Assignment 2
Assignment 2
Programming Assignment 2:
Burrows–Wheeler Transform and Suffix Arrays
Revision: July 3, 2017
Introduction
Welcome to your second programming assignment of the Algorithms on Strings class! In this programming
assignment, you will be practicing implementing Burrows–Wheeler transform and suffix arrays.
Recall that starting from this programming assignment, the grader will show you only the first few tests
(see the questions 6.4 and 6.5 in the FAQ section).
Learning Outcomes
Upon completing this programming assignment you will be able to:
Contents
1 Problem: Construct the Burrows–Wheeler Transform of a String 3
1
5 General Instructions and Recommendations on Solving Algorithmic Problems 13
5.1 Reading the Problem Statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
5.2 Designing an Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
5.3 Implementing Your Algorithm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
5.4 Compiling Your Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
5.5 Testing Your Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.6 Submitting Your Program to the Grading System . . . . . . . . . . . . . . . . . . . . . . . . . 15
5.7 Debugging and Stress Testing Your Program . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2
1 Problem: Construct the Burrows–Wheeler Transform of a String
Problem Introduction
The Burrows–Wheeler transform of a string Text permutes the symbols of Text so that it becomes well
compressible. Moreover, the transformation is reversible: one can recover the initial string Text from its
Burrows–Wheeler transform. However, data compression is not its only application: it is also used for solving
the multiple pattern matching problem and the sequence alignment problem.
BWT(Text) is defined as follows. First, form all possible cyclic rotations of Text; a cyclic rotation is
defined by chopping off a suffix from the end of Text and appending this suffix to the beginning of Text.
Then, order all the cyclic rotations of Text lexicographically to form a |Text| × |Text| matrix of symbols
denoted by 𝑀 (Text). BWT(Text) is the last column of 𝑀 (Text)
Problem Description
Task. Construct the Burrows–Wheeler transform of a string.
Input Format. A string Text ending with a “$” symbol.
Constraints. 1 ≤ |Text| ≤ 1 000; except for the last symbol, Text contains symbols A, C, G, T only.
Output Format. BWT(Text).
Time Limits.
language C C++ Java Python C# Haskell JavaScript Ruby Scala
time (sec) 0.5 0.5 0.75 0.5 0.75 1 2.5 2.5 1.5
3
Sample 2.
Input:
ACACACAC$
Output:
CCCC$AAAA
⎡ ⎤
$ 𝐴 𝐶 𝐴 𝐶 𝐴 𝐶 𝐴 𝐶
⎢𝐴 𝐶 $ 𝐴 𝐶 𝐴 𝐶 𝐴 𝐶⎥
⎢ ⎥
⎢𝐴 𝐶 𝐴 𝐶 $ 𝐴 𝐶 𝐴 𝐶⎥
⎢ ⎥
⎢𝐴 𝐶 𝐴 𝐶 𝐴 𝐶 $ 𝐴 𝐶⎥
⎢ ⎥
⎢𝐴
𝑀 (Text) = ⎢ 𝐶 𝐴 𝐶 𝐴 𝐶 𝐴 𝐶 $⎥⎥
⎢𝐶 $ 𝐴 𝐶 𝐴 𝐶 𝐴 𝐶 𝐴⎥
⎢ ⎥
⎢𝐶 𝐴 𝐶 $ 𝐴 𝐶 𝐴 𝐶 𝐴⎥
⎢ ⎥
⎣𝐶 𝐴 𝐶 𝐴 𝐶 $ 𝐴 𝐶 𝐴⎦
𝐶 𝐴 𝐶 𝐴 𝐶 𝐴 𝐶 $ 𝐴
Sample 3.
Input:
AGACATA$
Output:
ATG$CAAA
⎡ ⎤
$ 𝐴 𝐺 𝐴 𝐶 𝐴 𝑇 𝐴
⎢𝐴 $ 𝐴 𝐺 𝐴 𝐶 𝐴 𝑇 ⎥
⎢ ⎥
⎢ 𝐴 𝐶 𝐴 𝑇 𝐴 $ 𝐴 𝐺⎥
⎢ ⎥
⎢𝐴 𝐺 𝐴 𝐶 𝐴 𝑇 𝐴 $ ⎥
𝑀 (Text) = ⎢
⎢ ⎥
⎢𝐴 𝑇 𝐴 $ 𝐴 𝐺 𝐴 𝐶 ⎥
⎥
⎢𝐶 𝐴 𝑇 𝐴 $ 𝐴 𝐺 𝐴⎥
⎢ ⎥
⎣𝐺 𝐴 𝐶 𝐴 𝑇 𝐴 $ 𝐴⎦
𝑇 𝐴 $ 𝐴 𝐺 𝐴 𝐶 𝐴
Starter Files
The starter solutions for this problem read the input data from the standard input, pass it to a blank
procedure, and then write the result to the standard output. You are supposed to implement your algorithm
in this blank procedure if you are using C++, Java, or Python3. For other programming languages, you need
to implement a solution from scratch. Filename: bwt
Need Help?
Ask a question or see the questions asked by other learners at this forum thread.
4
2 Problem: Reconstruct a String from its Burrows–Wheeler Trans-
form
Problem Introduction
In the previous problem, we introduced the Burrows–Wheeler transform of a string Text. It permutes the
symbols of Text making it well compressible. However, there were no sense in this, if this process would
not be reversible. It turns out that it is reversible, and your goal in this problem is to recover Text from
BWT(Text).
Problem Description
Task. Reconstruct a string from its Burrows–Wheeler transform.
Input Format. A string Transform with a single “$” sign.
Constraints. 1 ≤ |Transform| ≤ 1 000 000; except for the last symbol, Text contains symbols A, C, G, T
only.
Output Format. The string Text such that BWT(Text) = Transform. (There exists a unique such string.)
Time Limits.
language C C++ Java Python C# Haskell JavaScript Ruby Scala
time (sec) 2 2 3 10 3 4 10 10 6
5
Starter Files
The starter solutions for this problem read the input data from the standard input, pass it to a blank
procedure, and then write the result to the standard output. You are supposed to implement your algorithm
in this blank procedure if you are using C++, Java, or Python3. For other programming languages, you need
to implement a solution from scratch. Filename: bwtinverse
What To Do
To solve this problem, it is enough to implement carefully the corresponding algorithm covered in the lectures.
Need Help?
Ask a question or see the questions asked by other learners at this forum thread.
6
3 Problem: Matching Against a Compressed String
Problem Introduction
Not only the Burrows–Wheeler transform makes the input string Text well compressible, it also allows one
to solve the pattern matching problem using the compressed strings instead of the initial string! This is
another beautiful property of the Burrows–Wheeler transform which allows us to avoid decompressing the
string, and thus to save lots of memory, while still solving the problem at hand.
The algorithm BWMatching counts the total number of matches of Pattern in Text, where the
only information that we are given is FirstColumn and LastColumn = BWT(Text) in addition to the
Last-to-First mapping. The pointers top and bottom are updated by the green lines in the following
pseudocode.
BWMatching(FirstColumn, LastColumn, Pattern, LastToFirst):
top ← 0
bottom ← |LastColumn| − 1
while top ≤ bottom:
if Pattern is nonempty:
symbol ← last letter in Pattern
remove last letter from Pattern
if positions from top to bottom in LastColumn contain an occurrence of symbol:
topIndex ← first position of symbol among positions from top to bottom in LastColumn
bottomIndex ← last position of symbol among positions from top to bottom in LastColumn
top ← 𝐿𝑎𝑠𝑡𝑇 𝑜𝐹 𝑖𝑟𝑠𝑡(𝑡𝑜𝑝𝐼𝑛𝑑𝑒𝑥)
bottom ← 𝐿𝑎𝑠𝑡𝑇 𝑜𝐹 𝑖𝑟𝑠𝑡(𝑏𝑜𝑡𝑡𝑜𝑚𝐼𝑛𝑑𝑒𝑥)
else:
return 0
else:
return bottom − top + 1
The Last-to-First array, denoted LastToFirst(𝑖), answers the following question: given a symbol at position
𝑖 in LastColumn, what is its position in FirstColumn? For example, if Text = panamabananas$,
BWT(Text) = smnpbnnaaaaa$a, FirstCol(Text) = $aaaaaabmnnnps, then we can rewrite
BWT(Text) = s1 m1 n1 p1 b1 n2 n3 a1 a2 a3 a4 a5 $1 a6 and FirstCol(𝑇 𝑒𝑥𝑡) = $1 a1 a2 a3 a4 a5 a6 b1 m1 n1 n2 n3 p1 s1 , and
now we see that a3 in BWT(Text) corresponds to a3 in FirstCol(Text).
If you implement BWMatching, you probably will find the algorithm to be slow. The reason for its
sluggishness is that updating the pointers top and bottom is time-intensive, since it requires examining every
symbol in LastColumn between top and bottom at each step. To improve BWMatching, we introduce
a function Countsymbol (𝑖, LastColumn), which returns the number of occurrences of symbol in the first 𝑖
positions of LastColumn. For example,
The green lines from BWMatching can be compactly described without the First-to-Last mapping by the
following two lines:
top ← (Countsymbol + 1)-th occurrence of character symbol in FirstColumn
bottom ← position of symbol with rank Countsymbol (bottom + 1, LastColumn) in FirstColumn
7
top ← FirstOccurrence(symbol) + Countsymbol (top, LastColumn)
bottom ← FirstOccurrence(symbol) + Countsymbol (bottom + 1, LastColumn) − 1
In the process of simplifying the green lines of pseudocode from BWMatching, we have also eliminated
the need for both FirstColumn and LastToFirst, resulting in a more efficient algorithm called BetterBW-
Matching.
BWMatching(FirstOccurrence, LastColumn, Pattern, Count):
top ← 0
bottom ← |LastColumn| − 1
while top ≤ bottom:
if Pattern is nonempty:
symbol ← last letter in Pattern
remove last letter from Pattern
if positions from top to bottom in LastColumn contain an occurrence of symbol:
top ← FirstOccurrence(symbol) + Countsymbol (top, LastColumn)
bottom ← FirstOccurrence(symbol) + Countsymbol (bottom + 1, LastColumn) − 1
else:
return 0
else:
return bottom − top + 1
Problem Description
Task. Implement BetterBWMatching algorithm.
Input Format. A string BWT(Text), followed by an integer 𝑛 and a collection of 𝑛 strings Patterns =
{𝑝1 , . . . , 𝑝𝑛 } (on one line separated by spaces).
Constraints. 1 ≤ |BWT(Text)| ≤ 106 ; except for the one $ symbol, BWT(Text) contains symbols A, C,
G, T only; 1 ≤ 𝑛 ≤ 5 000; for all 1 ≤ 𝑖 ≤ 𝑛, 𝑝𝑖 is a string over A, C, G, T; 1 ≤ |𝑝𝑖 | ≤ 1 000.
Output Format. A list of integers, where the 𝑖-th integer corresponds to the number of substring matches
of the 𝑖-th member of Patterns in Text.
Time Limits.
language C C++ Java Python C# Haskell JavaScript Ruby Scala
time (sec) 4 4 6 24 6 8 24 24 12
In this case, Text = GAGAGA$. The pattern GA appears three times in it.
8
Sample 2.
Input:
ATT$AA
2
ATA A
Output:
23
Text = ATCGTTTA does not contain any occurrences of two given patterns.
Starter Files
The starter solutions for this problem read the input data from the standard input, pass the Burrows–
Wheeler Transform to a preprocessing procedure to precompute some useful values, then pass each pattern
along with BWT and precomputed values to the procedure which counts the number of occurrences of the
pattern in the text, and then write the result to the standard output. You are supposed to implement these
two procedure which are left blank if you are using C++, Java, or Python3. For other programming languages,
you need to implement a solution from scratch. Filename: bwmatching.
What To Do
To solve this problem, it is enough to carefully implement the algorithm described in the lectures. However,
don’t forget that you need to do the preprocessing of the 𝑇 𝑒𝑥𝑡 only once, and then use the results. If you do
the preprocessing of the 𝑇 𝑒𝑥𝑡 each time, there is no point in such preprocessing, you don’t save anything.
But if you do the preprocessing once, save the results, and use them for searching each pattern, you save a
lot on each search.
Need Help?
Ask a question or see the questions asked by other learners at this forum thread.
9
4 Problem: Construct the Suffix Array of a String
Problem Introduction
We saw that suffix trees can be too memory intensive to apply in practice. This becomes a serious issue for
the case of massive datasets like the ones arising in bioinformatics.
In 1993, Udi Manber and Gene Myers introduced suffix arrays as a memory-efficient alternative to suffix
trees. To construct SuffixArray(Text), we first sort all suffixes of Text lexicographically, assuming that “$”
comes first in the alphabet. The suffix array is the list of starting positions of these sorted suffixes. For
example,
SuffixArray(“panamabananas$”) = (13, 5, 3, 1, 7, 9, 11, 6, 4, 2, 8, 10, 0, 12)
E.g., the suffix tree of a human genome requires about 60 Gb, while the suffix array occupies around
12 Gb.
Problem Description
Task. Construct the suffix array of a string.
Time Limits.
language C C++ Java Python C# Haskell JavaScript Ruby Scala
time (sec) 1 1 2 1 1.5 2 5 5 4
Sorted suffixes:
3 $
1 AC$
2 C$
0 GAC$
10
Sample 2.
Input:
GAGAGAGA$
Output:
875316420
Sorted suffixes:
8 $
7 A$
5 AGA$
3 AGAGA$
1 AGAGAGA$
6 GA$
4 GAGA$
2 GAGAGA$
0 GAGAGAGA$
Sample 3.
Input:
AACGATAGCGGTAGA$
Output:
15 14 0 1 12 6 4 2 8 13 3 7 9 10 11 5
Sorted suffixes:
15 $
14 A$
0 AACGATAGCGGTAGA$
1 ACGATAGCGGTAGA$
12 AGA$
6 AGCGGTAGA$
4 ATAGCGGTAGA$
2 CGATAGCGGTAGA$
8 CGGTAGA$
13 GA$
3 GATAGCGGTAGA$
7 GCGGTAGA$
9 GGTAGA$
10 GTAGA$
11 TAGA$
5 TAGCGGTAGA$
Starter Files
The starter solutions for this problem read the input data from the standard input, pass it to a blank
procedure, and then write the result to the standard output. You are supposed to implement your algorithm
in this blank procedure if you are using C++, Java, or Python3. For other programming languages, you need
to implement a solution from scratch. Filename: suffix_array
What To Do
To solve this problem, it is enough to just sort all suffixes of Text.
11
Need Help?
Ask a question or see the questions asked by other learners at this forum thread.
12
5 General Instructions and Recommendations on Solving Algorith-
mic Problems
Your main goal in an algorithmic problem is to implement a program that solves a given computational
problem in just few seconds even on massive datasets. Your program should read a dataset from the standard
input and write an answer to the standard output.
Below we provide general instructions and recommendations on solving such problems. Before reading
them, go through readings and screencasts in the first module that show a step by step process of solving
two algorithmic problems: link.
13
same way on your machine and on the testing machine (note that a buggy program may behave differently
when compiled by different compilers, or even by the same compiler with different flags).
If your C/C++ compiler does not recognize -std=c++14 flag, try replacing it with -std=c++0x flag
or compiling without this flag at all (all starter solutions can be compiled without it). On Linux
and MacOS, you most probably have the required compiler. On Windows, you may use your favorite
compiler or install, e.g., cygwin.
∙ Python 2 (CPython 2.7). File extensions: .py2 or .py (a file ending in .py needs to have a first line
which is a comment containing “python2”). No flags:
python2
∙ Python 3 (CPython 3.4). File extensions: .py3 or .py (a file ending in .py needs to have a first line
which is a comment containing “python3”). No flags:
python3
14
5.5 Testing Your Program
When your program is ready, you start testing it. It makes sense to start with small datasets (for example,
sample tests provided in the problem description). Ensure that your program produces a correct result.
You then proceed to checking how long does it take your program to process a massive dataset. For
this, it makes sense to implement your algorithm as a function like solve(dataset) and then implement an
additional procedure generate() that produces a large dataset. For example, if an input to a problem is a
sequence of integers of length 1 ≤ 𝑛 ≤ 105 , then generate a sequence of length exactly 105 , pass it to your
solve() function, and ensure that the program outputs the result quickly.
Also, check the boundary values. Ensure that your program processes correctly sequences of size 𝑛 =
1, 2, 105 . If a sequence of integers from 0 to, say, 106 is given as an input, check how your program behaves
when it is given a sequence 0, 0, . . . , 0 or a sequence 106 , 106 , . . . , 106 . Check also on randomly generated
data. For each such test check that you program produces a correct result (or at least a reasonably looking
result).
In the end, we encourage you to stress test your program to make sure it passes in the system at the first
attempt. See the readings and screencasts from the first week to learn about testing and stress testing: link.
15
6 Frequently Asked Questions
6.1 I submit the program, but nothing happens. Why?
You need to create submission and upload the file with your solution in one of the programming languages C,
C++, Java, or Python (see Subsections 5.3 and 5.4). Make sure that after uploading the file with your solution
you press on the blue “Submit” button in the bottom. After that, the grading starts, and the submission
being graded is enclosed in an orange rectangle. After the testing is finished, the rectangle disappears, and
the results of the testing of all problems is shown to you.
6.2 I submit the solution only for one problem, but all the problems in the
assignment are graded. Why?
Each time you submit any solution, the last uploaded solution for each problem is tested. Don’t worry: this
doesn’t affect your score even if the submissions for the other problems are wrong. As soon as you pass the
sufficient number of problems in the assignment (see in the pdf with instructions), you pass the assignment.
After that, you can improve your result if you successfully pass more problems from the assignment. We
recommend working on one problem at a time, checking whether your solution for any given problem passes
in the system as soon as you are confident in it. However, it is better to test it first, please refer to the
reading about stress testing: link.
6.3 What are the possible grading outcomes, and how to read them?
Your solution may either pass or not. To pass, it must work without crashing and return the correct answers
on all the test cases we prepared for you, and do so under the time limit and memory limit constraints
specified in the problem statement. If your solution passes, you get the corresponding feedback "Good job!"
and get a point for the problem. If your solution fails, it can be because it crashes, returns wrong answer,
works for too long or uses too much memory for some test case. The feedback will contain the number of
the test case on which your solution fails and the total number of test cases in the system. The tests for the
problem are numbered from 1 to the total number of test cases for the problem, and the program is always
tested on all the tests in the order from the test number 1 to the test with the biggest number.
Here are the possible outcomes:
Good job! Hurrah! Your solution passed, and you get a point!
Wrong answer. Your solution has output incorrect answer for some test case. If it is a sample test case from
the problem statement, or if you are solving Programming Assignment 1, you will also see the input
data, the output of your program and the correct answer. Otherwise, you won’t know the input, the
output, and the correct answer. Check that you consider all the cases correctly, avoid integer overflow,
output the required white space, output the floating point numbers with the required precision, don’t
output anything in addition to what you are asked to output in the output specification of the problem
statement. See this reading on testing: link.
Time limit exceeded. Your solution worked longer than the allowed time limit for some test case. If it
is a sample test case from the problem statement, or if you are solving Programming Assignment 1,
you will also see the input data and the correct answer. Otherwise, you won’t know the input and the
correct answer. Check again that your algorithm has good enough running time estimate. Test your
program locally on the test of maximum size allowed by the problem statement and see how long it
works. Check that your program doesn’t wait for some input from the user which makes it to wait
forever. See this reading on testing: link.
Memory limit exceeded. Your solution used more than the allowed memory limit for some test case. If it
is a sample test case from the problem statement, or if you are solving Programming Assignment 1,
16
you will also see the input data and the correct answer. Otherwise, you won’t know the input and the
correct answer. Estimate the amount of memory that your program is going to use in the worst case
and check that it is less than the memory limit. Check that you don’t create too large arrays or data
structures. Check that you don’t create large arrays or lists or vectors consisting of empty arrays or
empty strings, since those in some cases still eat up memory. Test your program locally on the test of
maximum size allowed by the problem statement and look at its memory consumption in the system.
Cannot check answer. Perhaps output format is wrong. This happens when you output something
completely different than expected. For example, you are required to output word “Yes” or “No”, but
you output number 1 or 0, or vice versa. Or your program has empty output. Or your program outputs
not only the correct answer, but also some additional information (this is not allowed, so please follow
exactly the output format specified in the problem statement). Maybe your program doesn’t output
anything, because it crashes.
Unknown signal 6 (or 7, or 8, or 11, or some other). This happens when your program crashes. It
can be because of division by zero, accessing memory outside of the array bounds, using uninitialized
variables, too deep recursion that triggers stack overflow, sorting with contradictory comparator, re-
moving elements from an empty data structure, trying to allocate too much memory, and many other
reasons. Look at your code and think about all those possibilities. Make sure that you use the same
compilers and the same compiler options as we do. Try different testing techniques from this reading:
link.
Internal error: exception... Most probably, you submitted a compiled program instead of a source
code.
Grading failed. Something very wrong happened with the system. Contact Coursera for help or write in
the forums to let us know.
17
make one’s programs work, one must test them really well. Sometimes, the programs still don’t work although
you tried really hard to test them, and you need to be both skilled and creative to fix your bugs. Solutions
to algorithmic problems are one of the hardest to implement correctly. That’s why in this Specialization you
will gain this important experience which will be invaluable in the future when you write programs which
you really need to get right.
It is crucial for you to learn to test and fix your programs yourself. In the real life, often there will be no
or only partial information about the failure of your program or service. Still, you will have to reproduce the
failure to fix it (or just guess what it is, but that’s rare, and you will still need to reproduce the failure to
make sure you have really fixed it). When you solve algorithmic problems, it is very frequent to make subtle
mistakes. That’s why you should apply the testing techniques described in this reading to find the failing
test case and fix your program.
6.6 My solution does not pass the tests? May I post it in the forum and ask
for a help?
No, please do not post any solutions in the forum or anywhere on the web, even if a solution does not
pass the tests (as in this case you are still revealing parts of a correct solution). Recall the third item
of the Coursera Honor Code: “I will not make solutions to homework, quizzes, exams, projects, and other
assignments available to anyone else (except to the extent an assignment explicitly permits sharing solutions).
This includes both solutions written by me, as well as any solutions provided by the course staff or others”
(link).
6.7 My implementation always fails in the grader, though I already tested and
stress tested it a lot. Would not it be better if you give me a solution to
this problem or at least the test cases that you use? I will then be able to
fix my code and will learn how to avoid making mistakes. Otherwise, I do
not feel that I learn anything from solving this problem. I am just stuck.
First of all, you always learn from your mistakes.
The process of trying to invent new test cases that might fail your program and proving them wrong
is often enlightening. This thinking about the invariants which you expect your loops, ifs, etc. to keep and
proving them wrong (or right) makes you understand what happens inside your program and in the general
algorithm you’re studying much more.
Also, it is important to be able to find a bug in your implementation without knowing a test case and
without having a reference solution. Assume that you designed an application and an annoyed user reports
that it crashed. Most probably, the user will not tell you the exact sequence of operations that led to a crash.
Moreover, there will be no reference application. Hence, once again, it is important to be able to locate a
bug in your implementation yourself, without a magic oracle giving you either a test case that your program
fails or a reference solution. We encourage you to use programming assignments in this class as a way of
practicing this important skill.
If you have already tested a lot (considered all corner cases that you can imagine, constructed a set of
manual test cases, applied stress testing), but your program still fails and you are stuck, try to ask for help
on the forum. We encourage you to do this by first explaining what kind of corner cases you have already
considered (it may happen that when writing such a post you will realize that you missed some corner cases!)
and only then asking other learners to give you more ideas for tests cases.
18