weight in controllability theory

 Xdot = AX + Bu

So, weight in matrix A are coefficients of the linear systems, which will influence the co-dependency but not the degree of freedom. So, it seems that weight will influence the which gene would be the driver nodes, but not the minimum number of driver node in theory. However, because aging biological network have noises, to weight would influence the sensitivity and robustness of gene network to aging noises.  

gene ontology annotation

 GAF

limit of human lifespan

 https://www.nature.com/articles/s41467-021-23014-1?utm_medium=affiliate&utm_source=commission_junction&utm_campaign=3_nsn6445_deeplink_PID100052172&utm_content=deeplink#Abs1 

https://www.livescience.com/human-life-span-limit-150-found.html

Jylhävä, J., Pedersen, N. L. & Hägg, S. Biological Age Predictors. EBioMedicine 21, 29–36 (2017).

information theory and aging

noise make the difference Weibull and Gompertz model of aging. 

noise also increase during aging, lead to loss of controllability

information was not lost during aging, in fact information increase during aging. 

Dang funding statement

3 R21 AG064345 02S1. 2020. $23.7 K

5 R21 AG064345 02. 2020 $200K

 5 R01AG052507-05.  2021, $325K 

5 R42 AG058368 03, 2019, $750K,

GO neural network discussion

 

Q: Can we merge sequential layers (GO-nodes)? 

Q: how many GO-nodes (layers) in DCell? After 3 criteria of filtering, only 2500 neurons are left. 

Can we use mask on the vectors to specify inputs at different layers? 

genome instability and aging

 

genome instability and aging, Laura Niedernhofer

https://youtu.be/wg5KZmDCWL8

AGE Presents: Laura Niedernhofer - Genome Instability and Aging.

bayesian network versus bayesian neural networks

 bayesian network inference infer topology (causal inference). 

Bayesian neural network is to apply Bayesian inference on parameters of stochastic nerual networks on the parameters. (not topology). 

NIH Bridge2AI

 This seems to be web-lab focused call, focusing on data generation. 

https://commonfund.nih.gov/bridge2ai/meetings

FFQ

 Python download meta information of SRA

https://github.com/pachterlab/ffq

UTC facility management, work request

 

Our online work request portal is back online!  Thank you for your understanding while it was down. 

 

If you need to submit a work request to facilities, please go here:

 

https://fpmis.utc.edu/

 

Also know that this link only works from the campus intranet.  If you are off campus, you can email us at work-control@utc.edu

 

If you have an emergency request or need immediate assistance, please call us at ext. 2254.

 

 

Michal W. Wells

Michal Wells, MSgt, USAF (Ret.)

Facilities Work Control Manager

Facilities Planning and Management, Dept.3553

Administration Building Room 228-L

University of Tennessee at Chattanooga

Phone: (423) 425-4075

Fax: (423) 425-4749

Webpage:  Facilities Planning & Management

GTEx phenotypes

 

Here is a sample phenotype data that I found from the public site of GTEx. The example file provide attributes of

Sample_id, age, sex, blood test results, postmortem information, sequence entries. 

 

This file is from 

https://storage.googleapis.com/gtex_external_datasets/eyegex_data/annotations/EyeGEx_meta_combined_inferior_retina_summary_deidentified_geo_ids.csv

 

I would expect the unreleased files may contain similar but more disease-related attribute based on GTEx-based publications. 

 

More on phenotypes: 

https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/dataset.cgi?study_id=phs000424.v8.p2&phv=169091&phd=3910&pha=&pht=2742&phvf=&phdf=&phaf=&phtf=&dssp=1&consent=&temp=1

The "Phenotype Datasets" describes 189 variables in the table for v8. 

DoD MINERVA research

 

https://minerva.defense.gov/Research/Research-Priorities/

R coding bootcamp day 1

all zoom breakout zoom for participants to join

ask TA to add TA in their zoom names

 

Doodly, Inkpad

 

Doodly

Illustrator -> draw -> save as SVG -> add into Doodly --> new object

InkPad --> draw -> export as SVG 

Molecular evolution and the decline of purifying selection with age

 

Molecular evolution and the decline of purifying selection with age

https://www.nature.com/articles/s41467-021-22981-9.pdf

This suggests that evolution and population may be informative for machine learning predictive model. 

biases in AI

 

from: 

https://twitter.com/PinakiLaskar/status/1393618068006342656/photo/1

GISAID

May 2021, GISAID genome alignment anymore does show up in my download tab. After contacting the online help, I logged out and re-logged in, the problem was solved. 

reference: 

hCoV-19/Wuhan/WIV04/2019 (WIV04) is the official reference sequence employed by GISAID (EPI_ISL_402124). 

GTEx data portals

 

DBGAP

https://dbgap.ncbi.nlm.nih.gov/aa/wga.cgi?page=pi_requests&filter=wlid&wlid=29078

ANVIL

https://anvil.terra.bio/#profile

https://app.terra.bio/#workspaces

UTC, submit grade change

 

https://mocsutc0.sharepoint.com/sites/gradechange/SitePages/Home.aspx 

https://youtu.be/AERQbvtwjr8

NIH research training grants

 

https://researchtraining.nih.gov/programs/training-grants

T32

T34, undergraduate

T90/R90

lifespan of cells and tissues

 

https://www.sciencefocus.com/the-human-body/what-cells-in-the-human-body-live-the-longest/#:~:text=Although%20the%20our%20bodies%20are,around%20for%20longer%20than%20others.&text=On%20average%2C%20the%20cells%20in,different%20organs%20of%20the%20body.

on average: 7-10

neutrophils, white cell: 2 days

cell in the middle eye lenses: entire lifespan of the host

brain cells: might live longer than the host

Brain cells: 200+ years?

Eye lens cells: Lifetime

Egg cells: 50 years

Heart muscle cells: 40 years

Intestinal cells (excluding lining): 15.9 years

Skeletal muscle cells: 15.1 years

Fat cells: 8 years

Hematopoietic stem cells: 5 years

Liver cells: 10-16 months

Pancreas cells: 1 year

GTEx protected data

 the GTEx portal,  "who can apply":

Extramural Investigators must be permanent employees of their institution at a level equivalent to a tenure-track professor or senior scientist with responsibilities that most likely include laboratory administration and oversight. Laboratory staff and trainees such as graduate students and postdoctoral fellows are not permitted to submit project requests.

Here is a link that walks through the application process: 

https://www.gtexportal.org/home/protectedDataAccess

Here is the "request access" page on the first step:

https://dbgap.ncbi.nlm.nih.gov/aa/wga.cgi?page=login

From there, you can click on the "How does one apply" link and it will walk though all the steps.

Tips: 

https://www.ncbi.nlm.nih.gov/projects/gap/cgi-bin/GetPdf.cgi?document_name=GeneralAAInstructions.pdf

dbGAP project PI: 

Research Use Statement. The approval of project requests depends primarily on a carefully written Research Use Statement (2,200 characters max). The statement should include the following components: • Objectives of the proposed research; • Study design; • Analysis plan, including the phenotypic characteristics that will be evaluated in association with genetic variants; • Explanation of how the proposed research is consistent with the data use limitations for the requested dataset(s); and • Brief description of any planned collaboration with researchers at other institutions, including the name of the collaborator(s) and their institution(s).

From T Day: 

I’ve been doing some digging to try and find the GTEx policy on data limitation. I found this document (attached) which highlights that they follow the NIH Genome Data Sharing GDS policy (found here: 

https://osp.od.nih.gov/scientific-sharing/genomic-data-sharing/ ) with slight variations. These variations are noted in the attached document which I was able to locate in the documentations section of the GTEx portal. 

GTEx Data Release and Publication Policy version 5-8-15

https://storage.googleapis.com/gtex-public-data/GTEx_Data_Release_and_Publication_Policy_v05-08-15.docx 

It is the intent of the NIH to promote the dissemination of research findings from NIH genomic dataset(s) as widely as possible through scientific publication or other appropriate public dissemination mechanisms. The Genotype-Tissue Expression (GTEx) project will follow the 8-27-14 Genomic Data Sharing (GDS) policy (gds.nih.gov/index.html) with the following exceptions:

1. Pilot data set (phs000424.v3.p1) - acceptance of the “Ft. Lauderdale” principles of rapid, pre-publication data release (see Sharing Data from Large-Scale Biological Research Projects: A System of Tripartite Responsibility, 2003). The continued success of rapid pre-publication data release relies on the scientific community to respect the data producer’s interest to publish a full analysis of their data first. Secondary users are asked to refrain from submitting manuscripts describing comprehensive analyses until the Consortium has published their analysis. 

May 8, 2015 - No restrictions - main manuscript published.

2. phs000424.v4.p1 - the dataset is subject to 9 months publication restriction starting from the date of the release; no restrictions after Jan 4, 2015. 

All datasets from phs00424.v5.p1 forward will follow the NIH GDS policy.  This means that once released through dbGaP, there are no restrictions on use or publication.  This document and an accompanying table of dataset releases can be found at http://www.gtexportal.org/home/documentationPage.

Overall, users of GTEx data are strongly encouraged to publish their results in peer-reviewed journals and to present research findings at scientific meetings, etc. Investigators planning to conduct analyses similar to those described at http://www.gtexportal.org/home/documentationPage may contact Consortium members or the NIH program staff at nhgrigtex@mail.nih.gov to discuss collaborations, if so desired

superposition versus entanglement

 From google: 

" superposition is necesary for entanglement, but they are not the same. On a formal level, superposition is just the sum of vectors in a Hilbert space. This vectors represent physical states. This gives the posibility of having a state which is the sum of two classically different states."

end to end deep learning

 

"End to End learning in the context of AI and ML is a technique where the model learns all the steps between the initial input phase and the final output result."

COVID-19 host genetics initiative

 

COVID19 

https://www.covid19hg.org/about/ 

https://github.com/rivas-lab/public-resources

https://github.com/rivas-lab/covid19

multi-task deep learning training

 

which tasks should be learned together?  https://arxiv.org/abs/1905.07553

L_total = a1 L1 + a2 L2 + a3 L3

Easy task versus hard task

Multi task learning using uncertainty to weight lossess for scence geometry and sematics

GradNorm: gradient normalization for adaptive loss balancing in deep multitask networks

laptops

 

Dell,  XPS 9700 , 16G RAM, 17 inch LCD, 512G Hard drive, solid state, NVIDIA GTX 1650 Ti (4G), $1939.97

Dell Mobile Precision 3560, 16G RAM, NVIDIA T500 (2G), 512G harddrive, $1996

So, Dell XPS 9700 is better than Dell Mobile Precision 3560. 

NVIDIA T500 versus GeForce GTX 1650

 Based on the following post, Hong think GTX1650 Ti is a better choice, because some specifics are higher in GTC1650 Ti than T500 mobile. 

https://www.notebookcheck.net/T500-Mobile-vs-GeForce-GTX-1650-Ti-Desktop-vs-GeForce-GTX-1650-Ti-Mobile_10605_9878_10227.247598.0.html

Ngyuen multi task corn yield prediction

 Nyguen mentioned that CUDA does not allow recursion

multiple data as input (normalization, input layers)? 

multi-task training (optimization procedure, weight? )

self-boosted forecasting time series 

slide 31: joining features, 

loss function: weight averaged. what weight? 

multi task, multi feature

 https://www.nature.com/articles/s41586-020-2496-1

A single-cell transcriptomic atlas characterizes ageing tissues in the mouse

• The Tabula Muris Consortium
• 
  

scRNA of cellular aging

Mol. Cells 2021; 44(3): 136~145  https://doi.org/10.14348/molcells.2021.2239
Transcriptomic Analysis of Cellular Senescence: One Step Closer to Senescence Atlas

Sohee Kim  and Chuna Kim 
http://www.molcells.org/journal/view.html?doi=10.14348/molcells.2021.2239
Recently, the Tabula Muris Senis (Mouse Aging Cell Atlas), which analyzes the aging in mice with scRNA-seq, has been released (Almanzar et al., 2020) (Table 3). In this study, 23 organs from 19 male and 11 female mice have been analyzed at 6 time points. The study observed the appearance of aging-specific cell types in various tissues, such as the liver and bladder, and the immune infiltration of various tissues. In addition, the number of T and B cells in gonadal adipose tissues increased with age, and the B cell cluster highly expressing immunoglobulin J was found specifically in old mice. Therefore, scRNA-seq enables the monitoring of the changes in cell population and the discovery of new rare populations associated with aging.

super position == linear combination

 super position == linear combination

linear combination of vectors in the hilbert space