Jack Downson

Jack Downson


How to Program UMAP from Scratch

This is the thirteenth article of my column Mathematical Statistics and Machine Learning for Life Sciences where I try to explain some mysterious analytical techniques used in Bioinformatics, Biomedicine, Genetics etc. in a simple way. In the previous post How Exactly UMAP works I started with an intuitive explanation of the math behind UMAP. The best way to learn it is to program UMAP from scratch, this is what we are going to do today. The idea of this post is to show that it is relatively easy for everyone to create their own neighbor graph dimension reduction technique that can provide even better visualization than UMAP. It is going to be lots of coding, buckle up!

Building High-Dimensional Probabilities

As a test data set, we will be using the Cancer Associated Fibroblasts (CAFs) scRNAseq data. We start with importing Python libraries (mainly numpy and scikit-learn will be used), having a look at the data matrix and checking the dimensions of the data set. Please note, that the rows are cells and columns are genes here, the last column contains the coding of the clustering results, i.e. each cell belongs to one of 4 clusters with IDs #1, 2, 3 and 4:


import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
from sklearn.metrics.pairwise import euclidean_distances

expr = pd.read_csv('CAFs.txt', sep='\t')
X_train = expr.values[:,0:(expr.shape[1]-1)]
X_train = np.log(X_train + 1)
n = X_train.shape[0]
print("\nThis data set contains " + str(n) + " samples")
y_train = expr.values[:,expr.shape[1]-1]
print("\nDimensions of the  data set: ")
print(X_train.shape, y_train.shape)

One important global variable to define from the very beginning is the matrix of squared pairwise Euclidean distances for the initial high-dimensional scRNAseq data set, this matrix will be used a lot in the future code. Here, we also define the local connectivity parameter rho, as a distance to the first nearest neighbor, see the previous post for more detailed explanations of the meaning of the parameter.


dist = np.square(euclidean_distances(X_train, X_train))
rho = [sorted(dist[i])[1] for i in range(dist.shape[0])]
print(dist[0:4, 0:4])

Using the matrix of squared pairwise Euclidean distances we will compute the matrix of probabilities in the high-dimensional space. Knowing the matrix, we can easily calculate the Entropy as a sum of probabilities for each cell, and the number of nearest neighbors, k, as k = 2^Entropy. Please note the argument σ of the function. This implies that the matrix of probabilities as a function of σ will be used later for the binary search procedure that computes the optimal σ for the fixed number of nearest neighbors.


def prob_high_dim(sigma, dist_row):
    For each row of Euclidean distance matrix (dist_row) compute
    probability in high dimensions (1D array)
    d = dist[dist_row] - rho[dist_row]; d[d < 0] = 0
    return np.exp(- d / sigma)

def k(prob):
    Compute n_neighbor = k (scalar) for each 1D array of high-dimensional probability
    return np.power(2, np.sum(prob))

A trick here is that for convenience we compute a 1D array of probabilities for each i-th cell, this means for each i-th row (or column) of the squared pairwise Euclidean distance matrix (distance from i-th cell to all other cells in the data set). This is done because of the unknown σ_i in the equation:

This is image title

We have to use binary search in order to find σ_i for each i-th cell in the data set. For each i-th cell, given the 1D array of the high-dimensional probabilities, we can sum up the elements of the array and compute the number of nearest neighbors, k, according to the definition:

This is image title

So now we have a function that produces the number of nearest neighbors, k, (scalar) value for each σ i_ for each i-th cell. We can fix the number of nearest neighbors value (desired k hyper-parameter) and input this function into the binary search procedure in order to compute the σ_i for each i-th cell.


def sigma_binary_search(k_of_sigma, fixed_k):
    Solve equation k_of_sigma(sigma) = fixed_k 
    with respect to sigma by the binary search algorithm
    sigma_lower_limit = 0; sigma_upper_limit = 1000
    for i in range(20):
        approx_sigma = (sigma_lower_limit + sigma_upper_limit) / 2
        if k_of_sigma(approx_sigma) < fixed_k:
            sigma_lower_limit = approx_sigma
            sigma_upper_limit = approx_sigma
        if np.abs(fixed_k - k_of_sigma(approx_sigma)) <= 1e-5:
    return approx_sigma

prob = np.zeros((n,n)); sigma_array = []
for dist_row in range(n):
    func = lambda sigma: k(prob_high_dim(sigma, dist_row))
    binary_search_result = sigma_binary_search(func, N_NEIGHBOR)
    prob[dist_row] = prob_high_dim(binary_search_result, dist_row)
    if (dist_row + 1) % 100 == 0:
        print("Sigma binary search finished {0} of {1} cells".format(dist_row + 1, n))
print("\nMean sigma = " + str(np.mean(sigma_array)))

Computing σ_i for each cell we get n-cells 1D arrays of probabilities, they all together build a matrix of high-dimensional probabilities. This matrix has to satisfy the symmetry condition according to the UMAP algorithm:

This is image title

However, to my experience the following simpler symmetry condition provides a better visualization as we will see later.


#P = prob + np.transpose(prob) - np.multiply(prob, np.transpose(prob))
P = (prob + np.transpose(prob)) / 2

I encourage the readers to play with different ways of obtaining the symmetric matrix of high-dimensional probabilities and check how the final visualization changes.

Building Low-Dimensional Probabilities

Next, we are going to compute the layout for the high-dimensional neighbor graph, i.e. the matrix of low-dimensional probabilities according to the UMAP definition:

This is image title

Since the goal of the UMAP algorithm is to find the optimal coordinates of the low-dimensional embeddings, y_i, which are updated through the Gradient Descent procedure, I would expect that the a and b parameters should also be updated at each Gradient Descent iteration. However, in the original UMAP algorithm, they seem to be fixed at the very beginning for a desired value of min_dist hyper-parameter:


MIN_DIST = 0.25

x = np.linspace(0, 3, 300)

def f(x, min_dist):
    y = []
    for i in range(len(x)):
        if(x[i] <= min_dist):
            y.append(np.exp(- x[i] + min_dist))
    return y

dist_low_dim = lambda x, a, b: 1 / (1 + a*x**(2*b))

p , _ = optimize.curve_fit(dist_low_dim, x, f(x, MIN_DIST))

a = p[0]
b = p[1] 
print("Hyperparameters a = " + str(a) + " and b = " + str(b))

Again, here I see a room for improvements and experiments. For simplicity, below I will set a=1 and b=1 in order to get a sort of hybrid of UMAP and tSNE. Then the matrix of low-dimensional probabilities is given by:


def prob_low_dim(Y):
    Compute matrix of probabilities q_ij in low-dimensional space
    inv_distances = np.power(1 + a * np.square(euclidean_distances(Y, Y))**b, -1)
    return inv_distances

Learning Low-Dimensional Embeddings

Finally, we are going to code the cost function of UMAP which is the Cross-Entropy, please see the previous post for details of how this cost function is capable of preserving global distances. We are actually interested not in the Cross-Entropy itself but in its gradient to be used in the Gradient Descent procedure later.


def CE(P, Y):
    Compute Cross-Entropy (CE) from matrix of high-dimensional probabilities 
    and coordinates of low-dimensional embeddings
    Q = prob_low_dim(Y)
    return - P * np.log(Q + 0.01) - (1 - P) * np.log(1 - Q + 0.01)

def CE_gradient(P, Y):
    Compute the gradient of Cross-Entropy (CE)
    y_diff = np.expand_dims(Y, 1) - np.expand_dims(Y, 0)
    inv_dist = np.power(1 + a * np.square(euclidean_distances(Y, Y))**b, -1)
    Q = np.dot(1 - P, np.power(0.001 + np.square(euclidean_distances(Y, Y)), -1))
    np.fill_diagonal(Q, 0)
    Q = Q / np.sum(Q, axis = 1, keepdims = True)
    fact=np.expand_dims(a*P*(1e-8 + np.square(euclidean_distances(Y, Y)))**(b-1) - Q, 2)
    return 2 * b * np.sum(fact * y_diff * np.expand_dims(inv_dist, 2), axis = 1)

Please note the Q matrix in the CE_gradient function. It is not supposed to be normalized according to the UMAP algorithm. However, to my experience this normalization improves the low-dimensional visualization. I encourage again everyone to experiment here. Finally, we initialize the coordinates of low-dimensional embeddings y_i with Graph Laplacian and launch the Gradient Descent procedure:


MAX_ITER = 200

model = SpectralEmbedding(n_components = N_LOW_DIMS, n_neighbors = 50)
y = model.fit_transform(np.log(X_train + 1))
#y = np.random.normal(loc = 0, scale = 1, size = (n, N_LOW_DIMS))

CE_array = []
print("Running Gradient Descent: \n")
for i in range(MAX_ITER):
    y = y - LEARNING_RATE * CE_gradient(P, y)
    plt.scatter(y[:,0], y[:,1], c = y_train.astype(int), cmap = 'tab10', s = 50)
    plt.title("UMAP on Cancer Associated Fibroblasts (CAFs): Programmed from Scratch", 
              fontsize = 20)
    plt.xlabel("UMAP1", fontsize = 20); plt.ylabel("UMAP2", fontsize = 20)
    plt.savefig('UMAP_Plots/UMAP_iter_' + str(i) + '.png')
    CE_current = np.sum(CE(P, y)) / 1e+5
    if i % 10 == 0:
        print("Cross-Entropy = " + str(CE_current) + " after " + str(i) + " iterations")

plt.title("Cross-Entropy", fontsize = 20)
plt.xlabel("ITERATION", fontsize = 20); plt.ylabel("CROSS-ENTROPY", fontsize = 20)

This is image title

The Cross Entropy seems to be decreasing and reaching a plateau. Please note that here for simplicity I implement the regular Gradient Descent rather than Stochastic Gradient Descent that is used by the original implementation of UMAP. This is because the goal of the post is to demonstrate the idea of UMAP and not to provide an optimized and pretty code. Now, we can visualize the low-dimensional embeddings:


plt.scatter(y[:,0], y[:,1], c = y_train.astype(int), cmap = 'tab10', s = 50)
plt.title("UMAP on Cancer Associated Fibroblasts (CAFs): Programmed from Scratch", 
          fontsize = 20)
plt.xlabel("UMAP1", fontsize = 20); plt.ylabel("UMAP2", fontsize = 20)

This is image title

We can also easily animate how the distinct clusters are formed starting from the Laplacian embeddings, this is done by merging plots from each iteration into a gif-file.


convert -delay 0 $(for i in $(seq 0 1 20; seq 21 10 199); do echo UMAP_iter_${i}.png; done) \
-loop 0 UMAP_animated.gif

This is image title

The from scratch implementation of UMAP seems to be working fine, the clusters of the 4 cell populations are clearly distinguishable. Clustering algorithms would not have problems identifying the cell populations if we were to run them on the low-dimensional embeddings above.

Let us compare the figure above with the original UMAP Python + numba implementation. We are going to use n_neighbors = 15 and min_dist = 0.25, i.e. identical values of UMAP hyperparameters as in the previous from scratch UMAP implementation.


from umap import UMAP
model = UMAP(n_neighbors = 15, min_dist = 0.25, n_components = 2, verbose = True)
umap = model.fit_transform(X_train)
plt.scatter(umap[:, 0], umap[:, 1], c = y_train.astype(int), cmap = 'tab10', s = 50)
plt.title('UMAP', fontsize = 20)
plt.xlabel("UMAP1", fontsize = 20)
plt.ylabel("UMAP2", fontsize = 20)

This is image title

What we can see here is that in a sense the from scratch implementation gives more distinct clusters compared to the original UMAP implementation. This is due to small changes applied on the way, such as slightly different symmetry condition and normalization of the Q matrix in the gradient of Cross Entropy.


In this post, we have learnt that it is relatively easy to implement UMAP from scratch in Python. Therefore, my prediction is that UMAP is just the beginning, and there going to be many more, and possibly better, dimension reduction techniques appearing in the Single Cell Genomics research area in the nearest future. It is quite straightforward to tune the low- and high-dimensional distributions, make better normalization and symmetry conditions, better cost functions, and play with attractive / repulsive forces for the N-body problem in order to get even better low-dimensional representations.

In the comments below let me know which analytical techniques in Life Sciences seem especially mysterious to you and I will try to cover them in the future posts. Check the codes from the post on my github. Follow me at Medium Nikolay Oskolkov, in Twitter @NikolayOskolkov and connect in Linkedin. Next time I will turn to Evolutionary Biology and show how to estimate population size back in time using Ancient DNA, stay tuned.

#DeepLearning # machine learning #TensorFlow

What is GEEK

Buddha Community

How to Program UMAP from Scratch

Coding 101: Programming Language Building Blocks

This article will introduce the concepts and topics common to all programming languages, that beginners and experts must know!

Do you want to learn a programming language for the first time?

Do you want to improve as a Programmer?

Well, then you’re in the right place to start. Learn any programming language without difficulty by learning the concepts and topics common to all programming languages.

Let me start by answering the following questions:

  • Why learn Programming?
  • What is Programming?
  • How to Learn a Programming Language?

Why learn Programming❔

Programming develops creative thinking

Programmers solve a problem by breaking it down into workable pieces to understand it better. When you start learning to program, you develop the habit of working your way out in a very structured format. You analyze the problem and start thinking logically and this gives rise to more creative solutions you’ve ever given.

Whether you want to uncover the secrets of the universe, or you just want to pursue a career in the 21st century, basic computer programming is an essential skill to learn.

_– _Stephen Hawking

Everybody in this country should learn how to program a computer… because it teaches you how to think.

_- _Steve Jobs

Programming Provides Life-Changing Experiences

Programming always provides you with a new challenge to take risks every time and that teaches you to take risks in your personal life too. The world is filled up with websites, apps, software and when you build these yourself you’ll feel more confident. When a programmer solves a problem that no one has ever solved before it becomes a life-changing experience for them.

What is Programming🤔?

program is a set of instructions to perform a task on a computer.

Programming is the process of designing and building an executable computer program to accomplish a specific task.

Well, according to me programming is like raising a baby. We provide knowledge (data) to help understand a baby what’s happening around. We teach a baby to be disciplined (and much more) by making rules.

Similarly, a computer is like a baby. We set rules and provide data to the computer through executable programs with the help of a Programming Language.

(Photo by Clément H on Unsplash)

That’s it👍. If you can understand this basic concept of programming, you’re good to go. Pick up a programming language and start learning. Read the following section to get an idea of where to start.

My recommendation is to choose Python Programming Language as a start, because it’s beginner-friendly.

#programming #programming-tips #programming-language #programming-top-story #computer-science #data-structures-and-algorithms #tips-for-programmers #coding

Programming In Acceleration: Levelling Up Programming Skills

Some require and some are not. But acceleration programs might require you to build one. I’ll tell you how I made a computer program for the competition.

1-Researched the source codes and pasted it on my VS code

Written on the internet “blockchain-based ticket codes” and found the Ethereum source codes in Github. Then, I’ve just copied and pasted on my VS code by naming with .sol extension. Then, I’ve got my hands on the code itself and started to correct the mistakes that the editor has shown so far. Managed to reduce 189 errors to 58 within two hours. The rest was handled by my teammate when I sent the code I’ve edited. He just fixed the codes in three more hours and my mistake was not to increase the gas price. We increased the gas price on the remix and everything worked. And he just tested the software on scalability and security. It was the perfect garment for us that everything worked except the indentation errors.

What should’ve been done by us

Found all the codes including testing, copied them, and pasted them to our text editor for further analysis. Still, we had the prototype and we could write all the test codes, migrations, etc. if needed. Even more, we should’ve researched the codes to our project before using one of the examples.

#acceleration-program #program-analysis #programming #startup #acceleration #data analysis

Abdullah  Kozey

Abdullah Kozey


Learning C: Input and Output and Two Program Templates

Before I get too deep into C, I need to show you how to get data into and out of your programs. Using assignment for data gets old after a while and you want to be able to have users enter their own data. And you definitely need to be able to see what happens to your data in a program so learning how to display data to the screen is important and necessary.

Besides demonstrating how to perform input and output in C, I will also be demonstrating two templates that are related to those topics — Prompt, Then Read and Input, Process, Output (IPO). The IPO template, in particular, is important because practically every C program you write will use this template.

When I talk about input and output in C, I’ll use the terms standard input and standard output. These terms refer to the default input and output devices on your computer. The standard input device is the keyboard. The standard output device is the computer’s monitor or screen. I will only use the terms input and output and when I use those terms I’m referring to standard input and standard output. If I want to refer to a different device for input and/or output, I’ll use the specific term for that device.

#c-programming-language #c-programming #c-program #c-programming-help

akshay L

akshay L


Top 5 Programming Languages to Learn in 2021 | Top Programming Languages | Intellipaat

In this video, you will know the top 5 Programming languages to learn in 2021. It is always confusing for a beginner to choose a programming language from the pool of tens of languages. So we have come up with this video to help you out chose the best one to start your career with and learn programming fast.

#programming #learn-programming #programming-languages #topprogramminglanguages #top5programminglanguages

Hal  Sauer

Hal Sauer


Thoughts on Visual Programming with Scratch

Scratch is a visual programming language developed by MIT where you program in similar fashion to how you put together lego blocks.
It is actually quite a clever system, as types of expressions are represented with different shapes and colors so that it is obvious which block can go where.
Consider the image below there is yellow control-flow structure called “repeat until” which takes a boolean expression. The loop will end when this expression is true.

#programming #scratch #game-development #makecode #teaching-kids-to-code