# Mon Dec 31 13:17:56 PST 2007 Kevin Karplus

Change picture generation at end to show models 1 through 5.

make_alignments needs to make multi-track alignments even for
sequences not in template library (as long as they have basic-setup done).

Make template library multi-track HMMs.
Add alignment methods for multi-track template HMMs.

Add consensus-based scoring for template models.
	[[ Tue Jan  1 13:49:13 PST 2008 Kevin Karplus
		Done for single-track HMMs, still need to do
		multi-track template-based scoring!
	]]


Add consensus-based scoring for target models
	[[ Tue Jan  1 13:32:34 PST 2008 Kevin Karplus
		Done for multi-track, but not w0.5 model.
	]]

Really need 2 scoring methods for each HMM:
	1) real sequence and structure for all of PDB (basic-setup)
	2) consensus sequence and real structure for template library


Modify undertaker script and  to do 5 optimizations: one with all the
alignments, the next 4 with just the first 4 templates.

Get George's latest contact prediction stuff.  Fix MQA_highE.costfcn
when George tells me how to make NN1000.

Modify rasmol display of solution to have coloring script controlled
by the Make.main file.  Better still, replace the static pictures with
jmol or other display technique.

------------------------------------------------------------
BELOW THIS LINE IS DONE
------------------------------------------------------------
	
	Add str4s and CB8-sep9s to pcem/indexes for t2k, t04, t06.
	Add pbs to pcem/indexes for t04, 
	Add n_seps, o_seps to pcem/indexes for t2k
	Add maxps to pcem/indexes for t2k,t04,t06
	[[ Tue Jan  1 13:20:24 PST 2008 Kevin Karplus
		Added to pcem/indexes, but not run yet,
		(need to make all the maxp files first.)
	   Wed Jan  2 16:07:30 PST 2008 Kevin Karplus
		Done.
	]]
	

	Compute consensus sequences for t2k, t04, t06.	
	[[ Wed Jan  2 16:09:09 PST 2008 Kevin Karplus
	   DONE, needed to fix a number of t2k models that had the wrong number of columns!
	]]

	in pcem/scripts/get-seqs	only gunzip if ends with .gz, to speed things up a bit
	[[ Wed Jan  2 16:10:31 PST 2008 Kevin Karplus
	   DONE, but it doesn't seem to speed things up.
	]]

    Missing several neural nets for predicting the chosen alphabets:
	    PURE_SECONDARY_ALPHABETS := str4 str2 alpha bys pb  n_notor2 o_notor2  n_sep o_sep dssp-ehl2
	    BURIAL_ALPHABETS :=   CB_burial_14_7 near-backbone-11 CB8-sep9

	    t2k	n_sep
	    t2k 	str2
	    t04	str2

    If we want to add dssp-ebghstl or stride-ebghtl, we'll need networks
    for them also.

	[[ DONE 14 Jan 2008 ]]


	Re-connect email.	[DONE 4 April 2008]
	
	remove .al alignments from top_reported_alignments.html, since we
	aren't making any .al alignments any more.
	[[Sun Apr  6 11:18:37 PDT 2008 Kevin Karplus	DONE]]


	Get Martin P's changes to undertaker and get his constraints and non-contacts.
	[[DONE ]]

	Modify pick_alignments not to reject multiple hits in SCOP library.
	[[DONE Sun Apr  6 11:35:17 PDT 2008 Kevin Karplus]]


	Make a "shit-list" for casp8 server to block particular users.
	[[DONE for SAM_T06 and SAM_T08
	Sun Apr  6 17:24:38 PDT 2008 Kevin Karplus
	]]



	Modify two-track-alignments and three-track-alignments to be called
	for all template sequences, not just ones in template library.
	This requires a new target, as consensus sequences are used for
	alignment in the target library.
	[[DONE Wed Apr  9 17:34:20 PDT 2008 Kevin Karplus
	  Didn't really need a new target, just a test whether maxp
		sequence exists.
	]]

	Make sure weights for 2-track alignments are right.
	[[ DONE Wed Apr  9 17:36:54 PDT 2008 Kevin Karplus
	   Fixed for both scoring and alignment.
	]]


	Need makelogo that understands str4 and CB8-sep9.
	Get version of hmmscore with new calibration.	
	[[ DONE Sun Apr 13 20:46:11 PDT 2008 Kevin Karplus
	   This version is not yet in the SAM cvs tree.
	]]


	Modify try1 to do selection of initial models with MQA cost
	function.  Modify try1 to change cost function gradually
	(increasing soft_clashes and breaks, and other 'physical' cost
	functions---may want to increase more components.
	[[ DONE Sun Apr 13 20:47:27 PDT 2008 Kevin Karplus
	]]