Utilizes 100% of tumor genome (SNV, CNV, genomic scars, structural rearrangements and more), professional expertise in big data analysis and AI approach combined with clinical information to dramatically improve patient stratification for cancer therapies.
The most powerful
genomic biomarker discovery platform
GEMMAai™
GEnoMic Marker Analyzer
AI-Powered platform for matching patients with right treatment
Most clinical trials fail
Most cancer treatments have low efficacy
GEMMAai™ PLATFORM
Our proprietary software
for data mining
5410
whole-genome tumor data
153
genomic patterns (features)
9
AI-powered assays
Our pipeline
Our AI-powered methodology can be used to build a predictive model for any drug-response diagnostic tool, including PARP, CDK4/6 and mTOR inhibitors. The best results are obtained when both tumor whole genome data and clinical informations are available. However our extensive tumor genome database allows for diagnostics development for drugs where no clinical data is available (mechanism of action for a drug is sufficient).
GEMMAai for clinical trial enrolment – available
GEMMAai for companion diagnostics development – available
GEMMAai for patients coming soon – pipeline *
TUMOR
TYPE
TUMOR GENOMES
TARGETED THERAPY
DISCOVERY
PROTOTYPE ASSAY
ANALITYCAL VALIDATION
CLINICAL VALIDATION
REGULATORY APPROVAL
UPCOMING
MILESTONES
BREAST
2152
CDK4/6i
Ant-HER2
PARPi
mTORi
PIK3i
2 analytically validated diagnostic tools in Q1 2021
OVARIAN
512
Carboplatin / PARPi
FDA breakthrough device submission Q2 2021
FASETM
GENOMIC FEATURE DATABASE
YOUR inhibitor
Discovery and validation
AI-POWERED
STRATIFICATION DIAGNOSTIC TOOL
* For Research Use Only. Not for use in diagnostic procedures.
TUMOR TYPE: OVARIAN
WGS DATA
512
TARGETTED THERAPY
UPCOMING MILESTONES
FDA breakthrough device in Q3 2021
TUMOR TYPE: BREAST
WGS DATA
2152
TARGETTED THERAPY
UPCOMING MILESTONES
2 analytically validated diagnostic tools in Q1 2021
TUMOR TYPE: LUNG
WGS DATA
1852
TARGETTED THERAPY
UPCOMING MILESTONES
2 prototype assays in Q3 2021
FUSETM
WGS DATA
GENOMIC FEATURE DATABASE
TARGETTED THERAPY
YOUR inhibitor
DISCOVERY AND VALIDATION
AI-POWERED STRATIFICATION DIAGNOSTIC TOOL
Individual "Decision Tree" Construction
Random Forest Artificial Intelligence methodology based on Training Dataset
INSTEAD OF >1 000 SAMPLES, GEMMAaiTM
requires only 100 training datasets for a reliable output and achieves the optimal balance of sensitivity (recall) and precision.
Discrimination threshold
Learning curves
CARBOPLATIN/PARPi CLASSIFIER
_ALICE
ArtificiaL Intelligence ClassifiEr
NEW GLOBAL STANDARD IN OVARIAN ONCOLOGY THROUGH CONNECTING AI TECHNOLOGY TO WHOLE-GENOME DATA
Our proprietary software
for data mining
_ALICE is a Random Forest model, consisting of 100 decision trees. It analyses 44 genomic features associated with ovarian cancer treatment response. Apart from ‘gold standard’ biomarkers for HRD like BRCA1/2 mutations, deletions with microhomologies and tandem duplications, we have identified 3 novel biomarkers correlated with HRD and 22 important genomic biomarkers associated with PARPi clinical response.
Biomarker imporance
3
novel biomarkers correlated with HRD
22
novel important genomic biomarkers associated with PARPi clinical response
19
‘gold standard’ biomarkers
TECHNICAL NOTE
Read our tech note to learn how we constructed the _ALICE classificatory – from whole-genome data mining, feature extraction and engineering, through AI model development and retrospective validation.
Download
