Be more forgiving of non-standard crystal frames

[sbliven]

Yes, non standard crystal frames should be fixed at the PDB level. However, they exist and include the important 4hhb case, so it would be really nice if owl were tolerant of them. Can't we just apply the scale matrix and still get eppic predictions?

[josemduarte]

As of commit af954e8 the tolerance used for comparing SCALE matrix values has been relaxed to 0.001. Thanks to that 27 structures (plus the recent one 4q8d) that had this problem are now accepted.

After this we still have a few PDBs failing the SCALE matrix comparison:

• 40 structures without a REMARK 285
• 151 structures with a REMARK 285 (see http://www.wwpdb.org/documentation/2011remediation_overview-061711.pdf for an explanation)

This is the full list:

1aat
1ana
1bab
1bbb
1bgt
1bgu
1cbl
1cbm
1chg
1coh
1d61
1d62
1dxt
1dxu
1dxv
1esc
1esd
1ese
1f47
1fdh
1gd1
1gdi
1gdj
1gdk
1gdl
1gli
1gof
1gog
1goh
1gp1
1gpd
1hba
1hbb
1hco
1hga
1hgb
1hgc
1hho
1hip
1hyh
1l01
1l02
1l03
1l04
1l05
1l06
1l07
1l08
1l09
1l10
1l11
1l12
1l13
1l14
1l15
1l16
1l17
1l18
1l19
1l20
1l21
1l22
1l23
1l24
1l25
1l26
1l27
1l28
1l29
1l30
1l31
1l32
1l33
1l34
1l35
1l36
1l37
1l38
1l39
1l40
1l41
1l42
1l43
1l44
1l45
1l46
1l47
1l48
1l49
1l50
1l51
1l52
1l53
1l54
1ldb
1llc
1lmc
1lmn
1lmo
1lmp
1lmq
1lrp
1lsp
1mbo
1mcp
1o7x
1pad
1phs
1sbt
1srd
1tgn
1thb
1tlp
1tml
1tmn
1wap
1xdw
253l
254l
255l
2bgt
2bgu
2c1c
2cab
2dhb
2gdm
2gls
2hco
2hhb
2ig2
2ldb
2ldx
2lzm
2mcp
2ovo
2pab
2pad
2pcd
2pfk
2pgk
2rmc
2sic
2sod
2taa
2tmn
3app
3dnb
3dpa
3hhb
3lzm
3pca
3pcb
3pcc
3pcd
3pce
3pcf
3pcg
3pch
3pci
3pcj
3pck
3pcl
3pcm
3pcn
3pgm
3tmn
3zs2
4cat
4cpa
4hhb
4mbn
4pad
4tln
4tmn
4tra
5csc
5dnb
5ldh
5mbn
5pad
5tln
5tmn
6pad
6tmn
6tna
7cat
7tln
8cat
9api
9ldb
9ldt

[sbliven]

Shouldn't it be possible to run EPPIC on entries which do have a SCALE matrix? Just apply the matrix to the coordinates during loading.