Converting ChEMBL to sequence

This page gives you access to a subset of ChEMBL that has been converted to Protein Line Notation. A total of 27142 structures out of 39123 "peptide-like" structures have been successfully converted. More details on the conversion process can be found beneath the structure table.

ChEMBL ID contains

Search is done within records with converted sequences only. Result listing is limited to max. 100 records.

Refresh page with an empty search box to get 100 random structures.

20 randomly selected converted structures

ChEMBL ID	ChEMBL structure	Converted sequence image	Proteax PLN (Protein Line Notation)
CHEMBL1170034			H-FKP{d}[Res_2201][Res_2201]{d}R-[NH2] name=CHEMBL1170034
CHEMBL3112678			H-KTNMKGMAGAM-OH name=CHEMBL3112678
CHEMBL2326567			H-[Res_2006]RLIF-[NH2] name=CHEMBL2326567
CHEMBL2371631			H-RPPGF{d}[Res_1788][Res_594]FR-OH name=CHEMBL2371631
CHEMBL67596			H-{d}AL{d}DAC{d}ED-[NH2] name=CHEMBL67596
CHEMBL1342083			[acetyl]-[Res_2684]GG-OH name=CHEMBL1342083
CHEMBL2147249			[acetyl]-EYFE[Res_1006]-[CTerm_1155] name=CHEMBL2147249
CHEMBL2372965			[NTerm_887]-VPV-[CTerm_707] name=CHEMBL2372965
CHEMBL2115556			(cyclo)-P[Nle]LG[Res_3000]V-(cyclo) name=CHEMBL2115556
CHEMBL418486			[acetyl]-SDK-OH name=CHEMBL418486
CHEMBL291602			[NTerm_820]-QQ{d}W-[NH2] name=CHEMBL291602
CHEMBL2333091			[NTerm_701]-[MeVal][MeVal][MeVal][Res_1700]{d}[Res_733]-[CTerm_450] name=CHEMBL2333091
CHEMBL1178333			H-[Res_2417]YIL-OH name=CHEMBL1178333
CHEMBL336553			[NTerm_68]-{d}F[Res_2350]R-[CTerm_1150] name=CHEMBL336553
CHEMBL598408			H-T{d}[Res_895]T-OH name=CHEMBL598408
CHEMBL386271			[acetyl]-[Nle]EHFRWGKPV-[NH2] name=CHEMBL386271
CHEMBL2393408			H-TFLAEDPKVTLTIFV-OH name=CHEMBL2393408
CHEMBL1814489			H-AVL-OH name=CHEMBL1814489
CHEMBL1784381			(cyclo)-SV[Res_733]GV[Res_1467]V[Res_1467]V{d}[Res_1700]GI[Res_678][MeVal]-(cyclo) name=CHEMBL1784381
CHEMBL2373264			H-SSC(1){d}FGGRIDRIG{d}AQ{d}SG{d}LG{d}C(1)NSFR-[NH2] name=CHEMBL2373264

Conversion process

All structures in the ChEMBL 19 database were downloaded as an SD file and, with the help of KNIME, probable "peptide-like" structures were identified. The "peptide-like" structures were defined as those containing a substructure of three connected glysines. This yielded a "peptide-like" subset of 39123 structures.

The peptide subset was loaded into a PostgreSQL database table and the Biochemfusion Proteax cartridge was used to convert structures, when possible, to sequences. The first conversion took 87 seconds and produced 27142 converted sequences.

You can download the full set of produced sequences (TAB-separated file with ChEMBL ID + PLN, ~11MB).

All found unknown residues and terminal structures were embedded as inline structures in the first round of produced PLN. The embedded structures were then de-duplicated and extracted. You can download the structure sets in gzip-ed SD file format from here:

N-terminals, 1699 structures, 420 kB
C-terminals, 1270 structures, 333 kB
Residues, 3030 structures, 726 kB

143 of the unknown residue structures have been assigned well-defined names by NextMove Software's great Sugar & Splice tool (press the "Biologics" button). Many thanks to Roger Sayle at NextMove Software for processing this subset of residues with Sugar & Splice.

NextMove's residue names can be applied to a Proteax modification database (where the above SD files have been imported) by the following SQL script:

Residue names update script, 143 name assigments, 18 kB

The currently produced PLN has been generated after all the above data has been loaded into Proteax's database of known structures. As you will see, most of the non-natural residues and terminals have auto-generated names based on sequential numbers.