TODO

Incremental

• Show ORFs from other direction too.
• Get rid of jiggle in the colored alignment (introduced by gaps).
• Change the coloring of the alignment graph as suggested by Derek.
• Make a panel of alignment graphs with the same X axis scale, see note above about heurstically selecting which subject sequences to display.
• Come up with heuristic for sequences to align with.
• Can we zoom on a rectangular section of an alignment graph and get a summary of the read info there? First do this a bit manually.
• Add temperature of HSP rank to alignment graphs (re david suggestion). But what about hits on (essentially) the same thing?
• Build consensus sequence from reads, given an alignment against a known target sequence.
• Compute total target sequence coverage by all hits, weighted by e-value (somehow like an integral).
• Can we use SVG for graphs?
• Look at odd axis problem with reversed ORFs
• Get rid of blurriness!
• Automate the splitting of read files for BLAST and putting the results back together.
• Add vertical line to alignment panel to show end of subject.
• What happens when we BLAST the stealth virus against all seqs?
• Look across the various read sets to see what's in common (in terms of sequences that are hit)
• What do the silo patterns mean in matching sequences?
• Look again at bloody blast scores and influence of read length
• Look at unrestricted sequence length matches in TJ data.
• Filter out integration site hits.
• Send BM the commands to make alignment graphs, etc
• Add an option to alignmentGraph to have it show e-values by rank.
• BLAST the stealth virus against allsequences to find out what it really is.
• Make it so that ORF plots only show the minStart to maxStop region
• Make it so that Feature plot only shows the minStart to maxStop region
• Add vertical feature lines to the reversed ORF plot.
• Can we put something on the RHS of alignment plots to show length of perfect match at this e-value?
• Make the alignment plots always start from y=0

questions from 2013-08-29

• What is the overhang that we see in Barbara's alignment graphs?
• Why are there internal whiskers and column boundaries on one of Barbara's plots?
• Why do we not see all the results from BLAST? Seqs too short? Yes, too short or low complexity.
• Is the topmost read not shown in an alignment graph?
• The right hand vertical bar in the alignment graphs is for the rightmost read not the end of the target seq.

science from 2013-08-29

• Make log/log plots of counts
• Do MERS cov reads against the MERS cov reads

Questions for Ron/Sander re merkel DNA

• Why aren't the 14 MCV reads of interest??? Is this end of story / computational diagnostics in this case?
• Show them the tools, emphasize non-cultured, default flow, etc
• Lack of whiskers => not a new virus?
• What should we be thinking about HEVs? Ignore? Interesting?

Questions for Sander and Ron

• How much Terry sample remains? Could we do more experiments?
• How can it be that some read counts are 1 when some others are 4000 ?
• WTF do the columns in the plots correspond to?
• What are the sharp edges in the columns?
• What happens if they sequence saline solution? If they run DNA/RNA killing on a sample?

Server / Admin

• Set up github account.
• Get Terry/Barbara iPython notebooks up on the acorg server.
• Populate git repo with code.
• Make small zip file of data and get BM working on it.
• Set us up to run BLAST on acorg servers not just locally.

Data

• Find out what other nucleotide/protein databases you should be BLASTing against.
• Ask Sander/Ron for more data sets
• Make bacterial subset of NCBI db & look for others who have done this.
• Can we better subset to make viral or bacterial databases?
• Look at data from Ron in comparing reads from various different technologies.

Misc

• Try assembly of puffinosis and others.
• Get an understanding of how other groups are putting viruses together.

Coord with

• Christian Drosten
• Lea at AMC
• Matt Cotton
• EBI "just down the freaking road" Euro Bioinformatics Institute