually used in computation.
  • Fixed bug in computation of LOD.
  • Labels can be turned on/off in graph.
  • LOD scores shown for individual fragments, with color highlighting to match figure.
  • Enhanced version allows spectral data to be:
    1. uploaded from a DTA file
    2. pasted from a file into a text-area
  • Prototype goes online.