Janjan

BS Bio | MS Bioinformatics | Data Engineer

[Janjan @dagsdags13]

ngl matlab is pretty decent and intuitive for image processing

[Janjan @dagsdags13]

A tip for people working with NGS data:

Compressing FASTQ files using gzip will result in a modest ~55% reduction in file size.

If possible, compress your read files.

[Janjan @dagsdags13]

🤔🤔🤔

[Rafael Irizarry @rafalab]

This UMAP plot was generated from data simulated without inherent clusters, meaning the observed clusters are artifacts.

It highlights how default Seurat clustering and UMAP settings can sometimes produce misleading patterns in data visualization.

[Janjan @dagsdags13]

Data preprocessing is a prerequisite to any good NGS-based analysis. Some important steps in quality control include adapter trimming and quality score filtering.

Here's a tutorial on how to get started with quality control: https://t.co/Q2eAmX1CAu

#bioinformatics #genomics

[Zhou Xian @zhou_xian_]

Everything you love about generative models — now powered by real physics!

Announcing the Genesis project — after a 24-month large-scale research collaboration involving over 20 research labs — a generative physics engine able to generate 4D dynamical worlds powered by a physics simulation platform designed for general-purpose robotics and physical AI applications.

Genesis's physics engine is developed in pure Python, while being 10-80x faster than existing GPU-accelerated stacks like Isaac Gym and MJX. It delivers a simulation speed ~430,000 faster than in real-time, and takes only 26 seconds to train a robotic locomotion policy transferrable to the real world on a single RTX4090 (see tutorial: https://t.co/bEkIlCKqdf).

The Genesis physics engine and simulation platform is fully open source at https://t.co/DhBv7NdyqH. We'll gradually roll out access to our generative framework in the near future.

Genesis implements a unified simulation framework all from scratch, integrating a wide spectrum of state-of-the-art physics solvers, allowing simulation of the whole physical world in a virtual realm with the highest realism.

We aim to build a universal data engine that leverages an upper-level generative framework to autonomously create physical worlds, together with various modes of data, including environments, camera motions, robotic task proposals, reward functions, robot policies, character motions, fully interactive 3D scenes, open-world articulated assets, and more, aiming towards fully automated data generation for robotics, physical AI and other applications.

Open Source Code: https://t.co/DhBv7NdyqH
Project webpage: https://t.co/SBNyhFB0yn

Documentation: https://t.co/3yuBoaealV

1/n

[Steven Salzberg 💙💛 @StevenSalzberg1]

The importance of peer-reviewed bioinformatics methods: a short rant about a recent paper by some leading scientists in my field. Most scientists would agree that well-engineered computational methods are critically important in genomics... 1/8

[Tom Ashbyトム アシュビー @tomaashby]

The Simpsons on PhD students - almost twenty years old & more cutting than ever

[Janjan @dagsdags13]

GATK is such a pain in the ass to use

[Neil Renic @NC_Renic]

Refusing to delete a crap paragraph because it took so long to research

[Janjan @dagsdags13]

The hesitation to include source code in ML-based research papers. Why do these even get published? 🥱

[The Nobel Prize @NobelPrize]

BREAKING NEWS
The Royal Swedish Academy of Sciences has decided to award the 2024 #NobelPrize in Chemistry with one half to David Baker “for computational protein design” and the other half jointly to Demis Hassabis and John M. Jumper “for protein structure prediction.”

[Janjan @dagsdags13]

Regardless, we try to simplify.

[Janjan @dagsdags13]

The bio ➡️ math and math ➡️ bio translations are asymmetrical. It's very challenging to reduce a biological problem into a simple algorithm or an elegant equation, especially if you're not classically trained as a computational person. A deep frustration of mine. 🥹

[Janjan @dagsdags13]

@BioProfBarker Wonderful! Thanks for sharing

[Julia Bauman @JuliaBauman2]

You’ve probably all seen the see-through mouse by now. Here’s a little more detail on how this optical trick works!

Amazing work from @HongNeuroTech in @ScienceMagazine

[Janjan @dagsdags13]

Wrote a short article explaining the two foundational algorithms for sequence alignment:

- Needleman-Wunsch for global alignments
- Smith-Waterman for local alignment

Here's the link for interested readers:
https://t.co/IYFmmUy7qE

[Janjan @dagsdags13]

Fits right into my note-taking workflow. Gotta love neorg.

[Bo Wang @BoWang87]

🚀 The Segment Anything Model (SAM) has been upgraded to SAM2, featuring an efficient image encoder for segmenting images and videos. But does SAM2 outperform SAM1 in medical image and video segmentation?

We're thrilled to present our paper "Segment Anything in Medical Images and Videos: Benchmark and Deployment"! We comprehensively benchmark SAM2 across 11 medical image modalities and videos.

📄 Paper: https://t.co/NSymKcOJ8q
💻 Code: https://t.co/9B7CG8J655

**Highlights:**
1. SAM2 doesn���t always outperform SAM1 in 2D medical images, but excels in video segmentation, making it more accurate and efficient for 3D images, such as CT and MR scans.

2. MedSAM still outperforms SAM2 on most 2D modalities, but SAM2 surpasses MedSAM for 3D image segmentation in a slice-by-slice approach.

3. Segmentation performance varies with model size; sometimes the smallest model outperforms larger ones.

4. Fine-tuning SAM2 significantly boosts its performance for medical image segmentation.

While SAM2 may struggle with challenging objects that have unclear boundaries or low contrast, it excels in generating good initial segmentation masks for common medical images and videos. However, the official interface doesn’t support medical data formats and has limitations on video length. To address this, we've developed a 3D Slicer Plugin and Gradio API for efficient 3D medical image and video segmentation. We invite you to try them out and provide feedback!

🔧 Deployment:
- 3D Slicer Plugin: https://t.co/j83JChav2r
- Gradio API: https://t.co/4zJUuPFR12

(Note: Due to GPU limitations, the online API is available for only 12 hours and may be slow. We highly recommend deploying the Gradio API with your own computing resources: https://t.co/q5UydWs6Xd

A big shoutout to Jun Ma (@JunMa_11) who recently joined our UHN AI hub (@UHNAIHUB) as Machine Learning Lead, and kudos to all co-authors: Sumin Kim, Feifei Li, Mohammed Baharoon (@BaharoonMS), Reza Asakereh, and Hongwei Lyu! This is true teamwork!

Looking forward to collaborating with the community to advance 3D medical image and video segmentation foundation models!

@UHN @UofTCompSci @UofT_LMP @UofT_TCAIREM @VectorInst

#MedTech #AIinHealthcare #DeepLearning #MedicalImaging #SAM2 #MedSAM #AIResearch

[Janjan @dagsdags13]

[2/30] Central Dogma of Molecular Biology

Part 2: Translation

1. Create a mapping between all unique codons and their corresponding amino acid.
2. Split the input RNA string into 3-mers to represent a codon list.
3. Iterate over the codon list. If it is a stop codon, terminate the loop and return the peptide string. Otherwise, append the mapped value of the current codon to the growing peptide string.

[Janjan @dagsdags13]

A good analysis starts with clean, wrangled data.

Here's a few tools for downloading biological data from the command line:

- curl, wget, aria2
- sra-toolkit, datasets
- genomepy
- refgenie

Learn how to use them by going thru this tutorial:

https://t.co/2HBfuHVu0f