Re: [Rdkit-discuss] list of failed chembl ids

On Aug 8, 2017, at 22:20, Peter S. Shenkin wrote: > But I would be curious to see the 51 CHEMBL SMILES that RDKit could not parse. As of ChEMBL 23, the following files are available: - the sdf.gz file - pre-computed RDKit Morgan fingerprints in fps.gz format - the

Re: [Rdkit-discuss] list of failed chembl ids

Thank you Andrew for the explanation. I was just commenting to my summer intern that you might weigh in. Brian From: Andrew Dalke [mailto:da...@dalkescientific.com] Sent: Tuesday, August 08, 2017 15:21 To: RDKit Discuss (rdkit-discuss@lists.sourceforge.net)

Re: [Rdkit-discuss] adding fragment to existing molecule

The other answers on this thread have been right on point with the exception of neglecting to explicitly encourage you to call Chem.SanitizeMol() on your joined molecule before you do anything else with it. In your case you'd call: Chem.SanitizeMol(back) This will lead to the error that Nik

Re: [Rdkit-discuss] RD-Kit with numpy 1.13.1 on macOS(10.12.6) and Python 2.7

Hi Sai, This recently came up somewhere else too. We may be able to change the requirements for the conda package so that it supports more recent versions of numpy too. I will take a look over the next day or so and reply again on this thread with an answer -greg

Re: [Rdkit-discuss] RD-Kit with numpy 1.13.1 on macOS(10.12.6) and Python 2.7

Cool, thanks! On Tue, Aug 8, 2017 at 1:25 PM Greg Landrum wrote: > Hi Sai, > > This recently came up somewhere else too. We may be able to change the > requirements for the conda package so that it supports more recent versions > of numpy too. > > I will take a look over

[Rdkit-discuss] list of failed chembl ids

Hello, If anyone is interested, the list of chembl ids for compounds that had such crazy 2D sd files are listed below. Several are just different formulations of the same parent compound. 181880 450200 1198593 1201364 1977677 1992520 2146259 2146289 2146290 2299271 3182693 3184182 3187332

[Rdkit-discuss] RD-Kit with numpy 1.13.1 on macOS(10.12.6) and Python 2.7

Hi, I’m trying to install PyTorch( a deep learning library) in the same conda environment as RD-Kit but PyTorch requires numpy 1.13.*. The numpy version is getting downgraded if the RD-Kit is in same environment as PyTorch and I’m unable to install the PyTorch. Is there any way I can compile or

Re: [Rdkit-discuss] list of failed chembl ids

I looked up a bunch of these. The ones I saw are ChEMBL activity records, not molecule records, so they do not contain structural data. But I would be curious to see the 51 CHEMBL SMILES that RDKit could not parse. -P. -P. On Tue, Aug 8, 2017 at 3:00 PM, Bennion, Brian