Software development process
How we work
The team that builds the software applications for the KPMP takes an iterative software development approach. In iterative development, features are designed, developed and tested in repeated cycles. With each iteration (sprint), additional features can be designed, developed and tested until there is a fully functional software application ready to be deployed. This approach delivers value with each iteration, is adaptable to changes in requirements throughout development, and allows course correction as user or research community needs change.
Iterative software development Phases
Software Requirements & Planning
- Understand and clearly define the problem (who it’s for and how it will be used)
- Define what the software will do and what it cannot do
- Prioritize features and plan project work
Design / Develop / Test
- Utilize user-centered design best practices
- Create mockups & prototypes
- Conduct user testing validation & gather feedback
Deploy / Evaluate
- Implementation and deployment of features
- Quality assurance testing
- Conduct user testing validation & gather feedback
Release
- Release features in small chunks in order to bring value to user community faster
Personas & Use cases
Why use Personas?
Personas are representations of types of real users. They help us to:
- Understand our audience and answer the question, “Who are we building for?”
- Visualize/humanize our user research
- Prioritize features and align on strategy
- Use the same language when we’re discussing our users
- Create a shared understanding of how the product would be used by customers
- Provide a reference point throughout the design and development process
All websites and applications produced by the KPMP are designed and developed for a primary persona and their requirements. Other secondary, tertiary, etc. persona types may be included as long as they do not interfere with the needs or workflow of the primary persona.
KPMP Personas & Use Cases
The following user personas were developed for the KPMP by conducting user interviews within and outside of the KPMP Consortium.
Patient
If KPMP could come up with clear answers about my disease, it would be great.
I need to...
- Understand what the KPMP is and how it’s helping patients
- Understand the information being presented without having to know "science-speak"
- Find clear answers about my kidney disease
- Know how my biopsy contributes to the KPMP
- See my biopsy slides so I can get a feel for my kidney and share it with friends and family
- Understand "what's going on in my body and what I can do about it"
- See my information as it goes through the project
- Compare my data with other patients that are like/not like me
I want KPMP/the Kidney Tissue Atlas to provide...
- An online resource with patient-specific content about kidney disease and KPMP
- An online tool to view and export my biopsy images
- Online portal/dashboard of my KPMP status and results throughout the study
- A summary of my baseline characteristics along with biopsy images
- A way to search the KPMP data to find patients like/not like me
Pathologist
I expect KPMP to help me link biopsies to outcomes and mechanism.
I need to...
- View the slides that need to be assessed/scored from a central repository
- Quickly know the quality of a digital image
- Identify structural characteristics and assess/score them accordingly
- Determine if there is an association between pathology features
- Learn more about the implications of specific structural patterns
- Link pathology analyses to molecular and clinical information
I want KPMP/the Kidney Tissue Atlas to provide...
- [In the Atlas] A way to point at a cell/structure/feature and bring up the associated molecular and/or clinical data
- An efficient way to assess/score descriptors
- Machine learning applied to digital pathology
- Automated quality validation
Clinician
I’m hoping KPMP gives me the ability to link my individual patients to the best treatments.
I need to...
- Understand the causality of my patient's kidney disease
- Determine a prognosis based on the biopsy findings using all data sources developed by KPMP
- Validate my patient's prognosis
- Link my patients disease to available and emerging targeted therapies
I want KPMP/the Kidney Tissue Atlas to provide...
- Recruitment materials that are brief and easy to understand
- [In the Atlas] Clear diagnostic categories - more than just traditional mixed-bag pathology
- Interactive system where I can search for a particular data point and look at what corresponding KPMP data exists
- Help me understand and compare various treatment decisions and outcomes for KPMP participants
- A tool for clinical decision support, such as a kidney disease management algorithm that can be integrated into a hospital EHR system
Researcher:
Summary DAta Consumer
I'm not a bioinformatician. I just need the highlights for a quick validation of my gene of interest in AKI and CKD.
I need to...
- Dig deeper into the KPMP data to get more detailed information within my [-omics] specialty
- See highlights of the KPMP data data that is outside of my [-omics] specialty
- Compare my research data to the data within the KPMP datasets
I want KPMP/the Kidney Tissue Atlas] to provide...
- Be easy-to-use, understandable
- Have a user interface that provides summary information
- Provide easily accessible tutorials or help information
- Allow me to filter down into the data using my own custom search parameters
- Understand the underlying KPMP data sources and quality
Researcher:
Data Analyst
I work closely with my wet lab and want to use my analytical skills to answer their biological questions using KPMP.
I need to...
- Validate my own data against KPMP data
- Discover and publish novel insights gleaned from KPMP data leading to new research in my own labs
- Learn about new analysis tools and data
I want KPMP/the Kidney Tissue Atlas] to provide...
- Make it easy to visualize and download specific datasets
- Provide rich metadata
- Have data in a format that I can easily read into my workflow
- Provide a mechanism that allows me to see "what's new"
- Allow me to create my own dataset using my own custom search [boolean] parameters
Researcher:
Data Miner
I develop tools and analytic methods for kidney researchers to use in KPMP.
I need to...
- Download KPMP data and run my own analyses
- Access the KPMP data through an API or the cloud
- Build an app that could potentially be hosted in the atlas
I want KPMP/the Kidney Tissue Atlas] to provide...
- Make it easy for me to download lots of data to test my tool
- Provide rich metadata
- Have clear documentation and easy-to-use APIs
Student
As a future clinician/scientist, I look for resources to help me understand kidney function and failure at a molecular level.
I need to...
- Explore nephron segments in normal and disease states
- Determine at the molecular level, where glucose is reabsorbed and how this is changed in disease states such as diabetes
I want KPMP/the Kidney Tissue Atlas to provide...
- Provide a way to dynamically explore kidney function and failure concepts
- Make it easy to compare kidney function and failure in healthy and disease states
Use Cases
The following use cases were developed as a result of feedback provided by the members of the KPMP Consortium. These use cases then used to inform and design the features and functionality of our software tools so that they tools answer the questions that our user personas are asking.
Use case group:
Change practice of medicine
Use case
Patient
Pathologist
Clinician
Summary data consumer
Data analyst
Data miner
I want to identify clinical features that can predict outcome and be implemented as an alert in the EHR
I want to know if KPMP data will change physician behavior
Use case group:
Download the raw data so I can analyze it
Use case
Patient
Pathologist
Clinician
Summary data consumer
Data analyst
Data miner
I want individual clinical data
I want to download the raw omics data to apply my clustering algorithm
I want to download the omics data to mine it myself
I want to download the raw data so that I can apply machine learning algorithms to it.
I want access to all of KPMP's raw and processed data so I can mine it myself.
I want to download raw KPMP data and mine it myself
Use case group:
Find where cell types & molecular markers are located in tissue
Use case
Patient
Pathologist
Clinician
Summary data consumer
Data analyst
Data miner
Where is there a receptor in one cell type and its associated ligand in a neighboring cell type?
I want to know the differences in gene expression based on different tissue locations
I want to know where in the tissue and in which cells and how much a given gene is expressed
I want to see transcriptomic data in specific cell types by patient characteristics and disease stage
I want to see how omics data changes over time
I want to compare omics data to healthy tissue data
I want to know where each cell type is located in the kidney
I want to investigate whether a pathway of interest from an animal model is seen in the KPMP patient data.
I want to know where certain transcripts are expressed in kidney tissue
I want to examine the transcriptomic data from a cell type based on its surrounding cells / location in the tissue
I want to know where different cell types are located in the tissue
Use case group:
find KPMP patients similar to my search criteria
Use case
Patient
Pathologist
Clinician
Summary data consumer
Data analyst
Data miner
I want to see how many participants in KPMP are like me (a non-participant) (perhaps search for my diagnosis, demographics, etc. and see some summary stats)
I want to know what samples and data are available for certain patient subsets
I want to link currently available clinical data (routine labs) to pathology and omics data
I want to use a KPMP machine learning algorithm to match my biopsy to KPMP patient biopsies
I want to identify how many patients match specific criteria
I want to see the outcomes and treatment for a group of patients who share similar features specified by me
I want to identify a pathology feature and find it in all pathology images and link to clinical outcomes
I want to use a KPMP machine learning algorithm to find patients similar to mine based on specified data
I want to see the outcomes and treatment for a group of patients who share similar features specified by me
I want to see the outcomes and treatment for a group of patients who share similar features specified by me
I want to identify a pathology feature and find it in all pathology images
I want to identify certain KPMP patients by various characteristics.
I want to identify a subset of patients of interest and visualize their other data
Use case group:
Learn about KPMP
Use case
Patient
Pathologist
Clinician
Summary data consumer
Data analyst
Data miner
I want to learn what the study is about and how to enroll in it
I want to learn how the study is progressing (for instance, are there any discoveries being made? how many participants have been enrolled?)
I want to learn about specific omics technologies & methods so that they can be improved
Use case group:
Link pathology to molecular markers
Use case
Patient
Pathologist
Clinician
Summary data consumer
Data analyst
Data miner
I want to link molecular data to pathologic changes on the biopsy and also to link to clinical phenotypes
I want to link pathology data (user defined image selection) to molecular data
I want to identify and quantitate and annotate certain pathology features from a digital image
I want to correlate molecular data to pathology features.
I want to link molecular pathways to pathologic features to understand pathogenesis of those changes
I want to identify a biomarker of of a specific pathology features (fibrosis)
Use case group:
Match patients to existing therapies
Use case
Patient
Pathologist
Clinician
Summary data consumer
Data analyst
Data miner
I want to provide KPMP with my clinical and molecular characteristics and have it suggest clinical trials.
I want to identify pathology data that can predict response to specific therapies.
I want to look for specific molecular activation signatures for known existing medications and supplements to see if they could be repurposed to kidney disease
I want to know what omics features could be stained in kidney tissue that better inform therapy and prognosis
I want to provide KPMP with the clinical and molecular characteristics for one of my patients and have it calculate the optimal treatment plan based on my patient's similarities to others in the KPMP cohort.
Use case group:
Participant questions
Use case
Patient
Pathologist
Clinician
Summary data consumer
Data analyst
Data miner
I want to see where my tissue went
I want to see my whole slide images and other interesting visual output (e.g. 3D imaging)
I want to look up terms from my [medical records/pathology report and/or visit with a nephrologist] and get simple definitions/explanations for them (with visuals!)
I want to know what happened to my tissue
I want to learn how my contributions (of time and tissue) are (or would be) advancing treatments for kidney disease
I want to hover over a whole slide image and be told what I'm looking at in plain English (and what it does)
I want to compare my KPMP data to the data of other people in the study
I want to have a "decoder ring" for the terminology in my KPMP pathology report
I want to understand my disease, its prognosis, and what I can do to optimize it
Use case group:
Predict progression of kidney disease
Use case
Patient
Pathologist
Clinician
Summary data consumer
Data analyst
Data miner
I want to know what transcriptional changes are induced in specific cell types in AKI and whether those changes predict clinical outcome
I want to understand how much of a patient's disease and outcome are related to their genetics.
I want to compare all data from patients who progress rapidly with DKD vs. patients who remaing stable to identify predictors of response.
I want to link molecular pathways to different AKI trajectories
I want to identify molecular pathways in AKI patients
I want to identify biomarkers from the blood or urine which predict risk of CKD progression
I want to associate gene expression with clnical outcomes
I want to identify non-invasive biomarkers that can predict progression from AKI to CKD.
I want to identify molecular mechanisms which are associated with progression of DKD
I want to correlate molecular and pathology data to AKI recovery trajectories.
I want to associate omic data with AKI trajectories
I want to identify the molecular mechanisms in specific cell types that predict transition from AKI to CKD
Use case group:
Use molecular data to identify new therapies
Use case
Patient
Pathologist
Clinician
Summary data consumer
Data analyst
Data miner
I want to identify biomarkers which predict or track response to new therapies so that we can see how they are working.
I want to compare transcriptomics data from various disease states against a reference dataset
I want to identify disease mechanisms within unbiased-ly classified patient clusters so we can identify new drug targets.
I want to use KPMP omics data to identify novel therapies for specific patient subtypes
I want to identify new pathways that could be targeted by new drugs to prevent diabetes and hypertensive kidney disease
I want to identify molecular pathways which drive all types of CKD that can be targeted by new therapies
I want to use the tissue omics data to identify new therapies for DKD
I want to correlate molecular data with clinical data to identify new biomarkers for drug targets.
Use case group:
Use molecular data to sub-classify people
Use case
Patient
Pathologist
Clinician
Summary data consumer
Data analyst
Data miner
I want to subclassify with hypertensive nephropathy patients based on their molecular features.
I want to see unbiased classifications of the AKI cohort's molecular and pathology data (to discover if AKI is one disease or lots of diseases).
I want to know if clinical features can predict molecular disease etiolgy
I want to see unbiased classifications of the AKI cohort's clinical, molecular, and/or pathology data (to discover if AKI is one disease or lots of diseases).
Use case group:
What is the biopsy diagnosis
Use case
Patient
Pathologist
Clinician
Summary data consumer
Data analyst
Data miner
I want to know what different types of pathologic lesions are seen in patients with hypertensive CKD
I want to see all of the pathologic changes for a given clinical diagnosis (AKI, etc.)
I want to search KPMP participant's clinical features to see the various pathologic diagnoses seen on biopsy
Use case group:
Which cells express in certain molecular markers
Use case
Patient
Pathologist
Clinician
Summary data consumer
Data analyst
Data miner
I want to compare transcriptomics data with proteomics data for a cell type / structure
I want to know which cells express certain transcripts
I want to describe the omics by cell type and disease stage
Which genes are expressed in which cells, and where are those cells located?
I want to look at the transcriptome for a specific cell type (in a specific dataset) so I can sub-cluster it myself.
I want transcriptomic data from a specific cell type to correlate with pathology, clinical and genetic features
I want to perform a gene-gene relationship within specific cell types
I want to know what cell types express my gene of interest
I want to know which cells express certain transcripts
I want to know which transcripts are expressed in which cells and the level of expression
I want to know which transcripts are expressed in which cells
I want to know which transcripts are expressed in which cells by different disease types
