DNA-encoded library screening
A library of 343 million synthesised, DNA-encoded compounds screened against a target in a single affinity selection — chemical space that plate-based screening could not sample in years.
Biopharmaceutical R&D
Finding the hits conventional screening misses
Chemtelliq screens a DNA-encoded library of 343 million synthesised compounds, identifies the genuine hits with its own AI models, and optimises them through structure-based design and bench validation.
What we do
Three business areas
One discovery engine feeding two development modalities.
Peptide-drug conjugates (PDC)
Targeting peptides joined to a payload through a tuned linker, so the drug concentrates where it is needed and the therapeutic window widens.
Targeted protein degradation (TPD)
PROTAC and degrader chemistry that recruits the cell’s own ubiquitin–proteasome machinery to remove a disease-driving protein rather than merely block it — reaching targets with no druggable active site.
DNA-encoded library
Every compound carries a DNA barcode recording how it was made. That lets 343 million of them meet a target in a single tube: what binds is kept, what does not is washed away, and sequencing reads back the answer.
Peptide-drug conjugates
The peptide finds the receptor, the linker holds the payload until it should not, and the payload does the work. The conjugate stays intact in circulation and comes apart only inside the target cell.
Targeted protein degradation
A degrader recruits the cell’s own disposal machinery to remove a protein rather than block it — and is released to do it again. Induced proximity can also disable a protein outright, acting only where the tumour-specific partner exists.
How we work
Prediction and experiment, one team
Physical chemistry, machine learning and bench work sit on the same clock. That loop is the platform.
Structure & docking
Our physical chemists model protein structures and predict molecular docking to explain how a hit binds — then modify it, round by round, through lead optimisation.
AI hit identification
Machine learning and deep learning models read the selection data from a 343-million-compound library and separate genuine binders from artefacts — the step that decides what is worth making.
Wet-lab execution
Protein production, binding characterisation and cell-based validation are run by our own team. Results come back in days, and every result — including the negatives — retrains the model.
Looking for a discovery partner?
We collaborate with pharmaceutical companies, academic groups and investors on targets that conventional screening has struggled with.