Skip to content

Biostatistics, IPR, Biosafety and Research Ethics

Accuracy, distributions, confidence intervals, t-test, ANOVA, chi-square, regression, multivariate basics, patents, biosafety, ethics, plagiarism and AI.

  • 23 min read
  • Updated
On this page · 13 sections
  1. Prerequisites
  2. The Hook: Three Mice and a p-Value
  3. Part 1: Biostatistics
  4. Part 2: Intellectual Property Rights
  5. Part 3: Biosafety
  6. Part 4: Research and Publication Ethics
  7. Memory Tricks
  8. Don’t Confuse
  9. Common Mistakes
  10. Practice
  11. Quick Revision Table
  12. Key Takeaways
  13. What’s Next

Prerequisites

The Hook: Three Mice and a p-Value

How CSIR NET asks it. Part B: accuracy vs precision, SD vs SE, properties of normal/binomial/Poisson distributions, which test for which design, degrees of freedom, Type I/II errors, patent criteria, Section 3(d), BSL examples, biosafety cabinet classes, the 3Rs, FFP. Part C: confidence intervals, t, χ² and F calculations against given critical values, regression standard curves, multiple-testing corrections, and multi-statement questions on IPR, biosafety and ethics.


Part 1: Biostatistics

Measurement: accuracy, precision and error

TermMeaning
AccuracyCloseness of measurements to the true value; reduced by systematic error (bias) such as an uncalibrated pipette
PrecisionCloseness of repeated measurements to each other; reduced by random error
Technical replicateSame sample measured again: tests the method
Biological replicateIndependent samples (different animals, cultures): tests the biology; the n for statistics

Tight but wrong values are precise and inaccurate; scattered values centred on the truth are accurate and imprecise.

Central tendency and dispersion

  • Mean \(\bar{x} = \sum x_i / n\); median (middle value, robust to outliers); mode (most frequent).
  • In a right-skewed (positively skewed) distribution: mode < median < mean.
  • Range; variance \(s^2 = \dfrac{\sum (x_i - \bar{x})^2}{n-1}\) (n − 1 for a sample); standard deviation \(s = \sqrt{s^2}\).
  • Standard error of the mean \(SE = s/\sqrt{n}\): the precision of the mean, which shrinks as n grows; SD describes the spread of individual values and does not shrink with n.
  • Coefficient of variation \(CV = (s/\bar{x})\times100\%\): unit-free comparison of variability.

Distributions

DistributionDescribesMeanVarianceBiological example
Normal (Gaussian)Continuous, symmetricμσ²Height, enzyme activity across replicates
BinomialNumber of successes in n independent trials, probability pnpnp(1 − p)Number of affected children among n offspring
PoissonCount of rare, independent events in a fixed intervalλλ (mean = variance)Mutations per genome, colonies per plate, phage per cell
tMean of a small sample with s estimated0Slightly > 1, tends to 1 as df risesSmall-sample tests and intervals
χ²Sum of squared standard normalsdf2 dfGoodness-of-fit, contingency tests
  • Normal rule: about 68% of values lie within μ ± 1σ, 95% within ±1.96σ (about ±2σ) and 99.7% within ±3σ. z-score \(z = (x - \mu)/\sigma\).
  • Poisson probability: \(P(k) = \dfrac{e^{-\lambda}\lambda^k}{k!}\); \(P(0) = e^{-\lambda}\). This is how the fraction of uninfected cells is predicted from the multiplicity of infection, and how Luria and Delbrück’s fluctuation test showed that mutant counts did not follow a Poisson distribution.
  • Central limit theorem: means of samples tend to a normal distribution as n grows, whatever the parent distribution.

Confidence intervals

A 95% confidence interval for a mean:

$$\bar{x} \pm t_{0.05,\,n-1}\,\frac{s}{\sqrt{n}}$$

(use z = 1.96 when σ is known or n is large). Interpretation: if the experiment were repeated many times, 95% of intervals built this way would contain the true mean. It is not “a 95% probability that this interval contains μ” in the strict frequentist sense. Larger n or smaller s gives a narrower interval.

Hypothesis testing

  • Null hypothesis H₀: no difference or no effect. The p-value is the probability of results at least as extreme as those observed if H₀ were true.
  • α (significance level): commonly 0.05; reject H₀ if p < α.
H₀ trueH₀ false
Reject H₀Type I error (false positive), probability αCorrect (power = 1 − β)
Do not reject H₀CorrectType II error (false negative), probability β
  • Power rises with sample size, effect size and α, and falls with variability. Power analysis sets n before the experiment.
  • One-tailed tests are used only when the direction is specified in advance.
  • A non-significant result is not proof of no effect; statistical significance is not biological importance (report effect sizes).

Choosing a test

graph TD
    Q["What are you comparing?"] --> M["Means of a measured variable"]
    Q --> C["Counts in categories"]
    Q --> R["Relationship between two variables"]
    M --> G2["Two groups"]
    M --> G3["Three or more groups"]
    G2 --> IND["Independent: unpaired t-test
(Mann-Whitney if non-normal)"] G2 --> PAIR["Same subjects: paired t-test
(Wilcoxon signed-rank)"] G3 --> AN["One-way ANOVA + post hoc
(Kruskal-Wallis)"] C --> GOF["Observed vs expected ratio:
chi-square goodness of fit"] C --> IN["Two classifications:
chi-square test of independence"] R --> COR["Correlation (r)"] R --> REG["Regression (predict y from x)"]

t-tests

TestStatisticdf
One-sample\(t = \dfrac{\bar{x} - \mu_0}{s/\sqrt{n}}\)n − 1
Unpaired (pooled variance)\(t = \dfrac{\bar{x}_1 - \bar{x}_2}{\sqrt{s_p^2\,(1/n_1 + 1/n_2)}}\), \(s_p^2 = \dfrac{(n_1-1)s_1^2 + (n_2-1)s_2^2}{n_1+n_2-2}\)n₁ + n₂ − 2
Paired\(t = \dfrac{\bar{d}}{s_d/\sqrt{n}}\) on the differences dn − 1

Assumptions: approximately normal data, independent observations, and (for the pooled test) similar variances; Welch’s t-test drops the equal-variance assumption. Pairing removes between-subject variation and is more powerful when the design allows it.

ANOVA

One-way analysis of variance compares k group means by partitioning total variation:

$$F = \frac{MS_{between}}{MS_{within}} = \frac{SS_{between}/(k-1)}{SS_{within}/(N-k)}$$
  • Large F means group means differ more than expected from within-group scatter. ANOVA says at least one mean differs; post hoc tests (Tukey, Bonferroni-corrected comparisons) say which.
  • Running many pairwise t-tests instead inflates the Type I error rate.
  • Two-way ANOVA tests two factors and their interaction (does the drug effect depend on genotype?).

Chi-square tests

$$\chi^2 = \sum \frac{(O - E)^2}{E}$$
  • Goodness of fit: observed counts vs an expected ratio (Mendelian 3 : 1, 9 : 3 : 3 : 1; Hardy–Weinberg). df = number of classes − 1 (minus one more for each parameter estimated from the data, e.g. allele frequency in a Hardy–Weinberg test).
  • Test of independence (contingency table, r × c): expected count = (row total × column total) / grand total; df = (r − 1)(c − 1).
  • Use counts, never percentages; expected counts should not be too small (a common rule is at least 5 per cell).

Worked example. 100 mice are challenged with a pathogen: 50 unvaccinated (30 infected) and 50 vaccinated (10 infected).

InfectedNot infectedTotal
Unvaccinated302050
Vaccinated104050
Total4060100

Expected: 20 and 30 in each row. χ² = (10²/20 + 10²/30) × 2 = (5 + 3.33) × 2 = 16.7, df = 1, far above the 0.05 critical value 3.84: infection is not independent of vaccination status.

Correlation and regression

  • Pearson’s r (−1 to +1) measures the strength of a linear association; r² is the fraction of variance in y explained by x. Spearman’s rank correlation for monotonic, non-normal data. Correlation is not causation.
  • Least-squares regression \(y = a + bx\):
$$b = \frac{\sum (x - \bar{x})(y - \bar{y})}{\sum (x - \bar{x})^2}, \qquad a = \bar{y} - b\bar{x}$$

Used for standard curves (Bradford, ELISA, qPCR efficiency), allometry and dose–response. Do not extrapolate beyond the range of x; inspect residuals.

Non-parametric and multivariate basics

  • Non-parametric tests use ranks and fewer assumptions: Mann–Whitney U (two independent groups), Wilcoxon signed-rank (paired), Kruskal–Wallis (several groups), Spearman’s ρ.
  • Multivariate methods analyse many variables at once:
    • PCA (principal component analysis): rotates correlated variables into uncorrelated principal components ordered by variance explained; used to visualise omics data, populations, morphometrics.
    • Cluster analysis (hierarchical, k-means): groups similar samples or genes; heat maps with dendrograms.
    • Discriminant analysis / PLS-DA: finds combinations that separate predefined groups.
    • MANOVA: ANOVA with several dependent variables; multiple regression with several predictors.
    • Ordination (PCA, correspondence analysis, NMDS) in community ecology.
  • Multiple testing: with m independent tests at α = 0.05, about 0.05m false positives are expected. Bonferroni: use α/m (controls family-wise error; conservative). False discovery rate (Benjamini–Hochberg) controls the expected proportion of false positives among discoveries, standard for omics.

Critical values used on this page (α = 0.05; two-tailed for t and z, upper tail for χ² and F): t(df 5) = 2.571, t(df 8) = 2.306, t(df 24) = 2.064; χ²(df 1) = 3.841, χ²(df 2) = 5.991, χ²(df 3) = 7.815; F(2, 6) = 5.14; z = 1.96.


Part 2: Intellectual Property Rights

Forms of IP

RightProtectsNotes
PatentInventions: products and processesExclusive right for 20 years from filing, in exchange for public disclosure; territorial
CopyrightOriginal expression (text, software code, figures, databases as compilations)Not ideas or facts
TrademarkBrand names, logosRenewable
Geographical indication (GI)Products tied to a place (e.g. Darjeeling tea)Collective right
Industrial designAppearance of a product
Trade secretConfidential know-howNo registration; lost if disclosed
Plant variety protectionNew plant varieties; breeders’ and farmers’ rightsIndia: Protection of Plant Varieties and Farmers’ Rights Act, 2001

What makes an invention patentable

  1. Novelty: not disclosed anywhere before the filing date (prior art includes your own publication, poster or thesis).
  2. Inventive step (non-obviousness): not obvious to a person skilled in the art.
  3. Industrial applicability (utility).
  4. Sufficient disclosure: the specification must enable a skilled person to perform the invention; for biological material that cannot be described in words, a deposit in a recognised depository under the Budapest Treaty satisfies this.
  • Publish after filing: publishing first can destroy novelty. A provisional specification secures a priority date; the complete specification follows within the allowed period.
  • The Patent Cooperation Treaty (PCT) offers one international application route, but patents are still granted nationally.
  • TRIPS (WTO) set minimum IP standards; India amended the Patents Act, 1970 (notably in 2005) to grant product patents for drugs, food and chemicals.

Biology-specific rules in India (Patents Act, Section 3)

  • Section 3(c): mere discovery of a living thing or non-living substance occurring in nature is not an invention.
  • Section 3(d): a new form of a known substance is not patentable unless it shows enhanced efficacy; this was central to the Supreme Court’s 2013 decision against a patent on a new crystalline form of imatinib (Glivec).
  • Section 3(i): processes for the medicinal, surgical, curative, prophylactic, diagnostic or therapeutic treatment of humans (or similar treatment of animals) are excluded.
  • Section 3(j): plants and animals in whole or part (other than micro-organisms), seeds, varieties and species, and essentially biological processes for producing plants or animals are excluded. Genetically modified micro-organisms can be patented.
  • Internationally, Diamond v. Chakrabarty (US Supreme Court, 1980) allowed a patent on a genetically engineered oil-degrading bacterium: “anything under the sun made by man”.
  • Biopiracy and traditional knowledge: patents based on traditional uses of turmeric (wound healing) and neem were challenged and revoked; India’s Traditional Knowledge Digital Library documents such knowledge as prior art. The Biological Diversity Act, 2002 and the Nagoya Protocol require access and benefit-sharing for biological resources.

Part 3: Biosafety

Biosafety levels (BSL) and risk groups

LevelAgents (risk group)ExamplesKey containment
BSL-1Not known to cause disease in healthy adultsE. coli K-12, Bacillus subtilis, Saccharomyces cerevisiaeOpen bench, good microbiological practice, handwashing
BSL-2Moderate hazard; treatable or preventableStaphylococcus aureus, Salmonella, hepatitis B virusRestricted access, Class II biosafety cabinet for aerosols, autoclave available, PPE
BSL-3Serious or lethal disease via aerosol, treatment usually availableMycobacterium tuberculosis, SARS-CoV-2 (culture work)Controlled access, directional inward airflow, HEPA-filtered exhaust, all open work in BSCs, respiratory protection
BSL-4Dangerous, often fatal, no vaccine or treatment, aerosol riskEbola, Marburg virusesIsolated building or zone, Class III cabinets or positive-pressure suits, airlocks, chemical showers, effluent decontamination

Biological safety cabinets (BSCs):

ClassProtectsDesign
IWorker and environment, not the sampleInward airflow, HEPA exhaust
II (most common)Worker, sample and environmentInward airflow plus HEPA-filtered downflow over the work surface
IIIMaximum: total containmentGas-tight glove box, for BSL-4 agents

A horizontal laminar-flow clean bench blows filtered air towards the user: it protects the sample (plant tissue culture, media pouring) but not the worker, so it must never be used for pathogens.

  • Primary containment (BSCs, sealed rotors, PPE) protects the worker; secondary containment (facility design, airflow, autoclaves) protects the outside world.
  • Decontamination: autoclaving (commonly 121 °C, about 15 psi, typically 15–20 min or longer for loads), chemical disinfectants (hypochlorite, 70% ethanol), incineration; segregated biomedical waste.

Regulation of GMOs and hazardous organisms in India

The Rules, 1989 for hazardous micro-organisms and genetically engineered organisms, framed under the Environment (Protection) Act, 1986, set up a tiered committee system:

CommitteeRole
IBSC (Institutional Biosafety Committee)In each institution; first-level review of rDNA work
RCGM (Review Committee on Genetic Manipulation, DBT)Reviews higher-risk research and confined field trials
GEAC (Genetic Engineering Appraisal Committee, MoEFCC)Approves large-scale use, environmental release and commercialisation (e.g. Bt cotton, approved in 2002)
RDAC, SBCC, DLCAdvisory and state/district-level monitoring

Internationally, the Cartagena Protocol on Biosafety (under the Convention on Biological Diversity) governs transboundary movement of living modified organisms using the precautionary approach and advance informed agreement.


Part 4: Research and Publication Ethics

Animal research

  • 3Rs (Russell and Burch, 1959): Replacement (cells, organoids, computer models, lower organisms), Reduction (fewer animals through better design and statistics), Refinement (minimise pain and distress: anaesthesia, analgesia, humane endpoints, enrichment).
  • In India, animal experiments require approval of an Institutional Animal Ethics Committee (IAEC) working under the national committee constituted under the Prevention of Cruelty to Animals Act, 1960 (known as CPCSEA, the name used in textbooks and past exam questions).
  • Reporting standards (such as ARRIVE guidelines) ask for sample-size justification, randomisation and blinding.

Human research

  • Nuremberg Code (voluntary consent), Declaration of Helsinki (World Medical Association), and the Belmont Report principles: respect for persons (informed consent), beneficence (maximise benefit, minimise harm), justice (fair selection and distribution of burdens).
  • In India, the ICMR National Ethical Guidelines govern biomedical and health research involving human participants; every study needs Institutional Ethics Committee approval, written informed consent (with special protection for vulnerable groups), privacy and confidentiality of data and samples. Clinical trials are registered in the Clinical Trials Registry – India (CTRI).

Research misconduct and publication ethics

IssueWhat it is
FabricationMaking up data or results
FalsificationManipulating materials, equipment, processes or images, or omitting data, so the record misrepresents the research
PlagiarismUsing others’ words, ideas, data or figures without credit; includes self-plagiarism (text recycling without disclosure)
Duplicate/redundant publicationSame results published twice; salami slicing splits one study into minimal units
Authorship misconductGift/honorary authors (no real contribution) and ghost authors (contributors left out); ICMJE criteria require substantial contribution, drafting/revision, approval and accountability
Conflict of interestUndisclosed financial or personal interests
Image manipulationSplicing, duplicating or selectively enhancing parts of gels or micrographs
Peer-review misconductBreaching confidentiality, stealing ideas, fake reviewers
Predatory journalsCharge fees with little or no peer review

Fabrication, falsification and plagiarism (FFP) are the core definition of research misconduct. Corrections and retractions repair the record; COPE (Committee on Publication Ethics) provides guidance to journals.

Plagiarism in Indian higher education. The UGC’s 2018 regulations on academic integrity and prevention of plagiarism grade similarity in theses and papers into levels (similarity up to 10% is not penalised; higher bands carry escalating penalties), with exclusions for quotations with attribution, references and common phrases. Similarity-checking software flags overlap; a human must judge whether it is plagiarism.

AI in research

  • Uses: structure prediction (AlphaFold’s impact was recognised in the 2024 Nobel Prize in Chemistry), image analysis and segmentation, omics analysis, literature search, coding help and language editing.
  • Responsibilities:
    • AI tools cannot be authors: they cannot take responsibility or approve a manuscript (position taken by COPE, ICMJE and major publishers); disclose substantive use in methods or acknowledgements.
    • Authors remain accountable for accuracy: generative tools can produce plausible but false statements and non-existent references.
    • Confidentiality: do not upload unpublished manuscripts, grant proposals under review, or identifiable patient data to public AI tools.
    • Many journals prohibit AI-generated or AI-altered images presented as data.
    • Bias and reproducibility: models reflect their training data; record versions, prompts and parameters; validate predictions experimentally.

See also Research Ethics and Scientific Communication for Part A’s coverage.


Memory Tricks

  • Poisson’s signature: mean = variance. Variance/mean > 1 means clumping (also used in field ecology).
  • 3Rs: Replace, Reduce, Refine.
  • FFP = Fabrication, Falsification, Plagiarism.

Don’t Confuse

ThisvsThat
Accuracy: near the true value (bias)vsPrecision: near each other (random error)
SD: spread of valuesvsSE: uncertainty of the mean
Technical replicatesvsBiological replicates (the true n)
Paired t-test: same subjectsvsUnpaired t-test: independent groups
ANOVA: ≥ 3 meansvsχ²: counts in categories
Correlation: association strengthvsRegression: predicting y from x
Type I (α): false positivevsType II (β): false negative
Bonferroni: family-wise errorvsFDR: proportion of false discoveries
Patent: inventions, 20 yearsvsCopyright: expression, not ideas
Discovery (Sec. 3(c), not patentable)vsInvention (GM micro-organism, patentable)
Class II BSC: protects worker and samplevsLaminar-flow clean bench: sample only
IBSC / RCGM (research)vsGEAC (environmental release, commercialisation)
Fabrication: inventing datavsFalsification: distorting real data

Common Mistakes


Practice

All questions below are original practice questions written in CSIR NET style. They are not previous-year questions.

Level 1: Foundation

  1. What is the variance of a Poisson distribution with mean 4?
  2. Which test compares the means of three or more independent groups?
  3. Which biosafety cabinet class protects worker, sample and environment?
Answers
  1. 4 (variance equals the mean).
  2. One-way ANOVA.
  3. Class II.

Level 2: Part B Style

Q1 (practice). Twenty-five plants have a mean height of 50 cm with SD 10 cm. Using t(0.05, 24) = 2.064, the 95% confidence interval for the mean is:

  • (a) 50 ± 20.6 cm
  • (b) 50 ± 10 cm
  • (c) 50 ± 4.1 cm
  • (d) 50 ± 2.0 cm
Answer and explanation

(c). SE = 10/√25 = 2 cm; half-width = 2.064 × 2 = 4.13 cm. (b) uses the SD, (d) forgets the t multiplier.

Q2 (practice). A dihybrid F₂ gives 190 : 58 : 52 : 20 (total 320) against an expected 9 : 3 : 3 : 1. The critical χ² at α = 0.05 for 3 df is 7.815. The χ² value and conclusion are:

  • (a) 1.69; the data fit 9 : 3 : 3 : 1
  • (b) 1.69; the data deviate significantly
  • (c) 16.9; the data deviate significantly
  • (d) 5.07; the data fit 9 : 3 : 3 : 1
Answer and explanation

(a). Expected = 180, 60, 60, 20. χ² = 10²/180 + 2²/60 + 8²/60 + 0 = 0.556 + 0.067 + 1.067 + 0 = 1.69 < 7.815; deviations are consistent with chance.

Q3 (practice). Phage are added to bacteria at an average of 2 phage per cell, distributed at random. The fraction of cells that receive no phage is about:

  • (a) 0.02
  • (b) 0.14
  • (c) 0.27
  • (d) 0.50
Answer and explanation

(b). Poisson with λ = 2: P(0) = e⁻² = 0.135. About 86% of cells are infected by at least one phage.

Q4 (practice). Under the Indian Patents Act, 1970, which of the following is most likely to be patentable?

  • (a) A newly discovered soil bacterium isolated as it occurs in nature
  • (b) A transgenic rice variety with improved drought tolerance
  • (c) A genetically engineered bacterial strain that degrades a pesticide, with a new process for using it
  • (d) A new salt of a known drug with no improvement in therapeutic efficacy
Answer and explanation

(c). An engineered micro-organism and process can be inventions. (a) is a discovery (Section 3(c)); (b) plants and varieties are excluded (Section 3(j); plant varieties are protected instead under the PPV&FR Act); (d) fails Section 3(d) without enhanced efficacy.

Level 3: Part C Style

Q1 (practice). An enzyme activity (U/mg) is measured in five control and five knockout mice:

  • Control: 12, 14, 11, 13, 15
  • Knockout: 16, 18, 15, 17, 19

(i) Compute the means, variances and the pooled-variance t statistic. (ii) Using t(0.05, two-tailed, df 8) = 2.306, what do you conclude? (iii) If the same ten values had come from five mice measured before and after a drug (pairing control 12 with knockout 16, and so on), would a paired test be more or less powerful here, and why?

Answer and worked solution
  1. Means: control 13, knockout 17. Variances: control: deviations −1, 1, −2, 0, 2 → squares sum 10 → s² = 10/4 = 2.5; knockout: same pattern → 2.5. Pooled s² = 2.5. \(t = \dfrac{17 - 13}{\sqrt{2.5\,(1/5 + 1/5)}} = \dfrac{4}{\sqrt{1}} = \mathbf{4.0}\), df = 8.
  2. 4.0 > 2.306: reject H₀. Knockout activity is significantly higher (p < 0.05).
  3. More powerful. Every paired difference is exactly +4 (16 − 12, 18 − 14, …), so the SD of differences is 0 and the paired t is effectively unbounded: pairing removes all between-animal variation. Pairing is only legitimate when the design really pairs measurements (same animal, littermates); it cannot be imposed on independent groups after the fact.

Q2 (practice). (a) Three diets are fed to three rats each; weight gains (g) are Diet 1: 4, 5, 6; Diet 2: 7, 8, 9; Diet 3: 10, 11, 12. Compute F for a one-way ANOVA and decide using F(2, 6) = 5.14 at α = 0.05.

(b) A Bradford standard curve gives:

Protein (µg)01234
A₅₉₅0.020.210.390.620.80

Fit y = a + bx and find the protein in a sample with A₅₉₅ = 0.50.

Answer and worked solution

(a) Group means 5, 8, 11; grand mean 8.

  • SS_between = 3[(5 − 8)² + 0² + (11 − 8)²] = 3 × 18 = 54; df = 2; MS = 27.
  • SS_within = each group has squared deviations 1 + 0 + 1 = 2, total 6; df = 9 − 3 = 6; MS = 1.
  • F = 27/1 = 27 > 5.14: diets differ significantly. A post hoc test (e.g. Tukey) would identify which pairs differ (here all pairs differ by 3 g with very little within-group scatter).

(b) x̄ = 2, ȳ = 0.408. Σ(x − x̄)(y − ȳ) = 1.97; Σ(x − x̄)² = 10.

  • b = 1.97/10 = 0.197 per µg; a = 0.408 − 0.197 × 2 = 0.014. (r ≈ 0.999.)
  • x = (0.50 − 0.014)/0.197 = 2.47 µg, within the calibrated range, so interpolation is valid.

Q3 (practice). A lab plans to (1) culture Mycobacterium tuberculosis to test a new compound, (2) test the compound in mice, (3) patent it and (4) publish, using a generative AI tool to polish the manuscript. Which statements are correct?

(A) The culture work requires BSL-3 containment with work inside a biological safety cabinet; a horizontal laminar-flow bench is not acceptable. (B) The mouse study needs prior approval from the Institutional Animal Ethics Committee, and sample size should be justified to satisfy the “Reduction” principle. (C) Presenting the results at a conference before filing the patent application does not affect novelty because the lab is the inventor. (D) The AI tool may be acknowledged and its use disclosed, but it cannot be listed as an author, and the authors remain responsible for every statement and reference.

  • (a) A, B and D only
  • (b) A and C only
  • (c) B, C and D only
  • (d) A, B, C and D
Answer and worked solution

(a) A, B and D only.

  • (A) Correct: M. tuberculosis is handled at BSL-3; a clean bench blows air at the worker and offers no personnel protection.
  • (B) Correct: IAEC approval (under CPCSEA) is mandatory; power analysis supports Reduction.
  • (C) Wrong: the inventor’s own public disclosure is prior art; file first, then present.
  • (D) Correct: consistent with COPE/ICMJE positions on AI tools.

Quick Revision Table

ItemFact
Accuracy / precisionBias (systematic) / random error
SD / SE / CVs; s/√n; (s/x̄) × 100%
SkewRight skew: mode < median < mean
Normal68–95–99.7; 95% at ±1.96σ
Binomial / PoissonMean np, var np(1−p) / mean = var = λ; P(0) = e⁻ˡ
CIx̄ ± t·s/√n
ErrorsType I = α (false +); Type II = β (false −); power = 1 − β
t-testsOne-sample, unpaired (df n₁ + n₂ − 2), paired (df n − 1)
ANOVAF = MS_between/MS_within; post hoc for which pairs; two-way for interaction
χ²Σ(O − E)²/E; df = classes − 1, or (r − 1)(c − 1); counts only
Critical valuesχ² 3.841 (1 df), 5.991 (2), 7.815 (3); z 1.96
Regressionb = Sxy/Sxx; a = ȳ − bx̄; r² = variance explained
Non-parametricMann–Whitney, Wilcoxon, Kruskal–Wallis, Spearman
MultivariatePCA, clustering, discriminant/PLS-DA, MANOVA
Multiple testingBonferroni α/m; FDR (Benjamini–Hochberg)
PatentNovelty, inventive step, industrial application, disclosure; 20 years; Budapest Treaty deposits
India Sec. 33(c) discoveries; 3(d) new forms need enhanced efficacy; 3(i) treatment methods; 3(j) plants, animals, essentially biological processes
Other IPCopyright, trademark, GI, design, trade secret, PPV&FR Act 2001
BSL1 E. coli K-12; 2 S. aureus; 3 M. tuberculosis; 4 Ebola, Marburg
BSCI worker + environment; II worker + sample + environment; III glove box
India GMORules 1989 under EPA 1986; IBSC → RCGM (DBT) → GEAC (MoEFCC)
InternationalCartagena Protocol (biosafety); Nagoya Protocol (access and benefit-sharing)
Animal ethics3Rs (Russell and Burch 1959); IAEC; CPCSEA
Human ethicsNuremberg, Helsinki, Belmont (respect, beneficence, justice); ICMR guidelines; CTRI
MisconductFFP; duplicate publication; gift/ghost authorship; image manipulation
AINot an author; disclose; authors accountable; confidentiality

Key Takeaways

What’s Next

Advertisement
Advertisement