BOL: Related items

Machine Learning for Genomics

Jit — Thu, 09 Sep 2021 11:26:32 -0500

Module 1: Statistics for genomics (2-8 August 2021)

A simple intro to statistical distributions
hypothesis testing
linear models.

reading: http://compgenomr.github.io/book/stats.html

slides: https://github.com/BIMSBbioinfo/compgen2021/tree/main/week1/compgen2021_stats.pdf

exercises+code: https://github.com/BIMSBbioinfo/compgen2021/tree/main/week1/

Module 2: Unsupervised learning for genomics (9-15 August 2021)

Understanding basic intuition behind machine learning approaches.
Using unsupervised learning to cluster and visualise data points
Dimension reduction techniques for visualisation and as input to clustering methods

reading: http://compgenomr.github.io/book/unsupervisedLearning.html

slides: https://github.com/BIMSBbioinfo/compgen2021/tree/main/week2/compgen2021_unsupervisedLearning.pdf

exercises+code: https://github.com/BIMSBbioinfo/compgen2021/tree/main/week2/

Module 3: Supervised learning for genomics (16-22 August 2021)

Understanding and using supervised learning methods for predictive purposes
How to measure prediction performance
Understand and use cross-validation and related concepts

reading: http://compgenomr.github.io/book/supervisedLearning.html

slides: https://github.com/BIMSBbioinfo/compgen2021/tree/main/week3/compgen2021_supervisedLearning.pdf

exercises+code: https://github.com/BIMSBbioinfo/compgen2021/tree/main/week3/

https://github.com/BIMSBbioinfo/compgen2021

Address of the bookmark: https://github.com/BIMSBbioinfo/compgen2021

NVIDIA and Arc Institute Unveil Evo 2: A Breakthrough AI for DNA Design

BioStar — Fri, 21 Feb 2025 10:39:47 -0600

NVIDIA and the Arc Institute have introduced Evo 2, a groundbreaking AI model designed to understand, predict, and generate DNA sequences. This marks a major advancement in computational biology, offering scientists an unprecedented tool to decode the genetic blueprint of life and even design entirely new biological systems.

The Power of Evo 2: AI Meets DNA

Evo 2 is the largest AI model for biology ever created, trained on an astonishing 9.3 trillion DNA "letters" (nucleotides) carefully selected from genomes spanning the entire tree of life. This massive dataset ensures that Evo 2 can recognize patterns and relationships in genetic sequences at an unparalleled scale.

For the first time, scientists can design DNA with AI, moving beyond simple sequence analysis to active DNA generation. Evo 2 enables researchers to predict, modify, and even create entire genetic sequences, opening new possibilities in medicine, agriculture, and synthetic biology.

Decoding the Dark Genome

One of the biggest challenges in genetics is understanding the non-coding regions of DNA—vast stretches of the genome that do not code for proteins but play crucial roles in regulating gene expression. These regions control when and how genes are activated, influencing everything from development to disease.

Evo 2 is designed to decode these non-coding elements, helping researchers uncover their functions and use this knowledge to develop gene-based therapies, synthetic life forms, and precision agriculture solutions.

From Reading DNA to Writing It

To put Evo 2’s impact into perspective:

Previous AI models could "read" DNA like a book, analyzing genetic sequences and identifying patterns.
Evo 2 can "write" entirely new DNA, designing functional genes, chromosomes, and even full genomes from scratch.

This means scientists can now engineer biological systems with AI, designing new proteins, metabolic pathways, and genetic circuits to address real-world challenges.

A Step Toward Generative Biology

The Arc Institute describes Evo 2 as a major step toward "generative biology"—a revolutionary approach where AI is used to create novel biological structures rather than just analyzing existing ones. This could lead to breakthroughs such as:

New medicines: AI-generated enzymes and proteins tailored for targeted therapies.
Disease-resistant crops: Genetically optimized plants for higher yield and climate resilience.
Synthetic organisms: Custom-designed microbes for bioremediation, biofuel production, and industrial applications.

An Open-Source Revolution

Unlike many proprietary AI models, Evo 2 is open source, making its capabilities accessible to researchers worldwide. This democratization of AI-driven biology means that scientists from different disciplines can collaborate, experiment, and innovate, accelerating discoveries in genetic engineering and synthetic biology.

With Evo 2, the boundaries of what’s possible in DNA design, genetic engineering, and biological innovation are being redrawn. The future of life sciences is no longer just about understanding life’s code—it’s about writing it.

Automatic Predictive Model Constructor - APMC

Jan Bińkowski — Mon, 16 Sep 2019 09:43:21 -0500

I would like to invite everyone interested in the subject of machine learning in life science, to test APMC module,

it`s a fully automatic tool (created by students) to simply create and develop supervised machine learning models

for classification and regression purposes. Links to tool, instruction and documentation bellow:

APMC: https://gene-calc.pl/apmc
How to use: https://gene-calc.pl/apmc/how-to-use
Documentation: https://gene-calc.pl/apmc/documentation

List of useful machine / ai learning resources !

Rahul Nayak — Tue, 30 Mar 2021 08:56:06 -0500

ML cheatsheet !

https://github.com/remicnrd/ml_cheatsheet

Visual AI / ML

https://setosa.io/ev/

Simple and efficient tools for predictive data analysis

https://scikit-learn.org/stable/

The Best Free Crash Course on Large Language Models (LLMs) I've Come Across

Neel — Tue, 04 Aug 2026 04:57:00 -0500

Stanford CME 295: Transformers & Large Language Models course by Afshine Amidi and Shervine Amidi. The official course website contains the syllabus, slides, and links to all lecture recordings.

Official course

Stanford CME 295 – Transformers & Large Language Models

Official YouTube playlist

Stanford Online – CME 295 Playlist

Course schedule

Transformer – Tokenization, embeddings, attention, Transformer architecture
Transformer-Based Models & Tricks – RoPE, MQA/GQA, BERT variants
Large Language Models – GPT, MoE, prompting, Chain-of-Thought
LLM Training – Pretraining, SFT, LoRA, optimization
LLM Tuning – RLHF, DPO, preference tuning
LLM Reasoning – Reasoning models, GRPO, scaling
Agentic LLMs – RAG, tool calling, ReAct, agents
LLM Evaluation – LLM-as-a-Judge, benchmarks, biases
Current Trends & Recap – Vision Transformers, diffusion-based LLMs, future directions

If you're looking for a research-level understanding of LLMs (beyond prompt engineering), this is one of the best freely available university courses currently available. It pairs particularly well with:

Project Associate-I | Project Associate-II | Senior Project Associate @ IGIB

Thu, 05 Aug 2021 16:11:32 -0500

Experience in Next Generation Sequencing (NGS) application and interest in Genomics/ Clinical / Translational Applications. OR Good computational programming skills and deep interest in working on interface of Genomics and Clinical application.

Project Scientist-I
Experimental / Computation analysis experience in highthroughput genomics/ clinical application.

Project Manager
Experience in handling large biological projects involving high-throughput genomics/ clinical application.

Scientific Administrative Assistant
Lab Work.

More at https://vinodscaria.genomes.in/positionsopen

Detect the sequence pattern and its location in fasta file with match and mismatches information.

Jit — Thu, 26 Sep 2013 15:02:53 -0500

This script is one of my old script to detect some centromeric pattern in chromosomes. User can also control the number of mismatches allowed through command line ..

To run:

perl centro.pl

AI Uncovers Hidden Secrets in Bacterial DNA, Opening New Frontiers in Genomic Research

Jitendra Narayan — Fri, 25 Sep 2026 22:38:42 -0500

Scientists are now using artificial intelligence in order to examine sections of bacterial DNA that have not been looked at before. This method is showing potential RNA interactions and previously unknown genetic systems which could alter our understanding of microbes. A new study presents Minerva, a genome language model, demonstrating that AI can assist researchers in identifying biological patterns that traditional techniques might fail to detect.

The study, which was made available as a preprint on bioRxiv on 23 September 2026, focuses on the non-coding sections of DNA that are still largely unknown. Although these areas do not produce proteins, they can contain important signals and instructions which have an effect on cell function. The research presents a novel approach to investigating how bacteria handle and control their genetic information.

AI maps previously unexplored genomic regions

The team developed Minerva in order to identify possible interactions between different regions in microbial genomes; rather than depending on similarities with known sequences, Minerva predicts these relationships directly from the DNA by using patterns learned by a genome language model.

On 150 bacterial genomes, Minerva identified a large number of interactions that were not included in the existing annotations. The researchers stated that 84.3 per cent of the predicted intergenic base-pairing interactions were not present in the current annotations, which demonstrates that AI can be of help in generating new ideas in biology.

Unusual RNA structures and viral genetic systems identified

The study also examined a bacterial RNA family known as TwoAYGGAY in Pseudomonas; the model anticipated longer RNA structures and identified associations with repeated DNA sequences, thus providing new insights into how these non-coding elements are organised and how they have evolved.

In a separate section of the study, the researchers examined reverse transcriptase systems associated with bacteriophages, which are viruses that infect bacteria. They identified RNA arrays that maintain their structure but have different sequences and were linked to Unknown Group 27 reverse transcriptases. The findings indicate that these RNAs could function as templates for the production of complementary DNA that is capable of forming hairpin shapes.

The researchers also observed that Minerva was able to detect patterns associated with protein-coding areas, even though it had not been trained to do so. This indicates that genome language models may pick up on biological signals that go beyond what they were intended to identify.

Implications for future genomic research

The fact that artificial intelligence is becoming increasingly important in the field of microbial genomics is shown by the fact that models such as Minerva are able to predict interactions and identify patterns in areas which have not been extensively studied, thus helping researchers to decide what to study next and enabling them to gain a better understanding of biological systems that are still not well understood.

Yet the predictions do not reveal the exact function of each element identified. In order to verify which of the predicted interactions actually take place in living cells and the way in which they affect the microbes, experiments will be necessary. Although the study has undergone peer review, it does nonetheless offer a promising illustration of how machine learning can complement traditional genomics and assist scientists in moving from the identification of known genes to the exploration of the complex relationships that shape microbial life.

More at https://www.biorxiv.org/content/10.64898/2026.09.22.753630v2.full.pdf

Bookmarks Biostatistics materials and books

Jit — Tue, 08 Nov 2016 07:42:42 -0600

Biostatistics did not spring fully formed from the brow of R. A. Fisher, but evolved over many years. This process is continuing, although it may not be obvious from the outside. It has been ten years since the first edition of this book appeared (and rather longer since it was begun). Over this time, new areas of biostatistics have been developed and emphases and interpretations have changed

Please bookmarks your favourate biostatistics books in commend sectons ...

Address of the bookmark: http://www.cos.ufrj.br/~bioestatistica/livros/Introduction%20to%20Biostatistics.pdf

Machine learning in Perl

Jit — Sun, 16 Feb 2020 15:32:03 -0600

this is a fourth blog post in the Machine learning in Perl series, focusing on the AI::MXNet, a Perl interface to Apache MXNet, a modern and powerful machine learning library.

If you're interested in refreshing your memory or just new to the series, please check previous entries over here: 1 2 3

https://metacpan.org/pod/AI::MXNet

Address of the bookmark: http://blogs.perl.org/users/sergey_kolychev/2018/07/machine-learning-in-perl-kyuubi-goes-to-a-modelzoo-during-the-starry-night.html