From 5c48157292024fca9013f9a62abe55a3bc789d89 Mon Sep 17 00:00:00 2001 From: Kaisar Dauyey Date: Thu, 8 Oct 2026 01:21:43 +0900 Subject: [PATCH 1/2] Redesign portfolio around published writing and thesis --- _data/navigation.yml | 8 +- _layouts/default.html | 3 +- _pages/about.md | 139 +++++++---------------------------- _pages/publications.md | 10 +++ _pages/research.md | 12 ++- assets/css/main.scss | 52 +++++++++++++ docs/writing-led-redesign.md | 21 ++++++ 7 files changed, 124 insertions(+), 121 deletions(-) create mode 100644 docs/writing-led-redesign.md diff --git a/_data/navigation.yml b/_data/navigation.yml index 3a9ab1a..b6edf49 100644 --- a/_data/navigation.yml +++ b/_data/navigation.yml @@ -8,15 +8,9 @@ # Markdown formatted page while the second is generated using JSON. main: - - title: "About" + - title: "Home" url: / - - title: "Clinical Practice" - url: /clinical-practice/ - - - title: "Diagnostic Medicine" - url: /diagnostic-medicine/ - - title: "Research" url: /research/ diff --git a/_layouts/default.html b/_layouts/default.html index 1c86667..de62a25 100644 --- a/_layouts/default.html +++ b/_layouts/default.html @@ -27,6 +27,7 @@ {% include browser-upgrade.html %} + {% include masthead.html %} {% if page.hero %}
@@ -40,8 +41,6 @@

{{ page.banner_title | default: page.title }}

{% endif %} - {% include masthead.html %} - {{ content }}
diff --git a/_pages/about.md b/_pages/about.md index 70ceeae..8d2205b 100644 --- a/_pages/about.md +++ b/_pages/about.md @@ -14,124 +14,41 @@ redirect_from: - /about.html --- -
-
-
-
- - Kaisar Dauyey - -
-

Physician · Translational researcher · Hokkaido University

-

Care informed by research.

-

I am a physician with general practice and residency experience in Kazakhstan and translational research training in Japan, Taiwan, Australia, and the United States.

-

I now study disease where it shows itself: in tissue, cells, molecules, and microbes. The best answers travel both ways.

- -
-
Clinical practice
General practitioner · Dau-Med Clinics, Kazakhstan
-
Research
Whole slide imaging and spatial biology, with AI and in silico gene perturbation in sequencing datasets · Hokkaido University
-
Training
MD · Nazarbayev University · PhD in Genetics · SOKENDAI · ECFMG Certified
-
- -
- Clinical practice - Diagnostic medicine - View CV -
+
+
+
+

Kaisar Dauyey · Physician and researcher

+

Questions from medicine.
Answers from biology.

+

My work starts with a question about a person, a population, or an organism. I use clinical observation, genetics, and computation to understand what the evidence can tell us.

+

My papers follow infections, cancer, immune disease, and the people who provide care. My doctoral thesis brings together ancient human genomes and the response of reef fish to a changing ocean.

+
Explore the research →Read the papers →
+

MD, Nazarbayev University · PhD in Genetics, SOKENDAI
Postdoctoral research, Hokkaido University

+
Kaisar Dauyey
Clinical experience in Kazakhstan. Research shaped by Japan and international collaborations.
-
-
-

What I do

-

Care for the person. Read the evidence.

-
-
-
-

Clinical practice

-

The first physician a person sees, and the one who stays

-

In general practice I cared for adults whose problems did not yet belong to any specialty. I learned to listen first, test sparingly, follow up, and coordinate care so patients did not have to. Continuity and context are the whole point.

-
    -
  • General practitioner, Dau-Med Clinics, Shymkent · 2025 to 2026
    • -
  • Case reports on late diagnosed colorectal cancer and aspirin intolerance in acute coronary syndrome
    • -
  • Study of burnout and wellbeing in Kazakhstan's palliative care workforce
    • -
- Clinical practice -
-
-

Diagnostic medicine

-

Where disease stops being a story and becomes something you can see

-

Every biopsy, culture, and genetic test I ordered was waiting on a laboratory answer. My research now lives there: whole slide imaging and spatial biology at Hokkaido University, molecular findings in hereditary cancer, and antimicrobial resistance in cultured H. pylori.

-
    -
  • Postdoctoral fellow, whole slide imaging and spatial biology · Hokkaido University
    • -
  • Co-author on BRCA2 and EHBP1 variant studies in prostate and colorectal cancer
    • -
  • First author, culture based H. pylori resistance study with regional meta-analysis
    • -
- Diagnostic medicine -
+
+

Start here

Follow the question.

+
+ Infection and resistanceWhat local evidence does a clinician need?Patient samples, culture, and antibiotic resistance in Kazakhstan. + Patients and careWhere does care become difficult?Diagnostic delays, cancer genetics, and the experience of palliative care professionals. + Ancient genomesHow far can a modern genetic score travel?Exploratory comparisons across ancestry and time. + Reef fish and adaptationHow does a changing ocean reach the genome?Transposable element related transcripts under elevated carbon dioxide.
-
-
-

How I work

-

One question, three scales

-
-
- -
Helicobacter pylori antimicrobial resistance summary from patients in Almaty, Kazakhstan
01 · PatientThe clinicStart with the person and the question their symptoms raise. Recruit, consent, and document with care.
- - -
Spatial molecular measurements aligned with whole slide tissue morphology
02 · TissueThe slideTest whether a predicted cell state makes sense in the tissue around it. Morphology is the referee.
- - -
Principal component plot of 1,402 intelligence-associated SNPs showing ancient individuals placed among modern populations
03 · GenomeThe scorePolygenic scores for intelligence applied to ancient genomes, with Naruya Saitou. On 1,402 top variants the ancient individuals sit inside modern population structure. Journal of Human Genetics, 2022.
- -
+
+

Selected writing

Read the evidence.

All publications →
+
Antibiotic resistance results from the Almaty H. pylori study

Frontiers in Microbiology · 2026

Antibiotic resistance in Almaty

Our study enrolled 150 patients and recovered 86 viable H. pylori isolates. Testing those isolates gave us a local resistance profile to compare with regional evidence.

Read the paper ↗
+
Themes from the study of Kazakhstan's palliative care workforce

BMC Palliative Care · 2025

The people behind the care

Interviews with 60 palliative care professionals in Kazakhstan explored emotional strain, burnout, and how people understood their work.

Read the paper ↗
+
Sea turtle among coral at Aka Island, Okinawa, providing field context for marine research

Doctoral thesis · SOKENDAI · 2024

Ancient humans and changing reefs

My thesis examines polygenic scores in ancient genomes and TE related expression in 72 reef fish brain transcriptomes. Both ask how genetic signals make sense in their biological context.

Read the thesis record ↗
-
-
-

From clinic to laboratory and back

-

Three patients, one loop

-

Each began as a question in a clinic, was answered in a laboratory, and came back as a decision about care.

-
-
-
- 01 -

Clinic → culture → treatment

-

A common infection, a local answer

-

Helicobacter pylori is everyday primary care. Recruiting patients, culturing isolates, and measuring antibiotic resistance in Almaty gave clinicians local evidence for first line therapy instead of borrowed guidelines.

-
primary careclinical microbiologyAMR
- Read the 2026 study -
-
- 02 -

Symptom → biopsy → diagnosis

-

A diagnosis that came too late

-

A young woman's colorectal cancer was recognised only after childbirth. The tissue diagnosis was unambiguous. The delay happened upstream, in how symptoms in young patients were weighed. I carry that lesson into both fields.

-
primary caretissue diagnosisoncology
- Read the case report -
-
- 03 -

Family history → variant → therapy

-

A variant that changed the plan

-

A pathogenic BRCA2 splice variant in a man with metastatic castration resistant prostate cancer. Molecular testing turned a family history into a treatment option, and a conversation with relatives.

-
molecular diagnosticshereditary cancergenetic counselling
- Read the case -
-
+
+

Current research

From tissue to cell states.

At Hokkaido University, I work with whole slide images, spatial biology, and computational models. Geneformer perturbations help propose genes to investigate. Donor consistency and tissue evidence guide the next step.

Explore the current work →
+ Geneformer cell embedding grouped by cell type and disease
- - -
-

Visitor locations worldwide

- Map showing the countries and regions from which this website has been visited -
-
+

Collaboration

What question are you working on?

Tell me what you want to understand and what evidence you have.

Get in touch →
+ +
diff --git a/_pages/publications.md b/_pages/publications.md index af16912..b8c7abe 100644 --- a/_pages/publications.md +++ b/_pages/publications.md @@ -12,6 +12,16 @@ banner_size: collage

Here is my publication record through 2026. You can check the latest identifiers and updates on ORCID or PubMed.

+
+

A guide to the writing

Choose a question, then read the study.

+
+ InfectionLocal evidence on antibiotic resistanceThe Almaty culture study and regional comparison. + Clinical medicineDiagnosis, treatment, and the work of careCase reports, reviews, and palliative care research. + Population geneticsAncient genomes and modern scoresThe 2022 paper and the first chapter of my thesis. + Doctoral thesisGene expression under environmental changeReef fish transcriptomes and TE related signals. +
+
+

Publication record

diff --git a/_pages/research.md b/_pages/research.md index 8bbed2e..cd9a433 100644 --- a/_pages/research.md +++ b/_pages/research.md @@ -11,7 +11,8 @@ banner_size: research-art --- -

A question from the clinic, an answer from tissue, molecules, or microbes, and a result that can return to the patient. Every result should show its assumptions and limits.

+

My research follows clinical questions into microbiology, genetics, and tissue biology. My thesis adds a wider view: how genomes relate to human history and environmental change.

+

Related public engagement: Oita Glocal feature on science, education, and regional collaboration.

@@ -44,6 +45,15 @@ same_sign = np.sign(by_donor["mean"]) == overall_sign
05

Model to mechanism

Machine Learning and Bioinformatics

I write workflows that another researcher can inspect and rerun. I document the inputs, the evaluation boundary, and the failures. Speed helps. Biological meaning matters more.

PythonRRNA sequencingGeneformerworkflow validation

Related work

+
+

Doctoral research · Environment and gene expression

+

Transposable elements in a changing ocean

+
+

My thesis examined 72 brain transcriptomes from Acanthochromis polyacanthus, a coral reef damselfish. The dataset compared control, acute, developmental, and transgenerational carbon dioxide exposure, with offspring from tolerant and sensitive parents.

I examined TE related transcripts alongside broader gene expression patterns. The work suggests questions about genome regulation and brain plasticity that need further investigation. Transcript abundance alone does not establish transposition or a regulatory mechanism.

Explore the public figures and evidence ↗
Read the thesis record ↗

+
Aka Island reef field context
Field context from Okinawa. The experimental dataset came from laboratory reared fish with parents collected near Palm Island, Australia.
+
+
+

People and places

Research is a team sport

Collaborators and mentors across Kazakhstan, Japan, Australia, and the United States.

diff --git a/assets/css/main.scss b/assets/css/main.scss index b051519..5b2cfdf 100644 --- a/assets/css/main.scss +++ b/assets/css/main.scss @@ -607,6 +607,58 @@ figcaption { font-family: var(--kd-sans); } .cv-photo { order: 2; flex: 0 0 auto; display: block; width: 192px; height: 192px; object-fit: cover; border-radius: 50%; border: 2px solid var(--kd-line); box-shadow: var(--kd-glow); } .about-profile { margin-bottom: 1rem; text-align: center; } .about-name { color: var(--kd-ocean); font-size: clamp(1.6rem, 3vw, 2.2rem); font-weight: 500; } + +/* Reading layout: quiet paper, clear routes, and evidence beside the text. */ +:root { --kd-paper: #faf8f4; --kd-surface: #fffdf9; --kd-line: #ddd8cf; --kd-grid: transparent; } +html[data-theme="dark"] { --kd-paper: #171c20; --kd-surface: #20272c; --kd-line: #39434a; --kd-grid: transparent; } +body { background-image: none; } +.masthead { position: sticky; top: 0; z-index: 100; } +.site-banner { min-height: 0; } +.site-banner--research-art, .site-banner--collage, .site-banner--reef { min-height: 0; height: auto; } +.site-banner__inner { padding-top: 2.8rem; padding-bottom: 2.8rem; } +.site-banner__title { font-size: clamp(2.1rem, 4.5vw, 3.6rem); max-width: 20ch; margin-left: auto; margin-right: auto; } +.site-banner__quote { max-width: 38rem; } +.reading-home, .writing-guide { max-width: 72rem; margin: 0 auto; } +.reading-intro { display: grid; grid-template-columns: 1.55fr 0.85fr; align-items: center; gap: clamp(2rem, 5vw, 5rem); padding: 1.5rem 0 3rem; } +.reading-intro h2 { font-family: Georgia, serif; font-size: clamp(2.2rem, 4.8vw, 4rem); line-height: 1.12; font-weight: 400; margin: 0.65rem 0 1.4rem; } +.reading-intro p:not(.eyebrow), .reading-study p:not(.eyebrow), .reading-current p { max-width: 62ch; font-size: 1rem !important; line-height: 1.75; color: var(--kd-muted); } +.reading-intro .reading-lead { color: var(--kd-ink); font-size: 1.2rem !important; } +.reading-portrait { margin: 0; } +.reading-portrait img { width: 100%; max-height: 28rem; object-fit: cover; border-radius: 4px; } +.reading-portrait figcaption, .reading-current figcaption { color: var(--kd-muted); font-size: 0.78rem; line-height: 1.6; margin-top: 0.7rem; } +.reading-actions { display: flex; flex-wrap: wrap; align-items: center; gap: 1.5rem; margin: 1.6rem 0; } +.reading-intro .reading-caption { font-size: 0.82rem !important; border-top: 1px solid var(--kd-line); padding-top: 1rem; } +.reading-section { border-top: 1px solid var(--kd-line); padding: 3rem 0; margin: 0; } +.reading-section h2, .writing-guide h2, .reading-next h2 { font-family: Georgia, serif; font-size: clamp(1.8rem, 3vw, 2.7rem); font-weight: 400; line-height: 1.2; margin: 0.4rem 0 1.5rem; } +.reading-paths { display: grid; grid-template-columns: 1fr 1fr; gap: 0 2.5rem; } +.reading-paths > a { position: relative; padding: 1.5rem 2rem 1.5rem 0; border-bottom: 1px solid var(--kd-line); text-decoration: none; color: var(--kd-ink); } +.reading-paths span { display: block; color: var(--kd-ocean); font-size: 0.76rem; margin-bottom: 0.6rem; } +.reading-paths strong { display: block; font-family: Georgia, serif; font-size: 1.35rem; font-weight: 400; line-height: 1.4; } +.reading-paths small { display: block; color: var(--kd-muted); font-size: 0.88rem; line-height: 1.65; margin-top: 0.7rem; } +.reading-paths b { position: absolute; right: 0; top: 1.5rem; color: var(--kd-ocean); } +.reading-paths > a:hover strong { color: var(--kd-ocean); } +.reading-heading, .reading-next { display: flex; align-items: center; justify-content: space-between; gap: 2rem; } +.reading-study { display: grid; grid-template-columns: 14rem 1fr; gap: 2.5rem; padding: 2rem 0; border-bottom: 1px solid var(--kd-line); align-items: center; } +.reading-study img { width: 100%; height: 12rem; object-fit: contain; background: #fff; border-radius: 4px; } +.reading-study h3 { font-family: Georgia, serif; font-size: 1.65rem; font-weight: 400; margin: 0.3rem 0 0.8rem; } +.reading-current { display: grid; grid-template-columns: 1fr 1fr; gap: 2.5rem; align-items: center; } +.reading-current > img, .reading-current figure img { display: block; width: 100%; border-radius: 4px; } +.reading-current figure { margin: 0; } +.reading-next { background: var(--kd-surface); border: 1px solid var(--kd-line); padding: 2rem; margin: 1.5rem 0; } +.reading-next h2 { margin-bottom: 0.8rem; } +.reading-related, .section-index { display: flex; flex-wrap: wrap; gap: 0.8rem 1.6rem; padding: 1.2rem 0; } +.reading-related a, .section-index a { font-size: 0.88rem; } +.section-index { border-top: 1px solid var(--kd-line); border-bottom: 1px solid var(--kd-line); margin-bottom: 2rem; } +[id] { scroll-margin-top: 6rem; } +.writing-guide { margin-bottom: 3rem; } +@media screen and (max-width: 46em) { + .reading-intro, .reading-current, .reading-paths, .reading-study { grid-template-columns: 1fr; gap: 1.5rem; } + .reading-portrait { max-width: 22rem; } + .reading-study img { height: 14rem; } + .reading-heading, .reading-next { align-items: flex-start; flex-direction: column; gap: 0.7rem; } + .reading-section { padding: 2rem 0; } + .site-banner__inner { padding-top: 1.8rem; padding-bottom: 1.8rem; } +} .about-visual .section-heading { text-align: center; } @media screen and (max-width: 36em) { .cv-photo { width: 128px; height: 128px; } diff --git a/docs/writing-led-redesign.md b/docs/writing-led-redesign.md new file mode 100644 index 0000000..939208a --- /dev/null +++ b/docs/writing-led-redesign.md @@ -0,0 +1,21 @@ +# Writing and structure + +The October 2026 redesign follows the supplied paper collection and the final 2024 doctoral thesis. Existing Huxley quotations and source links remain unchanged. + +## Visitor route + +Home introduces the author and gives visitors four research questions. Research provides section links and fuller explanations. Publications retains the chronological record and adds a guide by subject. CV and Contact complete the main navigation. Clinical Practice and Diagnostic Medicine remain available through Home and the research pages. + +## Source decisions + +| Source | Use in the site | +| --- | --- | +| Frontiers in Microbiology, 2026, DOI 10.3389/fmicb.2026.1747006 | The homepage explains the 150 patient study and 86 viable isolates. | +| BMC Palliative Care, 2025, DOI 10.1186/s12904-025-01798-z | The homepage introduces interviews with 60 professionals and the experience of care. | +| Journal of Human Genetics, 2022, DOI 10.1038/s10038-022-01039-8 | Ancient genomes form a distinct research route, with the existing ancestry and transferability limits retained. | +| Deciphering Intelligence in Ancient Humans and Adaptation in Coral Reef Fish, SOKENDAI, 2024 | The homepage connects both thesis topics. Research describes the 72 brain transcriptomes and four exposure conditions from Chapter 2, pages 42 and 43. | +| Clinical case reports and immune disease papers supplied with the collection | The clinical research route leads to the existing case reports and publication record. | + +The Aka Island photographs provide field context. The thesis experimental fish dataset came from laboratory reared offspring whose parents were collected near Palm Island in Australia. These places are stated separately in the TE section. + +No PDF from the supplied folders is copied into the public website. Readers follow existing journal identifiers and the public thesis record. From cc8e2169cb7c530feed8f6e454d1da487cbac096 Mon Sep 17 00:00:00 2001 From: Kaisar Dauyey Date: Thu, 8 Oct 2026 01:24:11 +0900 Subject: [PATCH 2/2] Point Geneformer links to the current public SCLC repository --- _pages/diagnostic-medicine.md | 8 ++++---- _pages/projects.md | 4 ++-- _pages/research.md | 6 +++--- 3 files changed, 9 insertions(+), 9 deletions(-) diff --git a/_pages/diagnostic-medicine.md b/_pages/diagnostic-medicine.md index b889f9b..d6bf9a9 100644 --- a/_pages/diagnostic-medicine.md +++ b/_pages/diagnostic-medicine.md @@ -34,7 +34,7 @@ banner_size: diagnostic-art

Whole slide imaging and spatial biology

I analyse whole slide images alongside spatial molecular measurements. The question is always whether a predicted cell state makes sense in the tissue around it. A model can propose. The slide decides.

whole slide imagingspatial transcriptomicstumour microenvironmentT cell states
- View the tissue and model workflow + View the tissue and model workflow
02 @@ -60,7 +60,7 @@ banner_size: diagnostic-art

My PhD in genetics and my current work with Geneformer give me tools that most diagnostic teams do not have yet. I use them to generate hypotheses about cell states, then hold those hypotheses to the standard of tissue evidence. A prediction is not a diagnosis.

Immunopathology

Expression profiling in vasculitis

Longitudinal profiling of CD4+ and CD8+ cells in giant cell arteritis, from active disease to remission. My first exposure to reading immune cell states as a marker of tissue disease.

Read the study

-

In silico perturbation

Ask the model, then check the tissue

Geneformer predicts how deleting or overexpressing one gene shifts a T cell representation in small cell lung cancer. I check that each donor points the same way, then look for the same signal in the spatial data.

View the workflow

+

In silico perturbation

Ask the model, then check the tissue

Geneformer predicts how deleting or overexpressing one gene shifts a T cell representation in small cell lung cancer. I check that each donor points the same way, then look for the same signal in the spatial data.

View the workflow

@@ -78,8 +78,8 @@ banner_size: diagnostic-art

Figures

From slide to model to microbe

-
Diagram of how ranked Geneformer gene inputs are deleted or overexpressed to estimate predicted cell state shifts
In silico perturbationRanked gene inputs are deleted or overexpressed, then the cell is read again to estimate a predicted shift. Workflow.
-
Geneformer embedding of lung cells organized by cell type and disease
Cell state embeddingLung cells organised by type and disease state in the Geneformer representation.
+
Diagram of how ranked Geneformer gene inputs are deleted or overexpressed to estimate predicted cell state shifts
In silico perturbationRanked gene inputs are deleted or overexpressed, then the cell is read again to estimate a predicted shift. Workflow.
+
Geneformer embedding of lung cells organized by cell type and disease
Cell state embeddingLung cells organised by type and disease state in the Geneformer representation.
Helicobacter pylori antimicrobial resistance in Almaty, Kazakhstan, summarising resistance rates and genomic markers
From patient to plate to policy. H. pylori resistance profile from Almaty. Analysis repository · Paper.
diff --git a/_pages/projects.md b/_pages/projects.md index d75c94f..402328d 100644 --- a/_pages/projects.md +++ b/_pages/projects.md @@ -10,7 +10,7 @@ published: false
Whole-slide tissue sections mapped with spatial T-cell identity and dysfunction scores
Tissue to model
-

Model to counterfactual

Geneformer In Silico Perturbation

Model-predicted deletion and overexpression effects in SCLC T cells, evaluated with donor separation and explicit experimental-validation boundaries.

Geneformergene perturbationT-cell statesvalidation
View workflow +

Model to counterfactual

Geneformer In Silico Perturbation

Model-predicted deletion and overexpression effects in SCLC T cells, evaluated with donor separation and explicit experimental-validation boundaries.

Geneformergene perturbationT-cell statesvalidation
View workflow
Historical map used to introduce population and ancient-genome research
Genome to risk
@@ -25,4 +25,4 @@ published: false
- + diff --git a/_pages/research.md b/_pages/research.md index cd9a433..3541206 100644 --- a/_pages/research.md +++ b/_pages/research.md @@ -19,7 +19,7 @@ banner_size: research-art

Projects at a glance

Four questions, four kinds of evidence

-
Diagram of Geneformer in silico gene perturbation
Tissue to modelGeneformer perturbation in SCLC T cells. Repository.
+
Diagram of Geneformer in silico gene perturbation
Tissue to modelGeneformer perturbation in SCLC T cells. Repository.
H. pylori antimicrobial resistance profile from Almaty
Clinic to microbeH. pylori resistance in Almaty. Repository.
Kisaburo Ohara's 1904 satirical map of Europe and Asia
Genome to riskPolygenic scores across ancient and present day populations. Map: Ohara, 1904, public domain. Repository.
Sea turtle among coral at Aka Island, Okinawa
Environment to genomeTransposable elements under ocean acidification, Aka Island fieldwork. Repository.
@@ -31,7 +31,7 @@ banner_size: research-art by_donor = cells.groupby("individual")["shift"].agg(["mean", "count"]) overall_mean = float(np.mean(shifts)) overall_sign = np.sign(overall_mean) -same_sign = np.sign(by_donor["mean"]) == overall_signRead the source in GitHub

View the current Geneformer T cell work

+same_sign = np.sign(by_donor["mean"]) == overall_signRead the source in GitHub

View the current Geneformer T cell work

02

Clinic meets microbiology

Infectious Diseases and Host Pathogen Dynamics

My Helicobacter pylori work starts with patients in Kazakhstan, a common primary care problem. I connect clinical records with culture, epidemiology, and antimicrobial resistance. That local evidence can sharpen regional surveillance and first line treatment decisions.

H. pyloriantimicrobial resistanceclinical microbiologyclinical epidemiologyprimary care

Open the analysis repository · Read the 2026 study

03

Clinical practice · Care to evidence

Clinical Practice and Evidence

General practice taught me to work with incomplete information. It also taught me that an elegant analysis is useless if it does not fit real care. I carry that lesson into research on primary care, infectious disease, palliative care, cancer genetics, and immune disease. Read more about my clinical practice.

general practiceprimary carepalliative careclinical researchstudy designmultidisciplinary care

Related clinical work

04

Genome to risk

Polygenic Risk Score Pipelines

I compare polygenic scores across ancient and present day populations. The hard part is not producing a score. It is handling variants clearly and judging whether an effect estimate can travel across ancestry and time.

polygenic scorespopulation genomicsancient DNARreproducibility
See how the score is calculated
R excerptpgs_calculator_1402snps.R

The script matches each observed allele to its reported effect allele, collects the matching effects, and sums them for each person.

for (v in 1:nrow(df2)) {
@@ -42,7 +42,7 @@ same_sign = np.sign(by_donor["mean"]) == overall_sign
Read the source in GitHub

Open the Ancient Intelligence repository

-
05

Model to mechanism

Machine Learning and Bioinformatics

I write workflows that another researcher can inspect and rerun. I document the inputs, the evaluation boundary, and the failures. Speed helps. Biological meaning matters more.

PythonRRNA sequencingGeneformerworkflow validation

Related work

+
05

Model to mechanism

Machine Learning and Bioinformatics

I write workflows that another researcher can inspect and rerun. I document the inputs, the evaluation boundary, and the failures. Speed helps. Biological meaning matters more.

PythonRRNA sequencingGeneformerworkflow validation

Related work