Bytes of Life

A language model enables accurate structural variant detection in whole-genome amplified long-read sequencing

Thu, 23 Jan 2025 00:00:00 +0000

Genomic Language Model Mitigates Chimera Artifacts in Nanopore Direct RNA Sequencing

Thu, 31 Oct 2024 00:00:00 +0000

Aurora Is a Web Application for Visualizing Non-linear Graph

Thu, 04 Apr 2024 00:00:00 +0000

Notes from learning everything

Thu, 21 Dec 2023 00:00:00 +0000

Autonamte Latex Build Processing

Fri, 01 Sep 2023 00:00:00 +0000

Latex
#

LaTeX functions as an intricate tool for the meticulous creation of scientific documents, enriched by a comprehensive ecosystem and an extensive assortment of templates. I habitually utilize LaTeX for a plethora of activities, including the annotation of notes, the authoring of scholarly articles, and the formulation of thesis. GitHub serves as an indispensable medium for version management and collaborative endeavors. As such, automating the assembly procedure of LaTeX documents through GitHub would confer significant benefits.

CUDA for Deep Learning Inference in Rust and C++

Tue, 29 Aug 2023 00:00:00 +0000

1. Deep Learning Inference
#

Currently, both Rust and C++ are emerging as noteworthy contenders in the realm of deep learning, primarily due to their efficiency despite Python’s prevailing dominance in model training. While Python continues to commandeer the training phase, it lags in performance during inference. Large Language Models (LLMs), such as ChatGPT, have burgeoned since their inception, instigating a competitive frenzy among corporations and research organizations alike. This has led to an explosion of various LLMs, although not all exhibit equal utility or value.

Aurora: A Web Application for Visualizing Non-linear Graph

Sat, 26 Aug 2023 00:00:00 +0000

Development Tips

Wed, 05 Jul 2023 00:00:00 +0000

my desktop snapshot

1 hidden desktop icons on macOS
#

defaults write com.apple.finder CreateDesktop false
killall Finder

Otherwise, you want to show the icons again

PxBLAT:An Efficient and Ergonomics Python Binding Library for BLAT

Sun, 25 Jun 2023 00:00:00 +0000

TikZ-Learning Journey

Fri, 02 Jun 2023 00:00:00 +0000

Overview
#

/ [pdf]

Why
#

Welcome to ‘My TikZ-Learning Journey’, a space where I document my exploration with the wonderful TikZ library in LaTeX. LaTeX is a typesetting system used widely in academia and industries for the production of scientific and technical documents. Of its many features, the one I’ve recently been fascinated with is the TikZ library - a versatile package for crafting detailed and professional-grade diagrams.

Efficient Genomic Interval Search Using SIMD-Enhanced COITree

Sun, 12 Mar 2023 00:00:00 +0000

Background
#

In bioinformatics, researchers frequently analyze various types of genomic data, such as DNA sequencing data, RNA sequencing data, and epigenetic data. Manipulating genomic intervals is a crucial task in comprehending the genetic basis of diseases and identifying potential therapeutic targets. Genomic intervals are defined as regions that span from a starting position to an ending position and can encompass genes, regulatory elements, and other functional elements of the genome. One primary application of genomic interval manipulation is analyzing ChIP-seq data. Moreover, manipulating genomic intervals allows for the integration of ChIP-seq data with other genomic data types, such as gene expression and genetic variations. This integration provides a more comprehensive understanding of biological processes and their contribution to normal development or disease. However, integrating these data types into a single data structure can pose challenges, especially when handling large datasets. Cache Oblivious Interval Trees (COITree), with cache-oblivious design and efficient query algorithms, have the potential to handle and integrate multiple types of genomic data into a single data structure. It stores the intervals in contiguous memory and employs in-order van Emde Boas layout to enhance query performance. The tree is designed to optimize cache performance by reducing the number of cache misses during traversal. However, COITree still suffer from performance bottlenecks, particularly when dealing with large datasets. One approach to addressing this bottleneck is to use Single Instruction Multiple Data (SIMD), which is optimized for vector operations, to improve the performance of COITree. Thus, I hypothesize that the approach is a viable solution for improving the speed and efficiency of genomic interval analysis.

How to Use Noodles Library in Rust

Sat, 04 Mar 2023 00:00:00 +0000

1. Introduction
#

Noodles and Rust-htslib are two widely used Rust libraries for genomic data handling. While both libraries are designed to work with genomic data, they take different approaches to achieve this goal. This blog explores Noodles and compares it with Rust-htslib, while also discussing its potential pitfalls.

Collections of Amazing Vim Tips

Wed, 21 Dec 2022 00:00:00 +0000

Why
#

This blog will evolve to including more valuable content. Using nvim tag to filter more content about nvim and others.

Tiling Windows Management in macOS

Tue, 20 Dec 2022 00:00:00 +0000

1. Aims
#

The philosophy of Vim’s motion has had a profound influence on me, leading me to fully embrace Neovim. The notion of having to reposition windows using a mouse is anathema to me. I greatly value the experience of utilizing solely the keyboard. As such, I set out to find a tool to aid me in managing windows seamlessly. Eventually, I discovered a quartet of exceptional tools that I highly recommend giving a try. To this day, they continue to surpass my expectations and enhance my productivity.

Make A Powerful Terminal Workspace

Thu, 15 Dec 2022 00:00:00 +0000

I recently made a change to my terminal setup, switching from using iTerm2 + Zsh + Tmux to Alacritty + Fish + Zellij. Now I use Wezterm + Fish + Zellij. I discovered that my new combination is extremely powerful and versatile. The flexibility to highly customize each of these tools was a significant factor in my decision to keep them.

macOS Setup for Development and Research

Sat, 26 Nov 2022 00:00:00 +0000

Gallery
#

TODO
#

add latex
add lsd
add fd
add fzf
add ripgrep
add procs
add zoxide
add gitui
add btop

TLDR
#


alacritty	office	aldente	one switch
alfred	pdf expert	alttab	picgo
bartender	pycharm	cheat.sh	rectangle
chrome	reeder 5	clion	rust
conda	snippetslab	default folder x	soundsource
docker	spacevim	dust	time sink
ferdi	tldr	fish	tmux
git	tmuxinator	hyperfine	vim
imagine	xcode	ishot	xmind
iterm2	zellij	lunarvim	zoom
magnet	zotero	mamba	zsh
micromamba	fisher	miniforge	ouch
monitorcontrol	topgrade	monodraw	imagemagick
neovim	youtobe-dl	notion	jetbrains
google drive	transmit	homebrew	tree
fluent reader	wezterm	iina	vs code

1. Introduction
#

Given the time and effort required for migrating configurations, I have decided to create a blog documenting the entire process. The configurations will be divided into two sections, the first of which will cover the software that I frequently use, and the second will contain the configuration files themselves. As I have recently acquired a MacBook Pro M1 model, here is a list of the tools I am currently utilizing.

Bioinformatics Algorithm Library aka BINARY

Mon, 26 Sep 2022 13:07:28 -0500

The library is a collection of algorithms and data structures that are designed for modern C++ bioinformatics applications. You can use the library in your own projects or as a part of a larger project.

Bioinformatics Toolbox Aka Boss

Sun, 25 Sep 2022 23:22:31 -0500

BOSS is a bioinformatics toolbox, which will contain efficient tools. It is written in modern C++ and is tested exhaustively. It is designed to be easy to use and time-efficient. BOSS is a free software and is distributed under the terms of the GNU General Public License V3.

Code Snippets

Thu, 22 Sep 2022 00:00:00 +0000

Get Random numbers
#

#include <algorithm>
#include <iostream>
#include <iterator>
#include <random>

int main() {
 std::random_device rd;
 std::mt19937 rng(rd());
 std::uniform_int_distribution<int> dist6(1, 6);
 std::generate_n(std::ostream_iterator<int>(std::cout, " "), 10,
 [&dist6, &rng]() { return dist6(rng); });
 return 0;
}

Get random numbers between min and max
#

// Generate a random number between min and max (inclusive)
// Assumes std::srand() has already been called
// Assumes max - min <= RAND_MAX
int getRandomNumber(int min, int max)
{
 static constexpr double fraction { 1.0 / (RAND_MAX + 1.0) }; // static used for efficiency, so we only calculate this value once
 // evenly distribute the random number across our range
 return min + static_cast<int>((max - min + 1) * (std::rand() * fraction));
}

Clear Input Stream
#

if (std::cin.fail()) // has a previous extraction failed or overflowed?
{
 // yep, so let's handle the failure
 std::cin.clear(); // put us back in 'normal' operation mode
 std::cin.ignore(std::numeric_limits<std::streamsize>::max(), '\n'); // and remove the bad input
}

Compare Float Number
#

#include <algorithm>
#include <cmath>

bool approximatelyEqualAbsRel(double a, double b, double absEpsilon, double relEpsilon)
{
 // Check if the numbers are really close -- needed when comparing numbers near zero.
 double diff{ std::abs(a - b) };
 if (diff <= absEpsilon)
 return true;

 // Otherwise fall back to Knuth's algorithm
 return (diff <= (std::max(std::abs(a), std::abs(b)) * relEpsilon));
}

Shelang for Python
#

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
@author: YangyangLi
@contact: yangyang.li@northwestern.edu
@version: 0.0.1
@license: MIT Licence
@file: cli.py.py
@time: 2020/12/28 10:21 PM
"""

Shelang for Bash
#

#!/bin/bash
set -e
set -u
set -o pipefail

CLI for Rust
#

#[allow(unused)]
use clap::Parser;
use env_logger::Builder;
use human_panic::setup_panic;
use log::info;


#[derive(Parser, Debug)]
#[command(name = "sv2gf")]
#[command(version = "0.1.0")]
struct Args {
 /// The input file of svs
 #[arg(short = 's', long)]
 sv: PathBuf,

 /// The input file of gfs
 #[arg(short = 'g', long)]
 gf: PathBuf,

 /// The Distance threshold
 #[arg(short = 'd', long, default_value = "1000000")]
 dis_threshold: u32,

 #[clap(flatten)]
 verbose: clap_verbosity_flag::Verbosity,
}

fn cli() -> (PathBuf, PathBuf, u32) {
 let args = Args::parse();
 let sv = args.sv;
 let gf = args.gf;
 let dis_threshold = args.dis_threshold;
 info!("sv: {:?}, gf: {:?}", sv, gf);
 (sv, gf, dis_threshold)
}

fn main() {
 setup_panic!();
 let args = Args::parse();

 Builder::new()
 .filter_level(args.verbose.log_level_filter())
 .init();

 let (input_sv, input_gf, dis) = cli();
}

Export port from two layers nodes
#

ssh -L 8899:localhost:8899 quest ssh -N -L 8899:localhost:8899 qgpu0101

Timer for C++
#

#include <chrono> // for std::chrono functions

class Timer
{
private:
 // Type aliases to make accessing nested type easier
 using clock_type = std::chrono::steady_clock;
 using second_type = std::chrono::duration<double, std::ratio<1> >;

 std::chrono::time_point<clock_type> m_beg { clock_type::now() };

public:
 void reset()
 {
 m_beg = clock_type::now();
 }

 double elapsed() const
 {
 return std::chrono::duration_cast<second_type>(clock_type::now() - m_beg).count();
 }
};

Gradient Clipping
#

def grad_clipping(net, theta):
 if isinstance(net, nn.Module):
 params = [p for p in net.parameters() if p.requires_grad]
 else:
 params = net.params

 norm = torch.sqrt(sum(torch.sum((p.grad ** 2)) for p in params )) # include all parameters

 if norm > theta:
 for param in params:
 param.grad[:] *= theta/nor

RST Docstring directive
#

:Example:

>>> print("hello world!")
hello world!

.. note:: can be useful to emphasize
.. seealso:: :class:`MainClass2`
.. warning:: arg2 must be non-zero.
.. todo:: check that arg2 is non zero.

Git remove large files
#

git filter-repo --force --strip-blobs-bigger-than 100M

Stop Colab Disconnect
#

function KeepClicking() {
 console.log("Clicking");
 document.querySelector("colab-connect-button").click();
}
setInterval(KeepClicking, 60000);

Open Colab
#

[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/googlecolab/colabtools/blob/master/notebooks/colab-github-demo.ipynb)

Srun for interactive shell
#

srun -n 1 -t 1:00:00 -p gpu --gres=gpu:tesla:1 --pty bash

Bash info for Slurm
#

#!/bin/bash -l
#SBATCH --job-name=MLHW3 # job name
#SBATCH --output=MLHW3_%j.log # log name
#SBATCH --time=8:00:00 # time
#SBATCH --ntasks=1 # tasks number
#SBATCH --cpus-per-task=8 # cpu number
#SBATCH --ntasks-per-node=1 # node number
#SBATCH --mem=30G # total memory
#SBATCH --tmp=30G #
#SBATCH --mail-type=ALL
#SBATCH --mail-user=li002252@umn.edu
#SBATCH --gres=gpu:v100:1 # gpu
#SBATCH -p v100 # partitions
cd $SLURM_SUBMIT_DIR
date;hostname;pwd

Conda export env
#

conda env export | grep -v '^prefix' > freeze.yml
conda env create -f freeze.yml

Multiple output in Jupyter
#

from IPython.core.interactiveshell import InteractiveShell
InteractiveShell.ast_node_interactivity = "all"

Multiple Images in Latex
#

\usepackage{subcaption}

\begin{figure}[h]
 \begin{subfigure}{0.5\textwidth}
 \includegraphics[width=0.9\linewidth, height=5cm]{rna_ttest.png}
 \caption{T-test for RNA-Seq data}
 \label{fig:subim1}
 \end{subfigure}
 \begin{subfigure}{0.5\textwidth}
 \includegraphics[width=0.9\linewidth, height=5cm]{rna_ranksum.png}
 \caption{Rank-sum for RNA-Seq data}
 \label{fig:subim2}
 \end{subfigure}

 \begin{subfigure}{0.5\textwidth}
 \includegraphics[width=0.9\linewidth, height=5cm]{micro_ttest.png}
 \caption{T-test for Microarray data}
 \label{fig:subim1}
 \end{subfigure}
 \begin{subfigure}{0.5\textwidth}
 \includegraphics[width=0.9\linewidth, height=5cm]{micro_ranksum.png}
 \caption{Rank-sum for Microarray data}
 \label{fig:subim2}
 \end{subfigure}
\caption{Histogram of the p-values of all genes}
\end{figure}

Show fonts in Jupyter
#

import matplotlib.font_manager
from IPython.core.display import HTML

def make_html(fontname):
 return "<p>{font}: <span style='font-family:{font}; font-size: 24px;'>{font}</p>".format(font=fontname)

code = "\n".join([make_html(font) for font in sorted(set([f.name for f in matplotlib.font_manager.fontManager.ttflist]))])

HTML("<div style='column-count: 2;'>{}</div>".format(code))

Display image in Jupyter
#

from IPython.display import Image
from IPython.core.display import HTML
PATH = "tree_default_max_depth.png"
Image(filename = PATH , width=900, height=900)

Short Code Example

Fri, 24 Jun 2022 00:00:00 +0000

Alert
#

{{< alert >}}
**Warning!** This action is destructive!
{{< /alert >}}

Warning! This action is destructive!

{{< alert cardColor="red" >}}
**Warning!** This action is destructive!
{{< /alert >}}

Warning! This action is destructive!

{{< alert textColor="red" >}}
**Warning!** This action is destructive!
{{< /alert >}}

Warning! This action is destructive!

{{< alert "twitter" >}}
**Warning!** This action is destructive!
{{< /alert >}}

Do not forget to follow me

{{< alert icon="fire" cardColor="#e63946" iconColor="#1d3557" textColor="#f1faee" >}}
**Warning!** This action is destructive!
{{< /alert >}}

Do not forget to follow me

Article
#

{{< article link="/posts/002-macos-configuration/" >}}

Badge
#

{{< badge >}}
New article!
{{< /badge >}}

New article!

Button
#

{{< button href="#button" target="_self" >}}
Call to action
{{< /button >}}

Call to action

{{< button href="https://blowfish.page/docs/shortcodes/#icon" target="_self" >}}
Call to action
{{< /button >}}

Open external links

Carousel
#

carousel is used to showcase multiple images in an interactive and visually appealing way. This allows a user to slide through multiple images while only taking up the vertical space of a single one. All images are displayed using the full width of the parent component and using one of the predefined aspect ratios of 16:9, 21:9 or 32:9.

C++ Development in Bioinformatics

Wed, 15 Jun 2022 23:17:30 -0500

1.1 Config Compile Environment
#

I am currently planning to develop a tool using C++ in both Linux and macOS environments. However, I frequently encounter obstacles in the form of lacking root access to download dependencies using apt-get install -y dependencies directly in Ubuntu. Navigating the complicated dependency chain and compiling each library individually can be time-consuming, often taking a night or even a week to complete. One solution to this issue is to use a package manager such as Conda, which is primarily used in the data science domain. Conda offers support for other languages such as C++, Rust and R as well. Concrete package names may change at any time, and it’s necessary to search for the real package name. Therefore, Conda can be useful tool for installing C++ dependencies, particularly in the bioinformatics domain. It’s worth mentioning that there are several other solutions available for managing C++ dependencies such as Vcpkg, Conan, and I use CPM as an alternative option.

Pybox: A Python Toolbox for Common Development Tasks

Fri, 03 Dec 2021 00:03:50 -0600

Introduction
#

Have you ever found yourself repeatedly searching for commands to download files from Google Drive or send messages to Slack? I created Pybox to solve this problem - a Python-based command-line toolbox that consolidates useful utilities for common development tasks.

PCA by Python

Sun, 04 Apr 2021 23:58:37 +0800

1. Introduction
#

This article records two methods of PCA analysis using Python, and visualizes 2-dimensional results

Face Recognition in Video

Thu, 20 Jun 2019 00:00:00 +0000

1. 准备工作
#

使用Python实现需求，环境需求比较严苛，所需要的主要依赖：

Requirements

dlib 由 C++编写，提供了和机器学习、数值计算、图模型算法、图像处理等领域相关的一系列功能
face-recognition 已经经过训练成熟的识别人脸的工具
imutils 用来操作系统文件和文件夹
Opencv是用来操作视频流，并将视频流转换格式

1.1 环境搭建
#

通过 pip 进行安装，在使用 pip 进行安装时，应该注意切换到国内源进行下载，提高下载速度，下面分享一下当前国内的 pip 源：

Nei Saitou Neighbor Joining

Wed, 03 Apr 2019 11:15:00 +0800

1. Background
#

Before diving into code, the description of NJ algorithm can be found in

, where first column indicates parent node, and second column is its children node, the last column is the value of edge.

About

Mon, 01 Jan 0001 00:00:00 +0000

About Me
#

Hello! I’m Yangyang Li, a fifth-year Ph.D. candidate in the DGP Program at Northwestern University Feinberg School of Medicine. As I approach graduation, I’m actively exploring opportunities where I can apply my expertise in AI and computational biology to solve real-world challenges.

Bytes of Life

A language model enables accurate structural variant detection in whole-genome amplified long-read sequencing

Genomic Language Model Mitigates Chimera Artifacts in Nanopore Direct RNA Sequencing

Aurora Is a Web Application for Visualizing Non-linear Graph

Notes from learning everything

Autonamte Latex Build Processing

Latex #

CUDA for Deep Learning Inference in Rust and C++

1. Deep Learning Inference #

Aurora: A Web Application for Visualizing Non-linear Graph

Development Tips

1 hidden desktop icons on macOS #

PxBLAT:An Efficient and Ergonomics Python Binding Library for BLAT

TikZ-Learning Journey

Overview #

Why #

Efficient Genomic Interval Search Using SIMD-Enhanced COITree

Background #

How to Use Noodles Library in Rust

1. Introduction #

Collections of Amazing Vim Tips

Why #

Tiling Windows Management in macOS

1. Aims #

Make A Powerful Terminal Workspace

macOS Setup for Development and Research

Gallery #

TODO #

TLDR #

1. Introduction #

Bioinformatics Algorithm Library aka BINARY

Bioinformatics Toolbox Aka Boss

Code Snippets

Get Random numbers #

Get random numbers between min and max #

Clear Input Stream #

Compare Float Number #

Shelang for Python #

Shelang for Bash #

CLI for Rust #

Export port from two layers nodes #

Timer for C++ #

Gradient Clipping #

RST Docstring directive #

Git remove large files #

Stop Colab Disconnect #

Open Colab #

Srun for interactive shell #

Bash info for Slurm #

Conda export env #

Multiple output in Jupyter #

Multiple Images in Latex #

Show fonts in Jupyter #

Display image in Jupyter #

Short Code Example

Alert #

Article #

Badge #

Button #

Carousel #

C++ Development in Bioinformatics

1.1 Config Compile Environment #

Pybox: A Python Toolbox for Common Development Tasks

Introduction #

PCA by Python

1. Introduction #

Face Recognition in Video

1. 准备工作 #

1.1 环境搭建 #

Nei Saitou Neighbor Joining

1. Background #

About

About Me #

Recent News 🚀

DeepChopper Is Published: A New Language Model

2026-01-19

Latex
#

1. Deep Learning Inference
#

1 hidden desktop icons on macOS
#

Overview
#

Why
#

Background
#

1. Introduction
#

Why
#

1. Aims
#

Gallery
#

TODO
#

TLDR
#

1. Introduction
#

Get Random numbers
#

Get random numbers between min and max
#

Clear Input Stream
#

Compare Float Number
#

Shelang for Python
#

Shelang for Bash
#

CLI for Rust
#

Export port from two layers nodes
#

Timer for C++
#

Gradient Clipping
#

RST Docstring directive
#

Git remove large files
#

Stop Colab Disconnect
#

Open Colab
#

Srun for interactive shell
#

Bash info for Slurm
#

Conda export env
#

Multiple output in Jupyter
#

Multiple Images in Latex
#

Show fonts in Jupyter
#

Display image in Jupyter
#

Alert
#

Article
#

Badge
#

Button
#

Carousel
#

1.1 Config Compile Environment
#

Introduction
#

1. Introduction
#

1. 准备工作
#

1.1 环境搭建
#

1. Background
#

About Me
#