Iteration

We often want a computer program to do a similar operation many times. For example, we may want to analyze many codons, one after another, and find the start and stop codons in order to determine the length of the open reading frame. Or, as a simpler example, we may wish to find the GC content of a specific sequence. We check each base to see if it is G or C and keep a count. Doing these repeated operations in a program is called iteration.

Introducing the for loop

The first method of iteration we will discuss is the for loop. As an example of its use, we compute the GC content of an RNA sequence.

[2]:

# The sequence we want to analyze
seq = 'GACAGACUCCAUGCACGUGGGUAUCUGUC'

# Initialize GC counter
n_gc = 0

# Initialize sequence length
len_seq = 0

# Loop through sequence and count G's and C's
for base in seq:
    len_seq += 1
    if base in 'GCgc':
        n_gc += 1

# Divide to get GC content
n_gc / len_seq

[2]:

0.5517241379310345

Let’s look carefully at what we did here. We took a string containing a sequence of nucleotides and then we did something for each character (base) in that string (nucleic acid sequence). A string is a sequence in the sense of the programming language as well; just like a list or tuple, the string is an ordered collection of characters. (So as not to confuse between biological sequences and sequences as a part of the Python language, we will always write the latter in italics.)

Now, let’s translate the new syntax in the above code to English.

Python: for base in seq:

English: for each character in the string whose variable name is seq, do the following, with that character taking the name base

This exposes a general way of doing things repeatedly in Python. For every item in a sequence, we do something. That something follows the for clause and is contained in an indentation block. When we do this, we say are “looping over a sequence.” In the context of a for clause, the membership operator, in, means that we consider in order each item in the sequence or iterator (we’ll talk about iterators in a moment).

Now, looking within the loop, the first thing we do is increment the length of the sequence. For each base we encounter in the loop, we add one to the sequence length. Makes sense!

Next, we have an if statement. We use the membership operator again. We ask if the current base is a G or a C. To be safe, we also included lower case characters in case the sequence was entered that way. If the base is a G or a C, we increment the counter of GC bases by one.

Finally, we get the fractional GC content by dividing the number of GC’s by the total length of the sequence.

Aside: the len() function

Note in the last example that we determined the length of the RNA sequence by iterating the len_seq counter within the for loop. This works, but Python has a built-in function (like print() is a built-in function) to compute the length of a string (or list, tuple, or other sequence). To find out the length of a sequence, simply use it as an argument to the len() function.

[3]:

len(seq)

[3]:

Example and watchout: modifying a list

Let’s look at another example of iterating through a list. Say we have a list of integers, and we want to change it by doubling each one. Let’s just do it.

[4]:

# We'll do one through 5
my_integers = [1, 2, 3, 4, 5]

# Double each one
for n in my_integers:
    n *= 2

# Check out the result
my_integers

[4]:

[1, 2, 3, 4, 5]

Whoa! It didn’t seem to double any of the integers! This is because my_integers was converted to an iterator in the for clause, and the iterator returns a copy of the item, not a reference to it. Therefore, the n inside the for block is not a view into the original list, and doubling it does nothing meaningful.

We’ve seen how to change individual list elements with indexing:

[5]:

# Don't do things this way
my_integers[0] *= 2
my_integers[1] *= 2
my_integers[2] *= 2
my_integers[3] *= 2
my_integers[4] *= 2

my_integers

[5]:

[2, 4, 6, 8, 10]

But we’d obviously like a better way to do this, with less typing and without knowing ahead of time the length of the list. Let’s look at a new concept that’ll help with this example.

Iterators

In the previous example, we iterated over a sequence. A sequence is one of many iterable objects, called iterables. Under the hood, the Python interpreter actually converts an iterable to an iterator. An iterator is a special object that gives values in succession. A major difference between a sequence and an iterator is that you cannot index an iterator. This seems like a trivial difference, but iterators make for more efficient computing than directly using a sequence with indexing.

We can explicitly convert a sequence to an iterator using the built-in function iter(), but we will not bother with that here because the Python interpreter automatically does this for you when you use a sequence in a loop. (This is incidentally why the previous example didn’t work—when the list is converted to an iterator, a copy is made of each element so the original list is unchanged.)

Instead, we will now explore how we can create useful iterators using the range(), enumerate(), and zip() functions. I know we have not yet covered functions, but the syntax should not be so complicated that you cannot understand what these functions are doing, just like with the print() and len() functions.

The range() function

The range() function gives an iterable that enables counting. Let’s look at an example.

[6]:

for i in range(10):
    print(i, end='  ')

0  1  2  3  4  5  6  7  8  9

We see that range(10) gives us ten numbers, from 0 to 9. As with indexing, range() inclusively starts at zero by default, and the ending is exclusive.

It turns out that the arguments of the range() function work much like indexing. If you have a single argument, you get that many integers, starting at 0 and incrementing by one. If you give two arguments, you start inclusively at the first and increment by one ending exclusively at the second argument. Finally, you can specify a stride with the third argument.

[7]:

# Print numbers 2 through 9
for i in range(2, 10):
    print(i, end='  ')

# Print a newline
print()

# Print even numbers, 2 through 9
for i in range(2, 10, 2):
    print(i, end='     ')

2  3  4  5  6  7  8  9
2     4     6     8

It is often useful to make a list or tuple that has the same entries that a corresponding range object would have. We can do this with type conversion.

[8]:

list(range(10))

[8]:

[0, 1, 2, 3, 4, 5, 6, 7, 8, 9]

We can use the range() function along with the len() function on lists to double the elements of our list from a bit ago:

[9]:

my_integers = [1, 2, 3, 4, 5]

# Since len(my_integers) = 5, this takes i from 0 to 4, exactly the indices of my_integers
for i in range(len(my_integers)):
    my_integers[i] *= 2

my_integers

[9]:

[2, 4, 6, 8, 10]

This works, but there’s an even better way to do this, with the next function.

The enumerate() function

Let’s say we want to print the indices of all G bases in a DNA sequence. We could do this by modifying our previous program.

[10]:

# Initialize GC counter
n_gc = 0

# Initialized sequence length
len_seq = 0

# Loop through sequence and print index of G's
for base in seq:
    if base in 'Gg':
        print(len_seq, end='  ')
    len_seq += 1

0  4  12  16  18  19  20  26

This is not so bad, but there is an easier way to do this. The enumerate() function gives an iterator that provides both the index and the item of a sequence. Again, this is best demonstrated in practice.

[11]:

# Loop through sequence and print index of G's
for i, base in enumerate(seq):
    if base in 'Gg':
        print(i, end='  ')

0  4  12  16  18  19  20  26

The enumerate() function allowed us to use an index and a base at the same time. To make it more clear, we can print the index and base type for each base in the sequence.

[12]:

# Print index and identity of bases
for i, base in enumerate(seq):
    print(i, base)

0 G
1 A
2 C
3 A
4 G
5 A
6 C
7 U
8 C
9 C
10 A
11 U
12 G
13 C
14 A
15 C
16 G
17 U
18 G
19 G
20 G
21 U
22 A
23 U
24 C
25 U
26 G
27 U
28 C

The enumerate() function is really useful and should be used in favor of just doing indexing. For example, many programmers, especially those first trained in lower-level languages, would write the above code similar to how we did the list doubling, with the range() and len() functions, but this is not good practice in Python.

Using enumerate(), the list doubling code looks like this:

[13]:

my_integers = [1, 2, 3, 4, 5]

# Double each one
for i, _ in enumerate(my_integers):
    my_integers[i] *= 2

# Check out the result
my_integers

[13]:

[2, 4, 6, 8, 10]

enumerate() is more generic and the overhead for returning a reference to an object isn’t an issue. The range(len()) construct will break on an object without support for len(). In addition, you are more likely to introduce bugs by imposing indexing on objects that are iterable but not unambiguously indexable. It is better to use the enumerate() function.

Note that we used the underscore, _, as a throwaway variable that we do not use. There is no rule for this, but this is generally accepted Python syntax and helps signal that you are not going to use the variable.

One last gotcha: if we tried to do a similar technique with a string, we get a TypeError because a string is immutable.

[14]:

# Try to convert capital G to lower g
for i, base in enumerate(seq):
    if base == 'G':
        seq[i] = 'g'

---------------------------------------------------------------------------
TypeError                                 Traceback (most recent call last)
<ipython-input-14-caac05059a6e> in <module>
      2 for i, base in enumerate(seq):
      3     if base == 'G':
----> 4         seq[i] = 'g'

TypeError: 'str' object does not support item assignment

The zip() function

The zip() function enables us to iterate over several iterables at once. In the example below we iterate over the names, nationality, and category of 2018 Nobel Prizes.

[15]:

names = (
    'Frances Arnold', 'George Smith', 'Gregory Winter',
    'postponed',
    'Denis Mukwege', 'Nadia Murad',
    'Arthur Ashkin', 'Gérard Mourou', 'Donna Strickland',
    'James Allison', 'Tasuku Honjo',
    'William Nordhaus', 'Paul Romer'
)

nationalities = (
    'USA', 'USA', 'UK',
    '---',
    'DRC', 'Iraq',
    'USA', 'France', 'Canada',
    'USA', 'Japan',
    'USA', 'USA'
)

categories = (
    'Chemistry', 'Chemistry', 'Chemistry',
    'Literature',
    'Peace', 'Peace',
    'Physics', 'Physics', 'Physics',
    'Physiology or Medicine', 'Physiology or Medicine',
    'Economics', 'Economics'
)

for name, nat, cat in zip(names, nationalities, categories):
    print(cat, end=': ')
    print(name, end=', ')
    print(nat)

Chemistry: Frances Arnold, USA
Chemistry: George Smith, USA
Chemistry: Gregory Winter, UK
Literature: postponed, ---
Peace: Denis Mukwege, DRC
Peace: Nadia Murad, Iraq
Physics: Arthur Ashkin, USA
Physics: Gérard Mourou, France
Physics: Donna Strickland, Canada
Physiology or Medicine: James Allison, USA
Physiology or Medicine: Tasuku Honjo, Japan
Economics: William Nordhaus, USA
Economics: Paul Romer, USA

For fun, check out Paul Romer’s feelings about Python and Jupyter.

The reversed() function

This function is useful for giving an iterator that goes in the reverse direction. We’ll see that this can be convenient in the next lesson. For now, let’s pretend we’re NASA and need to count down.

[16]:

count_up = ('ignition', 1, 2, 3, 4, 5, 6, 7, 8 ,9, 10)

for count in reversed(count_up):
    print(count)

10
9
8
7
6
5
4
3
2
1
ignition

The while loop

The for loop is very powerful and allows us to construct iterative calculations. When we use a for loop, we need to set up an iterator. A while loop, on the other hand, allows iteration until a conditional expression evaluates False.

As an example of a while loop, we will calculate the length of a sequence before hitting a start codon.

[17]:

# Define start codon
start_codon = 'AUG'

# Initialize sequence index
i = 0

# Scan sequence until we hit the start codon
while seq[i:i+3] != start_codon:
    i += 1

# Show the result
print('The start codon starts at index', i)

The start codon starts at index 10

Let’s walk through the while loop. The value of i is changing with each iteration, being incremented by one. Each time we consider doing another iteration, the conditional is checked: do the next three bases match the start codon? We set up the conditional to evaluate to True when the bases are not the start codon, so the iteration continues. In other words, iteration continues in a while loop until the conditional returns False.

Notice that we sliced the string the same way we sliced lists and tuples. In the case of strings, a slice gives another string, i.e., a sequential collection of characters.

Let’s try looking for another codon…. But, let’s actually not do that. If you run the code below, it will run forever and nothing will get printed to the screen.

# Define codon of interest
codon = 'GCC'

# Initialize sequence index
i = 0

# Scan sequence until we hit the start codon, but DON'T DO THIS!!!!!
while seq[i:i+3] != codon:
    i += 1

# Show the result
print('The codon starts at index', i)

The reason this runs forever is that the conditional expression in the while statement never returns False. If we slice a string beyond the length of the string we get an empty string result.

[18]:

seq[100:103]

[18]:

''

This does not equal the codon we’re interested in, so the while loop keeps going forever. This is called an infinite loop, and you definitely to not want these in your code! We can fix it by making a conditional that will evaluate to False if we reach the end of the string.

[19]:

# Define codon of interest
codon = 'GCC'

# Initialize sequence index
i = 0

# Scan sequence until we hit the start codon or the end of the sequence
while seq[i:i+3] != codon and i < len(seq):
    i += 1

# Show the result
if i == len(seq):
    print('Codon not found in sequence.')
else:
    print('The codon starts at index', i)

Codon not found in sequence.

for vs while

Most anything that requires a loop can be done with either a for loop or a while loop, but there’s a general rule of thumb for which type of loop to use. If you know how many times you have to do something (or if your program knows), use a for loop. If you don’t know how many times the loop needs to run until you run it, use a while loop. For example, when we want to do something with each character in a string or each entry in a list, the program knows how long the sequence is and a for loop is more appropriate. In the previous examples, we don’t know how long it will be before we hit the start codon; it depends on the sequence you put into the program. That makes it more suited to a while loop.

The break and else keywords

Iteration stops in a for loop when the iterator is exhausted. It stops in a while loop when the conditional evaluates to False. These is another way to stop iteration: the break keyword. Whenever break is encountered in a for or while loop, the iteration halts and execution continues outside the loop. As an example, we’ll do the calculation above with a for loop with a break instead of a while loop.

[20]:

# Define start codon
start_codon = 'AUG'

# Scan sequence until we hit the start codon
for i in range(len(seq)):
    if seq[i:i+3] == start_codon:
        print('The start codon starts at index', i)
        break
else:
    print('Codon not found in sequence.')

The start codon starts at index 10

Notice that we have an else block after the for loop. In Python, for and while loops can have an else statement after the code block to be evaluated in the loop. The contents of the else block are evaluated if the loop completes without encountering a break.

A note about use of the word “codon”

The astute biologists among you will note that we have not really been using the word “codon” properly here. We are taking it to mean any three consecutive bases, but the more precise definition of a codon means that it is three consecutive bases that code for an amino acid. That means that for a three-base sequence to be a codon, it must be in-register with the start codon. We will see how this distinction could lead to problems when we discuss test-driven development.

List comprehensions

Let’s say we wanted to build a list containing the information about the 2018 Nobel laureates. We could generate such a list as follows, using the += increment operator of lists.

[21]:

laureates_2018 = []
for name, nat, cat in zip(names, nationalities, categories):
    laureates_2018 += [(cat, name, nat)]

laureates_2018

[21]:

[('Chemistry', 'Frances Arnold', 'USA'),
 ('Chemistry', 'George Smith', 'USA'),
 ('Chemistry', 'Gregory Winter', 'UK'),
 ('Literature', 'postponed', '---'),
 ('Peace', 'Denis Mukwege', 'DRC'),
 ('Peace', 'Nadia Murad', 'Iraq'),
 ('Physics', 'Arthur Ashkin', 'USA'),
 ('Physics', 'Gérard Mourou', 'France'),
 ('Physics', 'Donna Strickland', 'Canada'),
 ('Physiology or Medicine', 'James Allison', 'USA'),
 ('Physiology or Medicine', 'Tasuku Honjo', 'Japan'),
 ('Economics', 'William Nordhaus', 'USA'),
 ('Economics', 'Paul Romer', 'USA')]

While this method of creating the list works, it is a bit awkward. For each entry we add, we create a new one-element list (containing, in this case, a 3-tuple) to concatenate to the list of scorers (though we can get around this using the append() method of lists, which we will discuss in a later lesson). We also have to create an empty list to start with. Under the hood, this means that the Python interpreter has to keep allocating memory as it creates and grows lists. So, in addition to being syntactically clunky, the above way of creating a list is inefficient.

Python offers an alternative, list comprehensions. This is, as usual, best seen through example.

[22]:

[(cat, name, nat) for name, nat, cat in zip(names, nationalities, categories)]

[22]:

[('Chemistry', 'Frances Arnold', 'USA'),
 ('Chemistry', 'George Smith', 'USA'),
 ('Chemistry', 'Gregory Winter', 'UK'),
 ('Literature', 'postponed', '---'),
 ('Peace', 'Denis Mukwege', 'DRC'),
 ('Peace', 'Nadia Murad', 'Iraq'),
 ('Physics', 'Arthur Ashkin', 'USA'),
 ('Physics', 'Gérard Mourou', 'France'),
 ('Physics', 'Donna Strickland', 'Canada'),
 ('Physiology or Medicine', 'James Allison', 'USA'),
 ('Physiology or Medicine', 'Tasuku Honjo', 'Japan'),
 ('Economics', 'William Nordhaus', 'USA'),
 ('Economics', 'Paul Romer', 'USA')]

The list comprehension is enclosed in brackets. The first part, (cat, name, nat), is an expression that will be inserted into the list. Next comes a for statement to produce the iterator.

Now, let’s say we are really interested in the prize in chemistry. We can add an if statement to the comprehension.

[23]:

[(cat, name, nat) for name, nat, cat in zip(names, nationalities, categories) if cat == 'Chemistry']

[23]:

[('Chemistry', 'Frances Arnold', 'USA'),
 ('Chemistry', 'George Smith', 'USA'),
 ('Chemistry', 'Gregory Winter', 'UK')]

We can also nest iterators. For example, let’s say the the chemistry and medicine prize winners got together in Sweden and wanted to play against each other in basketball. There are three chemistry winners, but only two medicine winners, so they will have to play two-on-two. Now, say we want to make a list of all possible pairs of chemistry winners so that we can get all possible Chemistry teams for the basketball game.

[24]:

# First get list of chemistry laureates
chem_names = [name for name, cat in zip(names, categories) if cat == 'Chemistry']

# List of all possible pairs of chemistry laureates
[(n1, n2) for i, n1 in enumerate(chem_names) for j, n2 in enumerate(chem_names) if i < j]

[24]:

[('Frances Arnold', 'George Smith'),
 ('Frances Arnold', 'Gregory Winter'),
 ('George Smith', 'Gregory Winter')]

To summarize this structure of list comprehensions, borrowing from Dave Beazley’s explanation in Python Essential Reference, a list comprehension has the following structure.

[expression_to_put_in_list for i_1 in iterable_1 if condition_1
                           for i_2 in iterable_2 if condition_2
                                     ...
                           for i_n in iterable_n if condition_n]

which is roughly equivalent to

my_list = []
for i_1 in iterable_1:
    if condition_1:
        for i_2 in iterable_2:
            if condition_2:
                ...
                for i_n in iterable_n:
                    if condition_n:
                        my_list += [expression_to_put_in_list]

What if you want an else statement in a list comprehension?

Now, let’s say that we deem “Physiology or Medicine” to be too long of a title for the category of the prize. We instead want to substitute that phrase with “Medicine” for brevity. We might construct the list like this:

[25]:

[
    ("Medicine", name, nat)
    for name, nat, cat in zip(names, nationalities, categories)
    if cat == "Physiology or Medicine"
]

[25]:

[('Medicine', 'James Allison', 'USA'), ('Medicine', 'Tasuku Honjo', 'Japan')]

This leaves out all of the other prizes. So, we need an else statement. (Note here that we split the list comprehension over many lines for readability, which is perfectly legal.) To include all prizes, we might try it like this.

[26]:

[
    ("Medicine", name, nat)
    for name, nat, cat in zip(names, nationalities, categories)
    if cat == "Physiology or Medicine" else (cat, name, nat)
]

  File "<ipython-input-26-0bfb7d47a7a6>", line 4
    if cat == "Physiology or Medicine" else (cat, name, nat)
                                       ^
SyntaxError: invalid syntax

Syntax error! This structure of a list comprehension does not match the template shown above. In the conditional expression of list comprehensions, you cannot have an else block.

However, the expression_to_put_in_list can be any valid Python expression. The following is a valid Python expression:

("Medicine", name, nat) if cat == "Physiology or Medicine" else (cat, name, nat)

So, we can still use a list comprehension to build the list.

[27]:

[
    ("Medicine", name, nat) if cat == "Physiology or Medicine" else (cat, name, nat)
    for name, nat, cat in zip(names, nationalities, categories)
]

[27]:

[('Chemistry', 'Frances Arnold', 'USA'),
 ('Chemistry', 'George Smith', 'USA'),
 ('Chemistry', 'Gregory Winter', 'UK'),
 ('Literature', 'postponed', '---'),
 ('Peace', 'Denis Mukwege', 'DRC'),
 ('Peace', 'Nadia Murad', 'Iraq'),
 ('Physics', 'Arthur Ashkin', 'USA'),
 ('Physics', 'Gérard Mourou', 'France'),
 ('Physics', 'Donna Strickland', 'Canada'),
 ('Medicine', 'James Allison', 'USA'),
 ('Medicine', 'Tasuku Honjo', 'Japan'),
 ('Economics', 'William Nordhaus', 'USA'),
 ('Economics', 'Paul Romer', 'USA')]

To be clear here, there is no conditional in the list comprehension; the conditional is in the expression to be added to the list, which we have called expression_to_put_in_list.

List comprehensions will prove very useful, and you will see many of them throughout the course.

Computing environment

[28]:

%load_ext watermark
%watermark -v -p jupyterlab

Python implementation: CPython
Python version       : 3.8.11
IPython version      : 7.26.0

jupyterlab: 3.1.7