openmolecules.org Forum: Functionality

Home » DataWarrior » Functionality » name vs score 2D plot

Show: Today's Messages :: Polls :: Message Navigator

name vs score 2D plot [message #1859]

Mon, 06 March 2023 08:36

sublimeuser
Messages: 6
Registered: June 2021

Junior Member

Hi everyone,

Just wanted to know how to represent a 2D plot - I have 2 columns, one containing scores and the other compound names.

I just wanted to know how to represent a simple 2D plot with names on the x axis and scores on y axis. In 2D view, I only see the score column assigned.

I'm sure this is a very simple thing to do, but for some reason it escapes me.

Many thanks!

Report message to a moderator

Re: name vs score 2D plot [message #1860 is a reply to message #1859]

Mon, 06 March 2023 09:57

nbehrnd
Messages: 224
Registered: June 2019

Senior Member

Hello sumblimeuser,

with the 2D plot already created, there will be a button «xy» (next to the wrench symbol). This allows to select the columns from the array / spread sheet, and if the corresponding columns possess a header line, this will be used / updated to annotate the 2D plot, too.

Regards,

Norwid

Attachment: column_selection.png
(Size: 14.99KB, Downloaded 143 times)

Report message to a moderator

Re: name vs score 2D plot [message #1861 is a reply to message #1859]

Mon, 06 March 2023 15:16

sublimeuser
Messages: 6
Registered: June 2021

Junior Member

Hi Norwid and thank you for the reply!

That's just the problem: while one of the columns is indeed assigned, the other one is unassigned (compound names). How do I assign the name column as well?

Many thanks!

[Updated on: Mon, 06 March 2023 17:04]

Report message to a moderator

Re: name vs score 2D plot [message #1862 is a reply to message #1861]

Mon, 06 March 2023 20:10

nbehrnd
Messages: 224
Registered: June 2019

Senior Member

Hi Sublimeuser,

access the representation of the array / the spread sheet. Mark the (empty) column header in question (left mouse click), then open the pull-down menu available by a right mouse click; select «Set Column Alias».

Regards,

Nprwid

Report message to a moderator

Re: name vs score 2D plot [message #1864 is a reply to message #1859]

Mon, 06 March 2023 22:14

sublimeuser
Messages: 6
Registered: June 2021

Junior Member

Defined new aliases for both columns. Unfortunately it still does not work. Sad

Report message to a moderator

Re: name vs score 2D plot [message #1867 is a reply to message #1864]

Tue, 07 March 2023 15:52

nbehrnd
Messages: 224
Registered: June 2019

Senior Member

Hi sublimeuser,

is it possible for you to share a typical, yet not confidential .sdf presenting the problems to you? I speculate perhaps the problem correlates with the syntax in the input file.

In an instance of Linux Debian 12/bookworm with openbabel (vesion 3.1.1) I created two .sdf files where the «compound name» is the standard InChI key

``` shell
$ obabel -:"O" -:"CCOCC" -:"c1ccncc1" -h --gen3d --append "inchikey" -O testV2000.sdf
$ obabel -:"O" -:"CCOCC" -:"c1ccncc1" -h --gen3d --append "inchikey" -x3000 -O testV3000.sdf
```

This is to cover both the elder V2000 format, as well as the contemporary V3000 one DW prefers on file I/O as .sdf. The reading from the CLI was in pattern of

``` shell
$ datawarrior ./testV3000.sdf &
```

run without noticeable difficulty, nor the adjust the header "compound name" by "InChI key" - there was a prompt (automatic) update in the 2D and 3D plot on abscissa and ordinate.

The four tests run DataWarrior 5.5.0 in said Debian, either the one provided by a pristine installation with the installer fetched by 2022-05-18 (i.e., the state by 2021-04-03), or by subsequent application of the update fetched today (2023-03-07) about the state reached by 2024-02-24.

With regards,

Norwid

Attachment: testV2000.sdf
(Size: 2.87KB, Downloaded 206 times)
Attachment: testV3000.sdf
(Size: 2.25KB, Downloaded 208 times)
Attachment: testV2000.dwar
(Size: 2.85KB, Downloaded 212 times)

[Updated on: Tue, 07 March 2023 15:53]

Report message to a moderator

Re: name vs score 2D plot [message #1868 is a reply to message #1867]

Wed, 08 March 2023 11:47

sublimeuser
Messages: 6
Registered: June 2021

Junior Member

Hi nbehrnd!

As an input, I'm simply using a .csv file with 2 columns (with headers) -> Ligand name and Docking Score. Could this be the source of the problem?

Thanks!

[Updated on: Wed, 08 March 2023 11:48]

Report message to a moderator

Re: name vs score 2D plot [message #1869 is a reply to message #1868]

Wed, 08 March 2023 14:36

nbehrnd
Messages: 224
Registered: June 2019

Senior Member

Hello,

reading a tabulator separated, two column ASCII file by DW (by Ctrl + O) did not yield a problem; the 2D plot generated on the fly was there, change of the column's alias provided the instantaneous update of the 2D plot, interchange of the selection for abscissa and ordinate of the 2D plot: so far, no replication of the problem you report.

Regards,

Norwid

Attachment: test_tsv.txt
(Size: 0.11KB, Downloaded 164 times)
Attachment: test_tsv.dwar
(Size: 1.90KB, Downloaded 215 times)

Report message to a moderator

Re: name vs score 2D plot [message #1870 is a reply to message #1869]

Wed, 08 March 2023 15:55

sublimeuser
Messages: 6
Registered: June 2021

Junior Member

Hi nbehrnd,

It works perfectly if I produce a test file with 5 ligand entries. If fails however,with the large number of rows in my original file (>300.000) so I suspect that DataWarrior is simply not able to plot such a large number of entries.

Thank you for your help once again!

Report message to a moderator

Re: name vs score 2D plot [message #1871 is a reply to message #1870]

Wed, 08 March 2023 21:32

nbehrnd
Messages: 224
Registered: June 2019

Senior Member

Hi sublimeuser,

a set of 300k indeed might a bit large (for DW). In such a case I recommend to have a look into AWK. Though initially written for text processing, this «Swiss pocket knive» understands some mathematics you can use to filter by thresholds. Based on the assumption your raw data file is organized as two-column ASCII like `test_tsv.txt` with docking score in the second column, it can be used e.g.

+ to report only the data with an entry in the second column higher than 0.2:

``` shell
$ awk '{if ($2 > 0.2) print}' test_tsv.txt
```

+ to report only the data where the second column's entries are in the in the interval between 0.2 and 0.8:

``` shell
$ awk '{if ($2 > 0.2 && $2 < 0.8) print}' test_tsv.txt
```

which you can redirect into a permanent record either by overwriting the old content (`>`), or by append (`>>`). In case of access to an installation of Linux, you can combine this with a line count (`wc -l`) you either can run on the newly written record

``` shell
$ awk '{if ($2 > 0.2 && $2 < 0.8) print}' test_tsv.txt > records.txt && wc -l records.txt
2 records.txt
```

or pipe

``` shell
$ awk '{if ($2 > 0.2 && $2 < 0.8) print}' test_tsv.txt | wc -l
2
```

Though this equally removes the headers, the newly written (filtered) record files should be less resource hungry and hence accessible to DW (Edit -> Paste Special -> Paste Without a Header Row).

With regards,

Norwid

Report message to a moderator

Re: name vs score 2D plot [message #1875 is a reply to message #1871]

Sat, 18 March 2023 17:55

thomas
Messages: 715
Registered: June 2014

Senior Member

Hi Norwid and Sublimeuser,

there are a couple of defined maxima applied to graphical views in DataWarrior to prevent an explosion of in-memory-data. Usually, these limits are meant to be high enough to not restrict any reasonable graph drawing. For instance category columns can be assigned to an axis if they have not more than 256 distinct categories (your compound name column should be classified as category column). The idea was that more than 256 distinct objects on an axis cannot be displayed well anymore. Admittedly, 256 seems very low, especially considering that the limitation stays in place when zooming in. I will increase the limit to 10000.

Thomas

Report message to a moderator

Previous Topic:	HBD/HBA
Next Topic:	Opening .CSV file

Goto Forum:

-=] Back to Top [=-

[ Syndicate this forum (XML) ] [

] [

]

Current Time: Thu Nov 21 23:59:35 CET 2024

Total time taken to generate the page: 0.04822 seconds