uiz Space

January 2024 term · Big Data and Biological Networks · BSBT4002

Big Data and Biological Networks End Term: 28 April 2024 (January 2024 term)

The IIT Madras BS Big Data and Biological Networks (Big Data and Biological Networks) End Term paper sat on 28 Apr 2024, in the January 2024 term: 42 questions for 100 marks in 180 minutes. Every question is below with its answer. Take it as a timed mock test to be marked, or read it through first.

Questions
42
Marks
100
Duration
180 min
MSQ
12
MCQ
30
Take as mock testPractise with answers

Updated

Official paper: IIT M DEGREE FN EXAM QDB1 28 Apr 2024 · No negative marking.

Question 1

+3 marksOne or more correct options

In the graph shown below

Select all that apply.

  1. A

    There is an Eulerian path starting from the node "a"

  2. B

    There is an Eulerian cycle starting from the node "a"

  3. C

    There are Eulerian paths from all the nodes

  4. D

    There are no Eulerian cycles

Show answer

Correct answers

  • A

    There is an Eulerian path starting from the node "a"

  • B

    There is an Eulerian cycle starting from the node "a"

  • C

    There are Eulerian paths from all the nodes

Question 1 on its own page

Question 2

+3 marksOne or more correct options

Assume you have been provided with well-curated highly accurate constraint-based genome-scale metabolic models. Which of the following studies can be carried out on the models provided

Select all that apply.

  1. A

    Metabolic engineering to increase yields of a target metabolite

  2. B

    Temporal evolution of cellular phenotypes

  3. C

    Kinetics of enzyme catalysed reactions

  4. D

    Prediction of interspecies interactions

Show answer

Correct answers

  • A

    Metabolic engineering to increase yields of a target metabolite

  • D

    Prediction of interspecies interactions

Question 2 on its own page

Question 3

+3 marksOne or more correct options

Signal transduction networks have long cascades. What advantages do they provide?

Select all that apply.

  1. A

    Gives more time for protein synthesis

  2. B

    Long cascades form more efficient modules

  3. C

    Many levels of control to ensure better regulation

  4. D

    Signal amplification from small input to high output of final product(s)

Show answer

Correct answers

  • C

    Many levels of control to ensure better regulation

  • D

    Signal amplification from small input to high output of final product(s)

Question 3 on its own page

Question 4

+3 marksOne or more correct options

The number of stuck/blocked reactions in the microbe ‘A’ when grown alone is 27. The number of stuck reactions is reduced to 15 when ‘A’ is grown together with ‘B’. Also, when ‘A’ is grown together with ‘C’, the number of stuck reactions in ‘A’ is 7. Which of the following is/are true?

Select all that apply.

  1. A

    MSI(A;AUB) = 0.55

  2. B

    MSI(A;AUB) = 0.44

  3. C

    MSI(A;AUC) = 0.74

  4. D

    MSI(A;AUC) = 0.26

Show answer

Correct answers

  • B

    MSI(A;AUB) = 0.44

  • C

    MSI(A;AUC) = 0.74

Question 4 on its own page

Question 5

+3 marksOne correct option

Consider the given below network and identify the possible gene-regulatory interactions

  1. A

    Gene 3 -> Gene 2

  2. B

    Gene 3 -> Gene 1

  3. C

    Gene 2 -> Gene 1

  4. D

    Gene 2 -> Gene 3

Show answer

Correct answer

  • A

    Gene 3 -> Gene 2

Question 5 on its own page

Question 6

+3 marksOne correct option

Which network model is obtained using the “preferential attachment” method of network growth?

  1. A

    Random network

  2. B

    Power law network

  3. C

    Erdős–Rényi network

  4. D

    Small world network

Show answer

Correct answer

  • B

    Power law network

Question 6 on its own page

Question 7

+3 marksOne correct option

Consider a toy model with 4 reactions, namely R1, R2, R3 and R4. The GPR (gene protein reaction) rules for each of the reactions are given below:
a) R1 = gene 1 OR gene 2
b) R2 = NOT gene 3
c) R3 = gene 4 AND gene 5
d) R4 = gene 1
If gene 1 and gene 4 are knocked out, which of the reactions mentioned below will be blocked?

  1. A

    R1 and R3

  2. B

    R2 and R3

  3. C

    R4 and R1

  4. D

    R4 and R3

Show answer

Correct answer

  • D

    R4 and R3

Question 7 on its own page

Question 8

+3 marksOne correct option

Based on the phylogenetic profiles given below, indicate which of the following proteins pairs interact

  1. A

    A and B

  2. B

    B and C

  3. C

    A and D

  4. D

    C and D

Show answer

Correct answer

  • C

    A and D

Question 8 on its own page

Question 9

+1 markOne correct option

Choose the correct pairs:

  1. A

    Drug-Drug interaction → Multi-label graph classification

  2. B

    Molecule side effect prediction → Graph classification

  3. C

    Protein interface prediction → Node prediction

  4. D

    Drug-Target binding affinity prediction → Graph classification

Show answer

Correct answer

  • B

    Molecule side effect prediction → Graph classification

Question 9 on its own page

Question 10

+1 markOne correct option
  1. A
  2. B
  3. C
  4. D
Show answer

Correct answer

  • A
Question 10 on its own page

Question 11

+1 markOne correct option

Enumerate the total number of solutions that are possible for a string reconstruction problem with 21 words?

  1. A
  2. B
  3. C
  4. D
Show answer

Correct answer

  • D
Question 11 on its own page

Question 12

+2 marksOne correct option

Which of the following statements about GraphSAGE and GAT (Graph Attention Network) is true?

  1. A

    GraphSAGE uses attention mechanisms to aggregate information from neighboring nodes

  2. B

    GraphSAGE is a node-level representation learning algorithm, while GAT is an edge-level representation learning algorithm

  3. C

    GAT assigns different importance weights to neighboring nodes during aggregation based on their relevance to the target node

  4. D

    GraphSAGE is a edge-level representation learning algorithm, while GAT is an node-level representation learning algorithm

Show answer

Correct answer

  • C

    GAT assigns different importance weights to neighboring nodes during aggregation based on their relevance to the target node

Question 12 on its own page

Question 13

+2 marksOne correct option

Which of the following is true?

  1. A

    Graph pooling refines the node features without changing the graph structure

  2. B

    For node-focused enumerate, a representation of the entire graph is desired

  3. C

    The assignment matrix is used to map super nodes to the original graph

  4. D

    Eigenpooling uses spectral clustering to generate super nodes

Show answer

Correct answer

  • D

    Eigenpooling uses spectral clustering to generate super nodes

Question 13 on its own page

Question 14

+2 marksOne correct option
  1. A
  2. B
  3. C
  4. D
Show answer

Correct answer

  • A
Question 14 on its own page

Question 15

+2 marksOne correct option

Why do we need specialized deep learning architectures for graphs?

  1. A

    The flattened adjacency matrix does not change with ordering of nodes

  2. B

    Graphs have variable-sized structures and irregular connectivity patterns

  3. C

    Feed-forward networks are permutation equivariant functions

  4. D

    All of these

Show answer

Correct answer

  • B

    Graphs have variable-sized structures and irregular connectivity patterns

Question 15 on its own page

Question 16

+2 marksOne correct option

Given two sequences TATTT and ATAT, in the graph representation of their sequence alignment, how many nodes are present?

  1. A

    7

  2. B

    36

  3. C

    30

  4. D

    45

Show answer

Correct answer

  • C

    30

Question 16 on its own page

Question 17

+2 marksOne or more correct options

In the case of flat graph pooling:

Select all that apply.

  1. A

    Directly generates a graph-level representation from the node representations

  2. B

    No new graph but a single node being generated

  3. C

    Used for node focused enumerate

  4. D

    Average pooling is a flat graph pooling method

Show answer

Correct answers

  • A

    Directly generates a graph-level representation from the node representations

  • B

    No new graph but a single node being generated

  • D

    Average pooling is a flat graph pooling method

Question 17 on its own page

Question 18

+2 marksOne or more correct options

Task-specific labels are scarce during the training phase. The pretraining helps to develop a general understanding of related enumerate with abundant unlabelled data. Then, fine-tuning can be performed for a downstream task of interest. For the molecular property prediction task, which of the following are valid pretraining strategies.

Select all that apply.

  1. A

    Reconstructing the adjacency matrix for each graph using an encoder-decoder approach

  2. B

    Mask nodes and predict the features of masked nodes during training

  3. C

    Predict the embedding similarity between different graphs

  4. D

    Use subgraphs to predict surrounding graph structures

Show answer

Correct answers

  • A

    Reconstructing the adjacency matrix for each graph using an encoder-decoder approach

  • B

    Mask nodes and predict the features of masked nodes during training

  • D

    Use subgraphs to predict surrounding graph structures

Question 18 on its own pageAlso asked in End Term 13 Apr 2025

Question 19

+2 marksOne or more correct options

Which of the following are valid regularization approaches for graphs:

Select all that apply.

  1. A

    Add L1 norm of trainable parameters to supervised loss term

  2. B

    Perturbing node features during training by adding gaussian noise

  3. C

    Force representations of nodes that are not connected to be different

  4. D

    Randomly drop nodes during training

Show answer

Correct answers

  • A

    Add L1 norm of trainable parameters to supervised loss term

  • B

    Perturbing node features during training by adding gaussian noise

  • C

    Force representations of nodes that are not connected to be different

Question 19 on its own page

Question 20

+2 marksOne or more correct options

Which of the following statements is not true?

Select all that apply.

  1. A

    Global alignment is more applicable for sequences with similar lengths

  2. B

    Traceback in the score matrix is the only place that the Smith-Waterman algorithm differs from the Needleman-Wunsch algorithm.

  3. C

    Affine gap penalties are more widely used than constant gap penalties

  4. D

    Sequence alignment cannot be used for identifying frameshift mutations

Show answer

Correct answers

  • A

    Global alignment is more applicable for sequences with similar lengths

  • B

    Traceback in the score matrix is the only place that the Smith-Waterman algorithm differs from the Needleman-Wunsch algorithm.

  • D

    Sequence alignment cannot be used for identifying frameshift mutations

Question 20 on its own page

Question 21

+3 marksOne correct option

Choose the most appropriate option:

Embedding MethodLearning ParadigmFeature Matrix
(1) node2vec(A) Supervised(i) X=IX = I
(2) Laplacian Eigenmaps(B) Unsupervised(ii) X≠IX \neq I
(3) GraphSAGE
(4) GCN

XX is the feature matrix and II is the identity matrix

  1. A

    (1), (A), (i); (2), (B), (ii); (3), (A), (ii); (4), (B), (i);

  2. B

    (1), (B), (i); (2), (B), (i); (3), (A), (ii); (4), (A), (ii);

  3. C

    (1), (A), (ii); (2), (A), (i); (3), (B), (i); (4), (A), (i);

  4. D

    (1), (B), (ii); (2), (B), (i); (3), (B), (i); (4), (A), (ii);

Show answer

Correct answer

  • B

    (1), (B), (i); (2), (B), (i); (3), (A), (ii); (4), (A), (ii);

Question 21 on its own page

Question 22

+3 marksOne correct option
  1. A
  2. B
  3. C
  4. D
Show answer

Correct answer

  • D
Question 22 on its own page

Question 23

+3 marksOne correct option
  1. A
  2. B
  3. C
  4. D
Show answer

Correct answer

  • D
Question 23 on its own page

Question 24

+3 marksOne correct option

A two-layer neural network with the DD features, x1, x2, …, xDx_1,\ x_2,\ \ldots,\ x_D is defined below:

h1=W1x+b1z1=ReLU(h1)h2=W2z1+b2y^=σ(h2)ℓ=−y.log(y^)−(1−y).log(1−y^)\begin{aligned} h_1 &= W_1 x + b_1 \\ z_1 &= ReLU(h_1) \\ h_2 &= W_2 z_1 + b_2 \\ \hat{y} &= \sigma(h_2) \\ \ell &= -y.log(\hat{y}) - (1 - y).log(1 - \hat{y}) \end{aligned}

compute ∂h2∂b2\frac{\partial h_2}{\partial b_2}, ∂z1∂h1\frac{\partial z_1}{\partial h_1} and ∂h1∂W1\frac{\partial h_1}{\partial W_1} .

  1. A
  2. B
  3. C
  4. D
Show answer

Correct answer

  • A
Question 24 on its own page

Question 25

+3 marksOne correct option

A two-layer neural network with the DD features, x1, x2, …, xDx_1,\ x_2,\ \ldots,\ x_D, is defined below:

h1=W1x+b1z1=ReLU(h1)h2=W2z1+b2y^=σ(h2)ℓ=−y.log(y^)−(1−y).log(1−y^)\begin{aligned} h_1 &= W_1 x + b_1 \\ z_1 &= ReLU(h_1) \\ h_2 &= W_2 z_1 + b_2 \\ \hat{y} &= \sigma(h_2) \\ \ell &= -y.log(\hat{y}) - (1 - y).log(1 - \hat{y}) \end{aligned}

Compute ∂ℓ∂y^\frac{\partial \ell}{\partial \hat{y}} and ∂y^∂h2\frac{\partial \hat{y}}{\partial h_2} .

  1. A
  2. B
  3. C
  4. D
Show answer

Correct answer

  • A
Question 25 on its own page

Question 26

+3 marksOne correct option

What is the average degree (regardless of the direction) of the 4-mer overlap graph constructed from “GAATTCAGTTA”?

  1. A

    2.22

  2. B

    2.45

  3. C

    3.12

  4. D

    1.75

Show answer

Correct answer

  • D

    1.75

Question 26 on its own page

Question 27

+3 marksOne correct option

What is the best local alignment score for the sequences "GTACTG" and "TAGTCTG" with the scoring scheme: matches = +2, mismatches = -1 and a constant gap penalty of -10.

  1. A

    -3

  2. B

    4

  3. C

    -7

  4. D

    6

Show answer

Correct answer

  • C

    -7

Question 27 on its own page

Question 28

+3 marksOne correct option

Elucidate the final sequence that is obtained from the kmers: ’ATCG’, ’TACA’, ’TCGG’, ’CGGT’, ’CAAC’, ’GGTA’, GTAC’, ’ACAA’, ’AACT’, ’ACTC’.

  1. A

    ATCTGTACAACTC

  2. B

    ATCGGTACAACTC

  3. C

    ATCGGGACAACTC

  4. D

    ATCGGTATAACTC

Show answer

Correct answer

  • B

    ATCGGTACAACTC

Question 28 on its own page

Question 29

+3 marksOne correct option

Consider the following statements:
Assertion (A): Eulerian Path algorithm is preferred over the Hamiltonian Path algorithm for the genome assembly problem.
Reason (R): A Hamiltonian Path passes through all the nodes of a graph.
Now which of the following options is true:

  1. A

    Both A and R are true and R is the correct explanation for A

  2. B

    Both A and R are true and R is not the correct explanation for R

  3. C

    A is true but R is false

  4. D

    Both the statements are false

Show answer

Correct answer

  • B

    Both A and R are true and R is not the correct explanation for R

Question 29 on its own page

Question 30

+3 marksOne or more correct options

What are the key features of convolutional layers in graph neural networks?

Select all that apply.

  1. A

    Equivariant representation

  2. B

    Sparse connections

  3. C

    Parameter sharing

  4. D

    Can be extended to data with high dimensions

Show answer

Correct answers

  • A

    Equivariant representation

  • B

    Sparse connections

  • C

    Parameter sharing

  • D

    Can be extended to data with high dimensions

Question 30 on its own page

Question 31

+3 marksOne or more correct options

What are the limitations of shallow embeddings?

Select all that apply.

  1. A

    Shallow embedding methods share parameters between nodes in the encoder

  2. B

    Shallow embedding approaches do not leverage node features in the encoder

  3. C

    Shallow embedding methods can only generate embeddings for nodes that were present during the training phase

  4. D

    Shallow embedding methods directly optimizes a unique embedding vector for each node

Show answer

Correct answers

  • B

    Shallow embedding approaches do not leverage node features in the encoder

  • C

    Shallow embedding methods can only generate embeddings for nodes that were present during the training phase

  • D

    Shallow embedding methods directly optimizes a unique embedding vector for each node

Question 31 on its own pageAlso asked in End Term 13 Apr 2025

Question 32

+3 marksOne or more correct options

Which of the following are true for graph filters:

Select all that apply.

  1. A

    Spatial-based graph filters explicitly leverage the graph structure

  2. B

    Spectral-based graph filters utilize spectral graph theory to design the filtering operation in the spectral domain

  3. C

    Cheby-Filter is a spatial-based graph filter

  4. D

    GAT-Filter is a spatial-based graph filter

Show answer

Correct answers

  • A

    Spatial-based graph filters explicitly leverage the graph structure

  • B

    Spectral-based graph filters utilize spectral graph theory to design the filtering operation in the spectral domain

  • D

    GAT-Filter is a spatial-based graph filter

Question 32 on its own pageAlso asked in End Term 13 Apr 2025

Question 33

+3 marksOne or more correct options

Which of the following can we use alignment for?

Select all that apply.

  1. A

    To identify mutations in a given sequence

  2. B

    To search in a database based on a query

  3. C

    To discover conserved regions in a gene

  4. D

    Generating overlap graphs from k-mer reads

Show answer

Correct answers

  • A

    To identify mutations in a given sequence

  • B

    To search in a database based on a query

  • C

    To discover conserved regions in a gene

  • D

    Generating overlap graphs from k-mer reads

Question 33 on its own page

Question 34

+2 marksOne correct option
  1. A

    i-a, ii-b, iii-c iv-d

  2. B

    i-d, ii-c, iii-a, iv-b

  3. C

    i-b, ii-c, iii-d, iv-a

  4. D

    i-c, ii-d, iii-b, iv-a

Show answer

Correct answer

  • C

    i-b, ii-c, iii-d, iv-a

Question 34 on its own page

Question 35

+2 marksOne correct option
  1. A

    Both the assertion and reason are true and the reason is the correct explanation for the assertion

  2. B

    Both the assertion and reason are true. But the reason is not correct regarding the assertion.

  3. C

    Only the Assertion is true.

  4. D

    Only the Reason is true.

Show answer

Correct answer

  • A

    Both the assertion and reason are true and the reason is the correct explanation for the assertion

Question 35 on its own page

Question 36

+2 marksOne correct option

Match the GWAS term with its description.

Column AColumn B
ASNP1Visualization of genomic regions associated with traits or diseases
BP-value2A statistical measure indicating the strength of association between a genetic variant and a trait.
CManhattan plot3Patterns of non-random association of alleles at different loci.
DLinkage disequilibrium4Common type of genetic variation analysed in GWAS
  1. A

    A - 2, B - 1, C - 4, D - 3

  2. B

    A - 4, B - 3, C - 1, D - 2

  3. C

    A - 4, B - 2, C - 1, D - 3

  4. D

    A - 3, B - 2, C - 4, D – 1

Show answer

Correct answer

  • C

    A - 4, B - 2, C - 1, D - 3

Question 36 on its own page

Question 37

+2 marksOne correct option

What is the significance of the ddNTPs in Sanger sequencing?

  1. A

    It allows for rapid sequencing of long DNA fragments.

  2. B

    It enables the synthesis of complementary DNA strands.

  3. C

    It enables rapid sequencing of DNA sample.

  4. D

    It provides sequence-specific terminaon of DNA synthesis.

Show answer

Correct answer

  • D

    It provides sequence-specific terminaon of DNA synthesis.

Question 37 on its own page

Question 38

+2 marksOne correct option

Which out of the following sequences is typically found at the end of a transcribed gene?

  1. A

    Start codon

  2. B

    Stop codon

  3. C

    Promoter sequence

  4. D

    Enhancer sequence

Show answer

Correct answer

  • B

    Stop codon

Question 38 on its own page

Question 39

+3 marksOne correct option

What is the sequence obtained from the gel picture given below.

  1. A

    5’-ATATTACGGTAG-3’

  2. B

    5’-ATAGTGTCGTAG-3’

  3. C

    5’-ATAGTCTCGTAG-3’

  4. D

    5’-ATATGCTAGTAG-3’

Show answer

Correct answer

  • B

    5’-ATAGTGTCGTAG-3’

Question 39 on its own page

Question 40

+1 markOne correct option

Answer the given subquestions from the given image below

Identify the plot

  1. A

    Circos Plot

  2. B

    Histogram

  3. C

    Manhatan Plot

  4. D

    Box plot

Show answer

Correct answer

  • C

    Manhatan Plot

Question 40 on its own page

Question 41

+1 markOne correct option

Answer the given subquestions from the given image below

What does the red horizontal line at y=6 represents?

  1. A

    Statistical cut off below which the values are significant.

  2. B

    Maximum threshold value of -Log10(P) value.

  3. C

    Statistical cut off above which the values are significant

  4. D

    None of these

Show answer

Correct answer

  • C

    Statistical cut off above which the values are significant

Question 41 on its own page

Question 42

+1 markOne correct option

Answer the given subquestions from the given image below

Based on the plot which of the chromosomes have significant SNP(s).

  1. A

    Chromosome 1

  2. B

    Chromosome 10

  3. C

    Chromosome 19

  4. D

    None of these

Show answer

Correct answer

  • D

    None of these

Question 42 on its own page