Sharpen specific weaknesses without burning a full mock-interview session. 106+ curated questions across 17 topics.
106 of 106 questions
Explain the difference between FDR and FWER. When would you pick one over the other?
Walk me through how you'd power a single-cell DE analysis between two conditions.
What's the difference between a p-value and an effect size? Why does the field over-rely on p-values?
Define a confounder. How do you detect one in an observational omics dataset?
When would you use a negative binomial vs Poisson for RNA-seq counts?
Explain Simpson's paradox with a biological example.
What is a permutation test and when does it beat a parametric test?
How would you correct for batch effects in a multi-site cohort?
What's the assumption behind a Cox proportional hazards model, and how do you test it?
Explain bootstrap vs jackknife. Which gives tighter CIs for small samples?
Walk me through your standard scRNA-seq QC pipeline.
Seurat vs Scanpy — which do you use day-to-day and why?
How do you choose number of PCs for UMAP?
Explain the math behind Harmony integration in 2 minutes.
When is RNA velocity misleading?
Compare scVI, scANVI, and totalVI. When would you reach for each?
How do you annotate cell types without a reference atlas?
Walk me through CellRanger output — what files matter and why?
How do you analyze CITE-seq or multi-modal data?
Pseudobulk DE vs single-cell DE — pros and cons.
How do you detect doublets and what tool do you trust?
Explain trajectory inference. When does Monocle3 fail?
Walk me through a germline variant calling pipeline.
GATK Best Practices: what changed in the last 3 years?
How do you call somatic mutations in tumor-normal? Tumor-only?
Explain VQSR. When do you skip it?
How would you find structural variants in long-read data?
What's the difference between population stratification and cryptic relatedness in GWAS?
Walk me through computing a polygenic risk score for a non-European cohort.
What does Hardy-Weinberg equilibrium tell you about a variant?
Explain imputation. What reference panel would you use for African genomes?
How do you handle multi-allelic sites in a VCF?
BWA-MEM vs Minimap2 vs STAR — when do you reach for each?
Salmon vs Kallisto vs STAR+featureCounts. Pros and cons of pseudo-alignment.
DESeq2 vs edgeR vs limma-voom. When would you pick limma-voom?
How do you handle low-count genes before DE?
Explain how DESeq2 estimates dispersion.
GSEA vs over-representation analysis. When is ORA misleading?
How would you detect novel splice junctions?
Bulk RNA-seq deconvolution — name 3 methods and their key assumption.
Walk me through AlphaFold2's architecture at a high level.
When would you NOT trust an AlphaFold prediction?
ESM-2 vs ProtBERT — which would you use for variant effect prediction?
Explain the difference between docking and molecular dynamics.
How would you design a protein binder de novo?
RFdiffusion vs ProteinMPNN — what does each solve?
Walk me through a CRISPR screen QC pipeline.
Pooled vs arrayed CRISPR screens — when to use each?
How do you design guide RNAs and which scoring metric do you trust?
Explain base editing vs prime editing. Off-target profile differences?
How would you analyze a Perturb-seq experiment?
What's the difference between CLIA, CAP, and ISO 15189?
Walk me through validating a clinical NGS assay end-to-end.
ACMG variant classification — what evidence codes matter most?
How is a 21 CFR Part 11 compliant pipeline different from a research pipeline?
What is GxP and which of GMP / GLP / GCP applies to bioinformatics?
Why does cross-validation leak in genomics data?
Explain SHAP vs LIME for interpreting a clinical risk model.
Class imbalance in a rare-disease classifier — what's your approach?
How would you fine-tune a protein language model on a 5K-sequence assay?
Compare diffusion models vs VAEs for molecule generation.
What's the difference between transduction and induction in ML?
How would you validate a deep-learning histopathology model for FDA submission?
Snakemake vs Nextflow vs WDL. Which would you pick for a 1000-sample run on GCP?
Design a scalable pipeline to process 100K WGS samples a year.
How do you handle failed jobs in a distributed pipeline?
Containers (Docker/Singularity) vs Conda envs — trade-offs for reproducibility?
Design a data lake for multi-omics + clinical metadata.
Write a function that returns the reverse complement of a DNA string.
Parse a VCF and return variants with AF > 0.05. Discuss complexity.
Given a BED file and a SAM file, compute coverage per interval. Stream it — assume both are huge.
Implement pairwise sequence identity for k sequences. O(k²L) is the obvious answer — make it faster.
Find the longest common subsequence between two protein sequences.
Given a GTF file, return all transcripts longer than 5kb that overlap a given region.
Implement k-mer counting with O(N) memory for an arbitrary genome.
SQL: given a `samples` table and a `variants` table, find variants present in >5 samples but absent in healthy controls.
Tell me about a time you disagreed with your PI or manager. How did you resolve it?
Walk me through a project that failed. What did you do differently next time?
Describe the most ambiguous problem you've worked on.
Tell me about a time you had to explain a complex result to a non-technical stakeholder.
Describe a time you had to push back on a wet-lab partner's experimental design.
Tell me about a time you missed a deadline.
How do you prioritize when 3 PIs want their analysis done first?
Tell me about your biggest scientific contribution to date.
Describe a time you mentored someone — what worked, what didn't?
How do you handle being the only computational person in a wet-lab team?
What's the most interesting paper you've read in the last 6 months and why?
Critique a paper from our group — pick any one.
Walk me through the original AlphaFold2 Nature paper's key innovation vs prior work.
Walk me through a TCR repertoire analysis pipeline.
How do you estimate tumor mutational burden (TMB)?
Explain HLA typing from short reads.
How would you predict neoantigens for a personalized cancer vaccine?
What is microsatellite instability and how do you call it from sequencing?
What computational metrics matter when designing an mRNA sequence?
Explain codon optimization. What are the trade-offs?
How would you model LNP formulation efficiency from sequence + lipid features?
What are your salary expectations?
Why are you leaving your current role?
Walk me through your decision matrix for choosing between two offers.
What's your notice period and earliest start date?
How important is equity vs base salary to you?
Why our company over a Big Pharma competitor?
How do you make a paper's analysis reproducible end-to-end?
What's wrong with `pip install` in a published Nature Methods paper?
Explain how you'd version both code AND data for a 5-year regulatory submission.