unexpected data in *_drugs.tsv

[ymahlich]

Three improve_drug_ids resolve to two different canSMILES strings:

• SMI_20644
• SMI_9830
• SMI_55606

SMI_20644

improve_drug_id	canSMILES
SMI_20644	C1CN(P(=O)(OC1)NCCCl)CCCl
SMI_20644	NaN

SMI_9830

improve_drug_id	canSMILES
SMI_9830	CC1C(C(CC(O1)OC2C(OC(CC2O)OC3C(OC(CC3O)OC4CCC5...
SMI_9830	NaN

SMI_55606

improve_drug_id	canSMILES
SMI_55606	COCCOC1=C(C=C2C(=C1)C(=NC=N2)NC3=CC=CC(=C3)C#C...
SMI_55606	COCCOC1=C(C=C2C(=C1)C(=NC=N2)NC3=CC=CC(=C3)C

Three improve_drug_ids resolve to NaN canSMILES:

improve_drug_id	canSMILES
SMI_56588	NaN
SMI_9830	NaN
SMI_20644	NaN

• SMI_9830 & SMI_20644 are an overlap with the "two canSMILES" per drug_id
• SMI_56588 only resolves to NaN

Once I have figured out which dataset this information is coming from I will update the Issue.

[ymahlich]

Dataset origin for the NaN values:

see last column of the table

improve_drug_id	chem_name	pubchem_id	formula	weight	InChIKey	canSMILES	data_set_origin
SMI_56588	mek1/2 inhibitor	NaN	NaN	NaN	NaN	NaN	beataml
SMI_56588	vargetef	NaN	NaN	NaN	NaN	NaN	beataml
SMI_56588	baiclein	NaN	NaN	NaN	NaN	NaN	beataml
SMI_9830	3-((2,6-dideoxy-4-o-[2,6-dideoxy-4-o-(2,6-dide...	NaN	NaN	NaN	NaN	NaN	mpnst
SMI_20644	n,3-bis(2-chloroethyl)tetrahydro-2h-1,3,2-oxaz...	NaN	NaN	NaN	NaN	NaN	mpnst

canSMILES for SMI_55606:

The outlier that produces the second canSMILES string is from the mpnst dataset (see row 8 / second last row)

improve_drug_id	canSMILES	data_set_origin
SMI_55606	COCCOC1=C(C=C2C(=C1)C(=NC=N2)NC3=CC=CC(=C3)C#C...	beataml
SMI_55606	COCCOC1=C(C=C2C(=C1)C(=NC=N2)NC3=CC=CC(=C3)C#C...	ccle
SMI_55606	COCCOC1=C(C=C2C(=C1)C(=NC=N2)NC3=CC=CC(=C3)C#C...	ctrpv2
SMI_55606	COCCOC1=C(C=C2C(=C1)C(=NC=N2)NC3=CC=CC(=C3)C#C...	fimm
SMI_55606	COCCOC1=C(C=C2C(=C1)C(=NC=N2)NC3=CC=CC(=C3)C#C...	gcsi
SMI_55606	COCCOC1=C(C=C2C(=C1)C(=NC=N2)NC3=CC=CC(=C3)C#C...	gdscv1
SMI_55606	COCCOC1=C(C=C2C(=C1)C(=NC=N2)NC3=CC=CC(=C3)C#C...	gdscv2
SMI_55606	COCCOC1=C(C=C2C(=C1)C(=NC=N2)NC3=CC=CC(=C3)C	mpnst
SMI_55606	COCCOC1=C(C=C2C(=C1)C(=NC=N2)NC3=CC=CC(=C3)C#C...	prism

[ymahlich]

as per Ryan the SMILES string for SMI_56588 should be C1=CC=C(C=C1)C2=CC(=O)C3=C(O2)C=C(C(=C3O)O)O. Once the issue has been fixed on the build backend we can check to make sure that SMI_56588 is populated accrodingly.