Skip to content

Commit

Permalink
revision and reorganization of the "saturation"
Browse files Browse the repository at this point in the history 
Instead of global variables, the scrip now uses return values.
Extended pytests monitor the script's retention of CIP's (R)/(S)
indicators, the preservation of atomic charges and processing of
multiple strings per model by the period character.  Equally, the
potential pitfall of tin organic molecules [Sn] being processed
to yield [SN] is shut.  The documentation was revised, too.
  • Loading branch information
@nbehrnd
nbehrnd committed Apr 29, 2021
2 parents 0b4141b + b026b31 commit 029f4d3
Show file tree
Hide file tree
Showing 19 changed files with 10,052 additions and 9,834 deletions.
Binary file removed 2020-04-26_kdiff3.png
View file
Binary file not shown.
198 changes: 111 additions & 87 deletions README.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -2,109 +2,133 @@

# Background

The Bemis-Murcko scaffold\[1\] provided by `DataWarrior`\[2\] retains
information about bond order and chirality. Sometimes, however, it
suffices to retain only atom connectivity, like an assumption «there are
only single bonds». Note `DataWarrior` equally offers the export of
Bemis-Murcko skeleton, however this simplifies e.g. the scaffold about
an imidazole into one of cyclopentane.
The Bemis-Murcko scaffold<sup><a id="fnr.1" class="footref" href="#fn.1">1</a></sup> provided by
`DataWarrior`<sup><a id="fnr.2" class="footref" href="#fn.2">2</a></sup> retains information about bond order
and chirality. Sometimes, however, it suffices to retain only atom
connectivity, like an assumption «there are only single bonds».
Note `DataWarrior` equally offers the export of Bemis-Murcko
skeleton, however this simplifies e.g. the scaffold about an
imidazole into one of cyclopentane.

![img](./pattern.png)

![](./pattern.png)

# Typical use

The script runs in the CLI of Python. File `listing_file.txt` contains
the SMILES to work with:
The script runs from Python's CLI with a file listing SMILES to
process as parameter. File `test_input.smi` (from sub-folder
`test_data`) is an example:

python saturate_murcko_scaffolds.py [test_input.smi]

``` shell
python saturate_MurckoScaffolds.py [listing_file.txt]
```
This generates `test_input_sat.smi` as permanent record; the
addition of `_sat` is only a reminder of the performed saturation.
The input file is preserved.

This generates `saturated_Murcko_scaffold_sat.smi` as permanent record,
which shares the file name with the input file *with the addition of
`_sat` (like saturated)*. Regardless of the three-character file
extension of the input, the plain ASCII-list of SMILES is recorded in a
file of extension `.smi`, accepted recognized e.g., by `openbabel` to
contain this specific information.
The file extension `.smi` of the input file is a suggestion,
because it is frequently seen (e.g., around
`OpenBabel`<sup><a id="fnr.3" class="footref" href="#fn.3">3</a></sup>). Internally, the script considers any
character prior to the first period as part of the name of the
input file. The name of the report file is a mere concatenation of
this and the string `_sat.smi`.

While written for current branch of Python 3.6+ (e.g.,
Python 3.9.1), the script equally works well with the legacy branch
of Python 2 like Python 2.7.16.

While intended to work with the current branch of Python 3.6+, the
script equally works well with the legacy branch of Python 2, e.g.,
Python 2.7.16.

# Example

For a collection of organic materials, the Bemis-Murcko scaffolds was
extracted with `DataWarrior` (then release 5.0.0 for Linux, January
2019) as listing `test_input.smi.txt` including higher bond orders (see
folder `test_data`). The «artificial saturation» was triggered by

``` shell numberLines
python saturate_MurckoScaffolds.py test_input.smi
```

to yield `test_input_sat.smi`. The effect of this operation is easy to
recognize while comparing the scaffold lists (fig. *file\_diff*) in a
difference view of the two `.smi` files.

![Difference view of the SMILES strings of a Murcko scaffold *prior*
(left hand column) and *after* an «artificial saturation» (right hand
column). Note the removal of explicit bond order indicators, e.g. double
bond (equality sign), triple bond bond (octohorpe), or about implicit
aromatization (lower case → upper case for atoms of carbon, nitrogen,
oxygen (depicted); or sulfur (not depicted). At the same time,
stereochemical indicators are removed, too (e.g., slashes in lines \#018
and \#019, @-signs line \#025).](./2020-04-26_kdiff3.png "file_diff")

Subsequently, `openbabel`\[3\] was used to illustrate the work
performed. While eventually automated (cf. script `test_series.py`,
deposit in folder `test_data`), instructions delivered to `openbabel` on
the command line follow the pattern of

``` shell numberLines
obabel -ismi test_input.smi -O test_input_color.svg -xc10 -xr12 -xl --addinindex
```
For a collection of organic materials, the Bemis-Murcko scaffolds
were extracted with `DataWarrior` (then release 5.0.0 for Linux,
January 2019) as listing `test_input.smi` including higher bond
orders (see folder `test_data`). The effect of the «artificial
saturation» is easy to recognize while comparing the scaffold lists
(fig. [8](#org2006179)) in a difference view of the two `.smi` files.

![img](./diffview.png "Difference view of the SMILES strings of a Murcko scaffold *prior* (left hand column) and *after* an «artificial saturation» (right hand column). The processing affects explicit bond order indicators, e.g. double bond (equality sign, e.g., line #14), triple bond bond (octohorpe, not shown); or about implicit aromatization (lower case &rarr; upper case) for atoms of carbon, nitrogen, oxygen (depicted); or phosphorus, sulfur (not depicted). Stereochemical indicators about double bonds will be removed (e.g., slashes in lines #18 and #19). Descriptors of stereogenic centers (@-signs, e.g., line #25) are copied verbatim.")

Subsequently, `OpenBabel`<sup><a id="fnr.3.100" class="footref" href="#fn.3">3</a></sup> was used to illustrate the
work performed. While eventually automated (cf. script
`test_series.py`, deposit in folder `test_data`), instructions
issued to `OpenBabel` on the command line followed the pattern of

2 obabel -ismi test_input.smi -O test_input_color.svg -xc10 -xr12 -xl --addinindex

to generate a `.svg` file (vector representation), or

``` shell numberLines
obabel -ismi test_input_sat.smi -O test_input_sat_color.png -xc10 -xr12 -xl --addinindex -xp 3000
```

to generate a bitmap `.png`. The instructions share the definition of
data input (`-ismi`) – eventually *a list of SMILES*, rather than only
one SMILES string – and the definition of the output format based on the
extension of the file to be written (`.svg`, or `.png`). Both put the
entries into an array of 10 columns and 12 rows (`-xc`, `xr`) and label
them in the consecution of their appearance in the input files
(`--addinindex`) as inner index. Contrasting to the vector output, in
case of `.png`, an explicit scaling by a factor of 10 (`-xp` parameter)
was required to yield an intelligible illustration.

It is remarkable how well `openbabel`'s algorithm displays the molecular
structures with advanced motifs (e.g., the scaffold of cyclophane
\[entry \#33 in the synopses, `test_data` folder\], sparteine \[\#38\],
or adamantane \[\#50\]), some equally depicted in the first illustration
of this guide.
3 obabel -ismi test_input_sat.smi -O test_input_sat_color.png -xc10 -xr12 -xl --addinindex -xp 3000

to generate a bitmap `.png` with structure formulae depicted in a
grid of 10 columns by 12 rows.

It is remarkable how well `OpenBabel`'s displays the molecular
structures with advanced motifs. In addition to those shown in the
first illustration of this guide, see sub-folder `test_data` for a
more extensive survey (e.g., the scaffold of cyclophane [entry
\#33], sparteine [#38], or adamantane [#50]).


# Known peculiarities

The script neither removes, nor newly assigns SMILES descriptors
about the absolute configuration of stereogenic centers (`@`).
Thus, the «reduction» of double bonds e.g., ketones to secondary
alcohols may yield new stereogenic centers with a description
incomplete in this regard.

To resolve implicitly described aromatic systems, the script
capitalizes the characters `c`, `n`, `o`, `p`, and `s` about the
elements more frequently involved in ring formation. To avoid
ambiguity processing data about tin &#x2013; typically described by SMILES
strings as `[Sn]`, which the algorithm would convert into `[SN]` &#x2013;
*any* SMILES string including either the pattern of `[Sn]` or `[sn]`
is excluded from the reduction. This rule is applied applied to
both tin analogues of benzene, e.g. `c1[sn]ccccc1`, and SMILES
strings describing tin in a side chain (e.g., about a Stille
reagent). Script `saturate_murcko_scaffolds.py` annotates these
SMILES strings accordingly in the output file.

The script will not actively alter a charge assigned to an atom. If
present (e.g., quaternary ammonium, carboxylate), this information
will be carried over to the newly written SMILES string. Given the
reduction of bond orders, depending on the substrate submitted, this
approach may be sensible (e.g., about N in cetyltrimethylammonium
bromide), or not (e.g., about N in pyridine *N*-oxide). Other
libraries than the current script (e.g., RDKit<sup><a id="fnr.4" class="footref" href="#fn.4">4</a></sup>) might
offer help to sanitize the processed SMILES strings.

If the input SMILES string describes more than exactly one molecule
by the concatenating "`.`" (period character), this special sign
equally is the newly written SMILES string. This permits working
with SMILES about e.g., co-crystals, like about 1,4-benzoquinone and
hydroquinone, `C1=CC(=O)C=CC1=O.c1cc(ccc1O)O`.


# License

Norwid Behrnd, 2019–21, GPLv3.
Norwid Behrnd, 2019&#x2013;21, GPLv3.


# Footnotes

1. Bemis GW, Murcko MA *J. Med. Chem.* 1996, ****39****, 2887-2893,
[doi 10.1021/jm9602928](https://pubs.acs.org/doi/10.1021/jm9602928)

2. Sander T, Freyss J, von Korff M, Rufener C, *J. Chem. Inf. Model.*
2015, ****55****, 460-473,
[doi 10.1021/ci500588j](https://pubs.acs.org/doi/10.1021/ci500588j).
The program, (c) 2002–2019 by Idorsia Pharmaceuticals Ltd., is
freely available under <http://www.openmolecules.org>. For the
source code, see <https://github.com/thsa/datawarrior>.

3. [www.openbabel.org](http://www.openbabel.org). The script initially
was developed for and tested with openbabel (release 2.4.1; Nov 12,
2018) and Python 2.7.17 provided by Linux Xubuntu 18.04.2 LTS. It
equally works with Python 3.9.1+ (released January 20, 2021) and
openbabel (release 3.1.1 by January 6, 2021) as provided in
Debian 10.
<sup><a id="fn.1" href="#fnr.1">1</a></sup> Bemis GW, Murcko MA *J. Med. Chem.* 1996, ****39****,
2887-2893, [doi 10.1021/jm9602928](https://pubs.acs.org/doi/10.1021/jm9602928).

<sup><a id="fn.2" href="#fnr.2">2</a></sup> Sander T, Freyss J, von Korff M, Rufener C,
*J. Chem. Inf. Model.* 2015, ****55****, 460-473, [doi
10.1021/ci500588j](https://pubs.acs.org/doi/10.1021/ci500588j). The program, (c) 2002&#x2013;2021 by Idorsia
Pharmaceuticals Ltd., is freely available under
<http://www.openmolecules.org>. For the source code (GPLv3), see
<https://github.com/thsa/datawarrior>.

<sup><a id="fn.3" href="#fnr.3">3</a></sup> [www.openbabel.org](http://www.openbabel.org). The script initially was developed
for and tested with OpenBabel (release 2.4.1; Nov 12, 2018) and
Python 2.7.17 provided by Linux Xubuntu 18.04.2 LTS. It equally works
with Python 3.9.1+ (released January 20, 2021) and OpenBabel
(release 3.1.1 by January 6, 2021) as provided in Debian 10.

<sup><a id="fn.4" href="#fnr.4">4</a></sup> For an overview about the freely available RDKit library,
see [www.rdkit.org](https://www.rdkit.org/). An introduction into the topic of «molecular
sanitization» is provided in the section of this very title in the
on-line [RDKit Book](https://www.rdkit.org/docs/RDKit_Book.html).
Loading

0 comments on commit 029f4d3

Please sign in to comment.