Explore Rancho Data Science Services
Rancho BioSciences provides data curation services for pharmaceutical and biotech companies, as well as for academic institutions, foundations, and the government. Rancho works with different life science data types, including clinical trials, genomics, gene variants, assays, chemistry, microbiome, flow cytometry, and imaging data. The data can be internal and/or public. Rancho BioSciences is platform-agnostic and has a lot of experience formatting data for many commercial, internal, and public (open-source) platforms.
AI, ML, and NLP Services
Rancho BiSciences full-cycle AI applications include data preparation, model training, and insights generation. We customize our data science workflows to address unique research needs, including support for clinical trials, dealing with unstructured text and multi-dimensional omics datasets. By analyzing large volumes of data using state-of-the-art AI models, our clients gain valuable insights that can be used to accelerate scientific progress, improve research efficiency, and drive innovation.
At Rancho BioSciences we empower businesses with a comprehensive suite of AI services through an expertly-crafted synergy of domain experts, data scientists, and data engineers. Our full-cycle AI solutions span every essential aspect – from data collection to robust model training, seamlessly merging cutting-edge technology and industry-specific knowledge. Together, we unlock the potential hidden within your data, enabling you to accelerate growth and drive innovation.
Get training data
Data curation is our core business. We work across all life science data types, are platform-agnostic and have robust manual and automated workflows to extract data from public or private sources, harmonize it, QC and prepare high quality machine-readable datasets for training or benchmarking your AI/ML algorithms.
Examples of training datasets include datasets for target liability, clinical trial patient cohorts, reagents (cell line, antibody, etc.), gene- disease associations, perturbation datasets.
Training AI/ML
-
Ranchos BioSciences team can identify relevant entities and attributes of data that are important for training and optimizing these algorithms.
-
Rancho BioSciences has experienced data scientists and SMEs who can build tailor-made ML/AI models:
- Predictive toxicology
- Survival analysis
- Cellular phenotype classification
- Disease signature analysis
-
Rancho BioSciences validates performance of AI/ML algorithms to ensure accuracy and reliability.
Applying AI to get insights
-
A combination of large language models (LLM) and classical NLP techniques is used to extract valuable information from text, e.g. pathology reports.
-
An embedding-based publication scoring algorithm enabled scientists to easily find relevant information based on incomplete information.
-
Rancho terminology mapping solution uses embedding, LLM and fuzzy to construct a semantic layer and enrich private datasets.
-
We are building Natural Language Query (NLQ) applications for our customers with AI-generated queries and scoring in diverse corporate data environments
AI, ML, and NLP Services
Rancho BioSciences’ full-cycle AI applications include data preparation, model training, and insights generation. We customize our data science workflows to address unique research needs, including support for clinical trials, extracting information from unstructured text, and processing and integrating multi-dimensional omics datasets. By analyzing large volumes of data using state-of-the-art AI models, our clients gain valuable insights that enable them to accelerate scientific progress, improve research efficiency, and drive innovation.
Our comprehensive suite of AI services empowers businesses by providing them with an expertly crafted synergy of domain experts, data scientists, and data engineers. Our full-cycle AI solutions span every essential part of the project — from data collection to robust model training — seamlessly merging cutting-edge technology and industry-specific knowledge. Together, we unlock the potential hidden within your data, enabling you to accelerate growth and drive innovation.
Get training data
Data curation is our core business. We work across all life science data types, are platform-agnostic, and have robust manual and automated workflows to extract data from public or private sources. We can then harmonize the data, run it through a rigorous QC protocol, and prepare high quality machine-readable datasets for training or benchmarking your AI/ML algorithms.
Examples of training datasets include datasets for target liability, clinical trial patient cohorts, reagents (cell line, antibody, etc.), gene-disease associations, and perturbation.
Training AI/ML
-
Ranchos BioSciences’ team can identify relevant data entities and attributes that are important for training and optimizing these algorithms.
-
Rancho BioSciences has experienced data scientists and SMEs who can build tailor-made ML/AI models, such as for:
- Predictive toxicology
- Survival analysis
- Cellular phenotype classification
- Disease signature analysis
-
Rancho BioSciences validates performance of AI/ML algorithms to ensure accuracy and reliability.
Applying AI to get insights
-
A combination of large language models (LLM) and classical NLP techniques is used to extract valuable information from text (e.g., pathology reports).
-
An embedding-based publication scoring algorithm enables scientists to easily find relevant information based on incomplete information.
-
Rancho terminology mapping solution uses embedding, LLM, and Fuzzy to construct a semantic layer and enrich private datasets.
-
We are building Natural Language Query (NLQ) applications for our customers with AI-generated queries and scoring in diverse corporate data environments.
Bioinformatics
Our Bioinformatics team can help answer scientific questions by analyzing your complex data. Whether you need a short-term boost in data analysis resources or engage a specialized team for a long-term project, Rancho can support your requirements with projects including:
-
Single-cell transcriptomic analysis in R and Python. Routine analysis includes QC, dimensional reduction, clustering, and differential expression. In-depth cell profiling with atlas-based cell annotation, trajectory inference, and RNA velocity.
-
Processing and analysis of multimodal data. Gene expression, proteomics, and metabolomics differential expression/abundance.
-
DNA or histone binding site analysis (ChIP-seq)
-
Genetic variant detection (single nucleotide variants, insertions/ deletions, copy number variants, genome structure variants)
-
Association of genetic variants or gene expression profiles with disease, drug response, or clinical outcome
-
Development of predictive models using machine learning or logistic regression
-
Pathway analysis and functional enrichment analysis
-
Phylogenetics and sequence analysis
By processing and integrating disparate data types, we can generate a comprehensive biological picture of a disease. This further informs practical biotech and pharmaceutical research and development, including applications in:
Precision Medicine:
Matching drug profiles and human genetic profiles, stratification of patient cohorts, biomarker identification
Drug discovery:
Mechanism of action studies, accelerating discovery pipelines by intelligently screening results, pharmacology applications
Translational medicine:
Drug repositioning for new indications, companions, and diagnostics, enabling cross-discipline communication
Rancho BioSciences provides expert data curation services to support biomedical research, drug discovery, and precision medicine.
Data Curation Services
Why is Rancho Biosciences different:
-
Expertise: Others crowd source or hire temporary curators while Rancho has a team of experienced, trained Ph.D. and M.D. scientists and curators who get the tools they need to deliver results.
-
Innovation: Rancho Biosciences is committed to staying at the forefront of biomedical data science and technology, with a focus on innovation and continuous improvement, such as supporting emerging AI/ML trends in the industry
-
Quality: delivers high-quality services and data that are accurate, reliable, and relevant, with a focus on quality control and validation. We have a strict QC process that is rigorous and is based on years of our collective experience. We develop QC tools to help us, and we often are called to arbitrate or review what other vendors provided.
-
Flexibility: Rancho Biosciences provides flexible and customized solutions that are tailored to the specific needs of each customer. By collaborating closely with clients to gain a deep understanding of their needs and goals, our team develops fit-for-purpose solutions that are unique to each project. We do not believe in a one-size-fits-all approach, and strive to deliver high-quality, customized services that exceed our clients' expectations.
At Rancho BioSciences, we cover nearly all types of Life Sciences data. Our curators have expertise in curating, annotating and enriching:
-
Genomic data: DNA sequencing data, gene expression data, and genomic variations
-
Clinical data: clinical data from clinical trials, biobanks, electronic health records (EHRs), and other sources, including patient demographics, diagnoses, procedures, and medications
-
Imaging data: medical and R&D imaging data, including radiology images, MRI scans, and other types of imaging data used in clinical practice or research
-
Omics data: large-scale datasets generated by high-throughput technologies in various areas of the life sciences, such as genomics, proteomics, metabolomics, and transcriptomics
-
Assay data: experimental data generated by various types of laboratory assays, such as biochemical assays, cell-based assays, and high-throughput screening assays including CRISPR screens, Cell Painting assays and other.
-
Pathway data: data on biochemical and molecular interactions in biological pathways, protein-protein interactions
-
Chemical data: chemical structures, chemical properties, and chemical interactions
-
Pharmacological data: drug targets, drug interactions, and drug metabolism (ADME, PK/PD, toxicity, metabolomics, and others.)
-
Other types of life science data: environmental data, microbiology, toxicology, etc.
Rancho Biosciences uses automated and manual curation services:
-
AI/ML/NLP development and application:
- Rancho BioSciences prepares high quality machine-readable datasets for training of your AI/ML algorithms.
- Rancho BioSciences validates performance of AI/ML algorithms to ensure accuracy and reliability.
- Ranchos BioSciences team can identify relevant entities and attributes of data that are important for training and optimizing these algorithms.
-
Data Harmonization: This involves identifying and resolving differences in data formats, structures, and semantics between different data sources, and creating a common data model that allows for easy integration and analysis of the data.
-
Aligning data to CDISC (SDTM, SEND and other domains), OMOP and other standard data models.
-
Quality assurance: ensuring integrity of data delivered by other vendors.
- Rancho Biosciences scientists are life science domain experts who help to guide AI/ML development and data interpretation.
-
Terminology services:
- Semantic data integration: Rancho Biosciences can integrate heterogeneous data sources using semantic technologies and standards.
- Customized terminology and ontology services: Rancho Biosciences can develop customized terminology and ontology services to meet the specific needs of clients, such as building a custom ontology for a specific project or application.
- Nomenclature mapping and cross-referencing: Rancho Biosciences can map and cross-reference data to standard nomenclatures, such as LOINC or SNOMED, to ensure consistency and interoperability.
- Training and support services: Rancho Biosciences can provide training and support services for clients to help them understand and effectively use biomedical terminologies and ontologies in their research or clinical practice.
- Data enrichment: Rancho Biosciences can enrich data with additional information, such as synonyms, cross-references, and metadata, to improve data quality and usability.
Our team at Rancho BioSciences has developed robust and reproducible curation workflows that utilize both automated and manual approaches. These workflows are designed to ensure that the data is accurate, complete, and of the highest quality. The team has extensive experience in curating various types of data, and their expertise allows them to tailor their workflows to the specific needs of each project.
Data Engineering
Rancho combines deep scientific knowledge and technical expertise in software engineering to provide both custom and off-the-shelf solutions.
Engineering services include, but not limited to:
-
Pipelines and Process Automation
-
Data Management and Migration Solutions
-
Data Integration
-
Data Visualizations
-
Custom Analytical Tools
-
FTEs To Assist Teams With Software Engineering Needs
-
Technical Consulting
Rancho leverages its managed toolkit to customize solutions that meet client needs while cutting down on delivery time and cost.
Data Imaging
Data Imaging Services
Whether you would like to increase the value of your medical imaging data through better annotations or extract new insights via analysis, Rancho’s scientists can provide curation and analytical support across several types of imaging modalities, such as neuroimaging and digital pathology, and data from a wide range of therapeutic areas. Rancho scientists have broad expertise in curation and analysis of different types of imaging data and have processed data from a range of platforms, such as UK Biobank, ADNI, AMP-AD, and PPMI.
Rancho can provide high-quality imaging data services that include:
-
Data Modeling
-
Data Aggregation and Curation
-
Data Analysis & Reports
1. Imaging data model
Data Modeling
Medical imaging creates vast amounts of data that need to be properly stored and organized to maximize its value. Rancho Biosciences can create conceptual, logical, or physical data models for imaging data that leverage existing models and can be customized to your needs. Our scientists are experienced in data modeling and have subject matter expertise in different imaging modalities. When they create a model, they consider the complexity of your imaging data. Rancho Biosciences can build a stand-alone model or embed it in a larger “omics” model, linking the data to other collected biospecimens.
2. Digital pathology entity examples
Data Curation
For accurate interpretation of medical images across datasets, the images need to be well documented, and their metadata standardized. With expertise in this area, Rancho scientists can ensure that data meets F.A.I.R. data standards. Rancho scientists have experience in the aggregation, automated extraction, and integration of imaging data such as H&E, single biomarker or multiplex IHC, structural and functional MRI, and PET. Our scientists have both clinical and technical knowledge that will help maximize the value of your imaging data. Rancho scientists have experience working with clinical data and research data from a variety of diseases such as Parkinson’s Disease, Alzheimer’s Disease, Traumatic Brain Injury, Huntington’s disease, and from oncology and immunology studies, and are well-versed in applying imaging data standards. Their understanding of the study process e.g., data lineage from biospecimen collection to scanning of an image, will ensure that the value of your annotations is maximized.
3. Digital pathology entity & attribute examples (sample lineage)
4. Example of MRI preprocessing pipeline comparison
5. Confounder evaluation MRI data
Data Management, Data Governance, and Data Modeling – What Is Data Governance, Data Modeling, and Data Management?
Data management is an IT practice encompassing an organization's practices across the data life cycle.
Data governance is a business practice that defines how data is processed across the organization to ensure that data is FAIR (compliant, private, etc.).
Data modeling is a process of creating conceptual, logical, or physical data models for one’s data.
Life sciences companies accumulate data with an ever-increasing pace, and data that they use to create new treatments or new diagnostics is more diverse and complex than in arguably any other domains. With multiple data sources, data elements, stakeholders, and data consumers, it gets progressively harder and harder to scale and to maintain control over the data life cycle. There is a strong awareness in the scientific community that more data must produce more insight and knowledge, but this goal has been difficult to attain. Life Sciences businesses are now relying on the practice of data management and data governance as best business practices to wrangle data assets and realize value from data.
There are many resources that deliver data management and data governance solutions. There are frameworks, platforms, and solutions; there are tools that one can use and knowledge bases to consult with. We are different. We provide people—scientists with subject matter expertise trained in data modeling—who can help. Our people have dozens of years of experience in their respective fields of study and participated in multiple projects where they did data modeling or provided data governance support. We have real-life experience, and we can support your projects with minimal ramp-up time by leveraging data models we have already built and know-how.
Data Science
Rancho BioSciences combines advanced data science methods with life science expertise to provide comprehensive data solutions. Our team is highly experienced in statistical modeling, machine learning, natural language processing, and other data science techniques. At the intersection of life science and data science, our experts deliver tailored solutions to unlock your data’s potential, keeping our clients at the forefront of scientific exploration and development.
Our full-cycle AI applications include data preparation, model training, and insights generation. We customize data science workflows to address your unique research needs, including support for clinical trials and navigating unstructured text and multi-dimensional omics datasets. By analyzing large volumes of data using state-of-the-art AI models, our clients gain valuable insights to accelerate scientific progress, improve research efficiency, and drive innovation.
We can help you with:
-
FAIRification of large datasets, transforming data from disparate sources and formats into Findable, Accessible, Interoperable, and Reusable data
-
Interactive data visualization, including R-shiny, to help you explore your data and develop hypotheses
-
Building custom AI/ML models and conducting statistical analyses to address your needs
-
Extracting information and gaining insights from unstructured or poorly structured private or public sources
Knowledge Mining
Our team of highly qualified Ph.D. and M.D. scientists can help you collect information to support your projects. If you have a need to assess a new target, or compile competitive intelligence, or assemble a dataset of toxicity endpoint – we can help. Our expertise includes:
-
Data gap analysis – we will find what is missing in your data collection and come up with ideas to fill in those gaps.
-
Crawling public resources to identify relevant datasets for your research needs. A custom-designed Crawling tool efficiently queries NCBI and EMBL resources and provides results in a user-friendly Excel format.
-
Developing custom scoring systems and generating profile reports for target prioritization.
-
Assessing therapeutic landscape to understand unmet needs, patient populations, competitors to make informed decisions about product development.
-
Building knowledge bases for various life science domains, including diseases, genetic variants, microbiomes, drugs, assays and other.
-
Extracting relevant information from publications such as interactions, assays, drug activity and toxicity, clinical trials, phenotypes, genetic variants, expression signatures, adverse events, biomarkers, patient population, and other domains.
-
Gathering information for drug repurposing.
-
Assisting with panel development by creating comprehensive lists of relevant genes based on literature and public databases and gathering information to facilitate gene prioritization.
The team's knowledge mining abilities are unparalleled, and their dedication to understanding each customer's unique requirements allows them to provide tailored solutions that deliver high-quality, actionable insights.
What Can Be Delivered?
Data Models:
Delivered as ERDs, in many cases with data dictionaries. We use a variety of tools to deliver data models, LucidCharts and ER studio by IDERA being the most common.
Data Governance as a Process:
We provide scientists trained in data modeling and experienced with data governance projects to support your team operationally. We bridge scientists and IT departments and provide oversight on the data governance process at your organization. We develop community outreach to educate your team on best practices. We write SOPs, post them, and check that people adhere to the recommendations
Documentation, Strategy Documents, Investigation of Third-Party Tools:
We help clients develop short and long data strategy. Often, young companies focus on research and work hard to develop a new drug or a diagnostic—and find themselves in need to quickly become ready for a regulatory finding, but they lack the skillset in-house to approach this stage of growth. Rancho team can evaluate the current state and find solutions for future data state 0- tailored to the specific needs of the client. We do not own software or platforms and will provide our opinions based on our experience and expertise to help clients select the best solution for their situation.