MD-PhD Letter of Recommendation From a PI

MD-PhD Letter of Recommendation From a PI

An MD-PhD letter from a PI is read as a forecast of independent research. Programs ask for it by name: Stanford's MSTP "wants to see letters of recommendation from your research mentor, or mentors, addressing your experiences in the laboratory", and requires three to six letters in all.

So the paragraph that carries your letter is the one where the student changed the plan: spotted the problem, proposed the control, asked the question that became a figure. A list of techniques does not do that work. A medical school letter about compassion and reliability does not either.

Every program page quoted below was opened on November 7, 2026. The AAMC's AMCAS letter instructions and its guidelines for letter writers returned HTTP 403 to our fetch that day, so this page quotes neither.

What MD-PhD programs say about research letters

Stanford states the purpose of the review plainly: its MSTP "is interested in identifying students with significant undergraduate research experience that would predict successful completion of a PhD program." That is the question your letter answers.

Stanford also says "All letters of recommendation must be submitted through the AMCAS Letters Service", and Harvard says "All letters of recommendation should be sent directly to AMCAS." McGill uses its own portal and gives referees 10 business days after the applicant submits.

Every research mentor writes

The Stanford wording, "your research mentor, or mentors", is the rule to plan around. If the applicant spent two years in your lab and a summer in another, both of you should write.

AMCAS sets a ceiling too. Penn's MD admissions page says "applicants may only submit up to 10 letters through the AMCAS Letter Writer Application." Inside that, the applicant still needs the letters their MD schools ask for, from science faculty and others.

The four paragraphs of a research-potential letter

This structure is our recommendation, built from what the programs above say they read for.

1. Role, duration and how you know. Months in the lab, hours per week, the project, and who supervised day to day. If a postdoc or graduate student did, say so and say which observations are theirs.
2. The intellectual contribution. One moment when the student changed what the lab did: a control they proposed, a question they asked that became an experiment, a flaw they found. Say how much of the plan was theirs and how much was yours.
3. The setback. An experiment that failed, and what the student did over the following weeks. This paragraph is your evidence on how the student handles work that does not succeed.
4. The comparison and the forecast. A count of the trainees you have supervised, where this student falls among them, and what you expect them to be able to do. End with an endorsement at the strength you mean.

Leave out the method list ("PCR, Western blots, confocal imaging"). The applicant's research description has it. One technique matters only if mastering it unusually fast is the evidence.

Our view: paragraph 2 decides the letter. A PI who can write it should spend most of the page there.

The countercase: some students did careful, valuable work inside a fixed protocol and had no room to change the plan. Do not invent independence. Describe the quality of execution, how fast they learned, and the questions they asked, and say plainly that the role was technical. A precise letter about a technical role reads better than an inflated one.

The same letter is read twice

At Stanford, "If an applicant is not chosen for an MSTP interview, their file is automatically routed to the MD Admissions Committee for consideration." Your letter then goes to readers who are judging a future physician, not a future scientist.

Give them one short paragraph they can use: how the student treats people in the lab, how they teach, whether they take feedback. Two or three sentences, one incident. The general guide to writing a medical school letter of recommendation covers what MD readers look for, and AAMC premed competencies in a rec letter shows which competencies a PI is placed to see.

Constructed sample PI letter, annotated

Constructed example, written by us, not a real recommendation letter. Context (invented): Mateo Ibarra worked two years as a research technician before applying MD-PhD. The letter is 310 words.

Dear MD-PhD Admissions Committee,

I am an associate professor of pharmacology and run a lab of nine people studying how kidney cells respond to low oxygen. Mateo Ibarra worked in my lab as a full-time research technician from July 2024 to June 2026. I met with him weekly for two years. My postdoctoral fellow, Dr. Ingrid Sato, supervised his bench work daily, and the observations below are mine unless I say they are hers.

Mateo was hired to run our imaging pipeline. Six months in, he noticed that our cell counts dropped on days when plates waited longer before fixation. He proposed and ran a timing control that the whole lab now uses. In his second year he asked whether the response we study differs in cells from older donors. The question was his, he wrote the plan, and I changed one step. The result is a figure in a manuscript in preparation, with Mateo as second author.

That project failed for three months: the donor cells would not grow. Dr. Sato reports that Mateo logged every condition he tried, asked a neighboring lab for their protocol, and presented the failure at lab meeting with a ranked list of next steps. The fourth option worked.

He teaches well. When we took on two summer undergraduates, he wrote the training document they used and checked their first ten imaging runs himself.

In 14 years I have supervised 22 technicians and 11 graduate students. Mateo is one of three technicians I would have admitted to our graduate program on their lab work alone. His gap is statistics beyond the methods he has used. He knows this and is taking a course this fall.

I recommend him with enthusiasm for MD-PhD training, and I expect him to run his own research program.

Sincerely,

[Name], PhD, Associate Professor of Pharmacology

[University letterhead, phone, email]

What each paragraph does:

• Paragraph 1 names who saw what. Daily supervision by a postdoc is stated, not hidden.
• Paragraph 2 is the forecast paragraph: one problem he found, one question he owned, and the PI's exact share ("I changed one step").
• Paragraph 3 is the setback, attributed to the postdoc who watched it.
• Paragraph 4 gives the MD readers their paragraph: teaching, with an observable act.
• Paragraph 5 compares with a count, 3 of 22 technicians, and names a real gap. A named gap makes the praise around it believable.
• The close states the strength of the endorsement and the forecast.

When the postdoc knows the student better

If a postdoc or graduate student supervised the work, you have two clean options:

• One letter, two signers. Stanford asks for letters "signed by the recommender(s) or committee members", so a jointly signed letter fits its wording. The letter must say which observations belong to whom. Interfolio's letter writer article says that for AMCAS letters "A physical (wet or handwritten) signature is mandatory." If the letter goes through Interfolio, both signatures should be handwritten.
• Two letters. If the program allows enough letters and both of you saw different things, write separately.

What does not work is a PI signature on a postdoc's letter with no attribution. Joint PI and postdoc letters covers the attribution in detail, and who can write a medical school letter covers co-signing by residents and others.

Checklist before you upload

1. The first paragraph says how long, how often, on what project, and who supervised day to day.
2. One paragraph shows the student changing the plan, with your share of it stated.
3. One paragraph covers a setback and what the student did about it.
4. Two or three sentences give MD readers an incident about how the student treats people.
5. The comparison has a count.
6. Methods appear only where they are the evidence.
7. Letterhead and handwritten signature, submitted through the service the program names.

Questions about MD-PhD letters

How many letters does an MD-PhD applicant need?

It depends on the program. Stanford requires three to six. Harvard and Rutgers ask for research mentor letters without a number on the pages we opened. Through AMCAS, Penn says applicants can submit up to 10 letters.

Is an MD-PhD letter different from a PhD admissions letter?

Yes, in two ways. At Stanford an MD-PhD file not chosen for an MSTP interview goes on to the MD committee, so the letter needs a paragraph for each reader. And AMCAS letters sit beside the medical school file, not a graduate application. The PhD recommendation letter sample shows the graduate-school version.

Should I mention the student's publications?

Mention the student's role in a paper or poster, not the citation. The applicant lists the publications. Your job is to say what part of the work was theirs.

Can the student add a research letter after an interview?

At Stanford, yes: updates are accepted only after an interview invitation, and its page lists "additional letters of recommendation" among acceptable updates. Other schools differ; see extra letters of recommendation after an interview.

Get a free review of the letter

Get a free Quick Review of your recommendation letter against GradPilot's medical school letter of evaluation rubric. It is the MD letter rubric, so it reads your letter the way the MD committee will if the file is routed there: whether each claim stays inside what you observed, and whether your comparison names its group. Other letter types are on the recommendation letter review page.

The applicant's side of the same file is in the MD-PhD essay guide, and more letter guides are in the medical school essays hub.

Sources

Retrieved November 7, 2026, except where noted.

• Stanford Medicine, MSTP Admissions and Eligibility: research mentor letters, three to six letters, letterhead and signature "by the recommender(s)", AMCAS Letters Service, routing to the MD committee, post-interview letters.
• Harvard Medical School, MD-PhD Admissions Process: research mentor letters expected; letters to AMCAS.
• Rutgers New Jersey Medical School, MD-PhD Admission Requirements: at least one research mentor letter.
• Washington University in St. Louis, MSTP Admissions: two years of research and outstanding letters of reference.
• McGill University, MDCM-PhD: Apply now: two research mentor references; 10 business days.
• Perelman School of Medicine, Application Procedure: up to 10 letters through the AMCAS Letter Writer Application.
• Interfolio, Submit a Letter of Recommendation (For Letter Writers): handwritten signature for AMCAS letters.
• Not quoted: the AAMC's AMCAS letters of evaluation instructions and Guidelines for Writing a Letter of Evaluation (students-residents.aamc.org) returned HTTP 403 on November 7, 2026.