Jumping in here before Geoff... On 02/11/2016 11:38, Mungo Carstairs (Staff) wrote: > The bottom line seems to be that Jalview code treats Threonine as not > hydrophobic. OK. So that's in agreement with 'most' googleable sources.
> This is a bug if the intention is to match the diagram and matrix > published to support the method. If anything, the bug is only that the matrix doesn't reflect what Jalview uses. It would be good to also add some additional documentation and links to appropriate reference sources too. > Jim is of the opinion that we should correct this, but keep the default > behaviour backwards compatible with past versions. Contrary to what I said in our standup today, there is precedence for simply changing the matrix in the code in line with the diagram. When I was working on N-glycosylation patterns with Jalview, I noticed that N was marked as charged in the matrix, but in the diagram it is simply small and polar. I silently changed it. Whether that was the 'right thing' 7 years ago is a good question. > A question on thresholding the conservation calculation: > > The AMAS paper refers to ignoring counts representing less than /N%/ of > the column [the 'N%' is a missing image in the html version] but doesn't > state the choice of /N/ as far as I can see. Jalview has /N/=3. Is it > possible to view the original code to compare? https://svn.lifesci.dundee.ac.uk/svn/barton/software/amas/ (log in with your Lifecci password) > A question on treatment of gaps: (over to Geoff) > Jalview also applies a threshold requiring less than 25% gaps in a > column to compute any conservation. This doesn't seem to be mentioned in > the paper or Jalview help. I would be interested to know if it is from > the original implementation or an addition. I thought it was mentioned somewhere... odd that it never made it! Easily fixed of course. As I've emailed in our private thread on this - there are experts who would like to change these thresholds (e.g. compute conservation over just the non-gapped sequences in each column, and don't ignore residues with abundances less than a certain threshold). Jim. -- ------------------------------------------------------------------- Dr JB Procter, Jalview Coordinator, The Barton Group Division of Computational Biology, School of Life Sciences University of Dundee, Dundee DD1 5EH, UK. +44 1382 388734 | www.jalview.org | www.compbio.dundee.ac.uk _______________________________________________ Jalview-dev mailing list [email protected] http://www.compbio.dundee.ac.uk/mailman/listinfo/jalview-dev
