Closed ifeoluwafavour closed 2 years ago
Why I am interested in Ersilia Ersilia's ultimate goal is to effectively lower the barrier to drug discovery, encouraging academic groups and enterprises to pursue the development of new medicines and this excites me.
I'm an Industrial Chemistry undergraduate and I attended a public lecture in my university here in Nigeria. In that event, lecturers and research scientists complained about the limitations they faced when it came to making substantial progress in their research work which would help them solve pressing issues in the healthcare system.
I have also listened to students who have ideas that can greatly assist in solving these problems but these limitations hold the students back.
However, if they are presented with a tool like Ersilia that makes pre-trained machine learning models available to assist in drug discovery and development of new meficines, more than half of these scientists' problems would be solved.
This is why I'm interested in Ersilia. I would love to be part of the initiative that will take research in the healthcare system in low and middle income countries especially, to a new level of making discoveries with ease.
Initial contributions:
I corrected a typo at Line 47 in the README file https://github.com/ersilia-os/ersilia/pull/18#issue-1181659850
I created an issue template for third-party model incorporation to Ersilia's Model to be suggested by contributors. https://github.com/ersilia-os/ersilia/issues/23
I created an issue to include a Windows Installation guide on Ersilia's Model Hub page. https://github.com/ersilia-os/ersilia/issues/26
I created a pull request where I listed the steps to install Ersilia on Windows operating system through Windows Subsystem for Linux or through a Virtual Machine. I also described the problems I faced while trying to install Ersilia on my system in the pull request description. This pull request solves the issue I raised on including a Windows Installation guide on Ersilia's Model Hub page. https://github.com/ersilia-os/ersilia/pull/67
Hi @ifeoluwafavour ,
Great work so far. I will incorporate the template for model request once the discussion is closed. For the Windows installation, I am only missing a couple of sentences explaining a bit what a virtual machine is. For example, users might get confused that they have to install python or conda again even if they have it in their computers already. Can you add one line to explain that a virtual machine is like a separate computer?
Next you can continue working on any of the tasks listed above!
Hi @ifeoluwafavour ,
Great work so far. I will incorporate the template for model request once the discussion is closed. For the Windows installation, I am only missing a couple of sentences explaining a bit what a virtual machine is. For example, users might get confused that they have to install python or conda again even if they have it in their computers already. Can you add one line to explain that a virtual machine is like a separate computer?
Next you can continue working on any of the tasks listed above!
Okay. Thank you very much for your feedback. I will go ahead to start making changes now and also start working on a task.
Hi @ifeoluwafavour ,
Great work so far. I will incorporate the template for model request once the discussion is closed. For the Windows installation, I am only missing a couple of sentences explaining a bit what a virtual machine is. For example, users might get confused that they have to install python or conda again even if they have it in their computers already. Can you add one line to explain that a virtual machine is like a separate computer?
Next you can continue working on any of the tasks listed above!
I have created a pull request where I updated the installation steps file explaining why users still need to download Conda, Python and Git for Linux.
Thanks @ifeoluwafavour I have merged your changes! I think the WSL installation is much clearer now. I will let @victorabba finish his work, then we can work to add your work to the official installation guide, I will let you know next steps in this regard. Meanwhile, please check the list of tasks and select one that you find interesting to continue the contribution!
Thank you very much @GemmaTuron I'm currently working on one of the tasks.
Docstring for the ErsiliaModel class
class ErsiliaModel(model_identifier):
" " "The ErsiliaModel class fetches a model from the Ersilia Model hub.
Attributes:
model_identifier(str): Provide model identifier in eos0abc format.
" " "
def serve():
' ' 'The serve function provides the URL, Process ID and scl of model session.' ' '
def predict(molecule_name):
" " "
Arg:
molecule_name(str): Provide molecule name in SMILE format or in a .csv file
Returns:
predictions: Outcomes could be Antibiotic activity, Target, Vector, Physiochemical property, ADME, CYP450, Toxicity.
" " "
def close():
' ' 'The close function ends a model session.' ' '
Hi @ifeoluwafavour !
I am taggin @miquelduranfrigola for feedback on the docstrings, meanwhile continue working on some other task !
Hi @ifeoluwafavour !
I am taggin @miquelduranfrigola for feedback on the docstrings, meanwhile continue working on some other task !
Okay. Thank you for your feedback. I will move on to another task
I have written a summary of EOSI's strategic plan in the form of a blog post.
Here's the link to the document: https://docs.google.com/document/d/18sPRsVA7GmRMQdR5DELKpcy6Jn1pmqvfF_7_s2ZNZeo/edit?usp=drivesdk
Hi @GemmaTuron I didn't see the need to rewrite everything. I just highlighted the main points from the main strategic plan page.
Hey @ifeoluwafavour you need to grant an access to 3rd party users to your blog post.
Here is what it says
Hey @ifeoluwafavour you need to grant an access to 3rd party users to your blog post. Here is what it says
Oh I thought I was to give access to the mentor only.
I'll change it to be publicly accessible now. Is this okay with you @GemmaTuron ?
That's why It's an open-source project Every thing should be transparent
On Fri, Apr 8, 2022, 10:02 AM Ife @.***> wrote:
Hey @ifeoluwafavour https://github.com/ifeoluwafavour you need to grant an access to 3rd party users to your blog post. Here is what it says [image: ife1] https://user-images.githubusercontent.com/18760267/162338591-4b2b4323-b3ba-46e6-8636-e03ef87f47e4.PNG
Oh I thought I was to give access to the mentor only.
I'll change it to be publicly accessible now.
— Reply to this email directly, view it on GitHub https://github.com/ersilia-os/ersilia/issues/97#issuecomment-1092625012, or unsubscribe https://github.com/notifications/unsubscribe-auth/AEPEES2FK6U42TSDGQKOFW3VD7Y23ANCNFSM5SNEYLSQ . You are receiving this because you commented.Message ID: @.***>
It's publicly accessible now @dchidindu5
Twitter Template for The Release of New Models
We have a new model! 💃🕺
Ever wondered how a molecule moves? 🤔
With
Learn more:
The code is of course open sourced:
Source paper:
Work by: <contributors' or developers' names>
Ersilia Open Source Initiative welcomes model contributions.
You can check out our README in our repository for more information to get started as a contributor.
Hey @ifeoluwafavour you need to grant an access to 3rd party users to your blog post. Here is what it says
Oh I thought I was to give access to the mentor only.
I'll change it to be publicly accessible now. Is this okay with you @GemmaTuron ?
Hello @ifeoluwafavour that's a nice write-up, but don't you think you can make it more catchy by adding a suitable title?
Hi @ifeoluwafavour,
Thanks for making the blogpost public. You can choose to make it readable only to the mentor, but of course we encourage to publicly share the information. I have added a few comments in it, specially I think you should focus in summarizing the lists into one or two sentences. You do not need to have all the information, is is a summary, but it makes it nicer for readers if blogposts do not have too many bullet points.
thanks for the comments @victorabba and @dchidindu5. Giving it a title as Victor suggests would be nice as well! Hope you like my comments and if you need further clarification tag me!
Twitter Template for The Release of New Models
Introducing
, A model that predicts activity in molecules with 70% accuracy. (Brief information about the model) Learn more: *picture/video of the model being used 1/3
The code is of course open sourced:
Source paper: online article or publication that inspired the development of the model.
Work by: <contributors' or developers' names> 2/3
Ersilia Open Source Initiative welcomes model contributions.
You can check out our README in our repository for more information to get started as a contributor.
3/3
Thanks for starting off in this task! Good first ideas, as a general suggestion, think that people scrolls through twitter and only half reads things, so it has to be very catchy. let's try to use a very catchy title, Introducing XXX model might not catch the attention of people scrolling through twitter. Something more like "We have a new model...."
I would not mention accuracy or other numbers in a twitter post, many people may not understand to what are we referring to, or videos as they wont have time to check them.
Refering to authors etc is good!
Hi @ifeoluwafavour,
Thanks for making the blogpost public. You can choose to make it readable only to the mentor, but of course we encourage to publicly share the information. I have added a few comments in it, specially I think you should focus in summarizing the lists into one or two sentences. You do not need to have all the information, is is a summary, but it makes it nicer for readers if blogposts do not have too many bullet points.
thanks for the comments @victorabba and @dchidindu5. Giving it a title as Victor suggests would be nice as well! Hope you like my comments and if you need further clarification tag me!
Yes I've seen your comments and I like them too 😄 I will make the changes as soon as possible. I will also go ahead to add the title and make the blogpost more engaging.
Twitter Template for The Release of New Models
Introducing
, A model that predicts activity in molecules with 70% accuracy. (Brief information about the model) Learn more: *picture/video of the model being used 1/3
The code is of course open sourced:
Source paper: online article or publication that inspired the development of the model.
Work by: <contributors' or developers' names> 2/3
Ersilia Open Source Initiative welcomes model contributions.
You can check out our README in our repository for more information to get started as a contributor.
3/3
Thanks for starting off in this task! Good first ideas, as a general suggestion, think that people scrolls through twitter and only half reads things, so it has to be very catchy. let's try to use a very catchy title, Introducing XXX model might not catch the attention of people scrolling through twitter. Something more like "We have a new model...."
I would not mention accuracy or other numbers in a twitter post, many people may not understand to what are we referring to, or videos as they wont have time to check them.
Refering to authors etc is good!
Alright! Noted. I'll make everything more catchy and make the other changes. Thank you for your feedback.
Hi @ifeoluwafavour . Thanks for making your work readable. Great job!
Hi @ifeoluwafavour, I saw you ticked the community task. Please could you help me with some clarifications? When they say we should look at other projects, do they mean other Erisilia's projects on outreachy ? or other ersilia's projects in general on their webpage?
Hi @ifeoluwafavour, I saw you ticked the community task. Please could you help me with some clarifications? When they say we should look at other projects, do they mean other Erisilia's projects on outreachy ? or other ersilia's projects in general on their webpage?
It means you have to check other contributors' projects. Just as you've commented on mine
Oh ok! Great !!! I was thinking far😅.
Thank youuuu. Let me catch some sleep now.
@ifeoluwafavour great work.
Oh ok! Great !!! I was thinking far😅.
Thank youuuu. Let me catch some sleep now.
😂😂 okay.
@ifeoluwafavour great work.
Thank you 😁
Twitter Template for The Release of New Models
We have a new model! 💃🕺
Ever wondered how a molecule moves? 🤔
With
you can see how the molecule of interest moves and also, with what it moves! 😃 Learn more:
*picture/video of the model being used
The code is of course open sourced:
Source paper:
Work by: <contributors' or developers' names>
Ersilia Open Source Initiative welcomes model contributions.
You can check out our README in our repository for more information to get started as a contributor.
Hi @GemmaTuron, I have edited the Twitter post template
I have written a summary of EOSI's strategic plan in the form of a blog post.
Here's the link to the document: https://docs.google.com/document/d/18sPRsVA7GmRMQdR5DELKpcy6Jn1pmqvfF_7_s2ZNZeo/edit?usp=drivesdk
@GemmaTuron I also incorporated the corrections you made in the blog post
Great work @ifeoluwafavour
Twitter Template for The Release of New Models We have a new model! 💃🕺 Ever wondered how a molecule moves? 🤔 With
you can see how the molecule of interest moves and also, with what it moves! 😃 Learn more: *picture/video of the model being used The code is of course open sourced: Source paper: Work by: <contributors' or developers' names>
Ersilia Open Source Initiative welcomes model contributions. You can check out our README in our repository for more information to get started as a contributor.
Hi @GemmaTuron, I have edited the Twitter post template
Great work @ifeoluwafavour , absolutely love it.
Thank you so much @Jaya3112 @loweyvana
I wrote a blog post on artificial intelligence in biomedical research. Here's the link. https://docs.google.com/document/d/13jQqMVx85TuVAiE00flmGymiC3_pUdJSF_WMdF3fu08/edit?usp=drivesdk
Hi @ifeoluwafavour
Thanks for adding our comments and being so active in the community in both GitHub and Slack, your work is really appreciated. The AI post is well thought, good that you introduce the risk of bias in AI, is a topic we are really interested in. The installation guidelines for Ersilia in Windows are great, we will incorporate them in our guide. I will work with @victorabba to make a section on virtual machines.
I think you are ready to work towards your final application! Perhaps one last task would be to choose a model and suggest a technical card for it, if you want to work further on the project
Okay! 😃
Thank you very much for your kind words @GemmaTuron
I will start working on the technical card.
MODEL TECHNICAL CARD
IDENTIFICATION Model ID: eos1vms Model name: chembl-multitask-descriptor Model description: the model predicts the main target of a small molecule based on ChEMBL data.
GENERAL INFORMATION ML algorithm used to train model: Naive Bayes (specifically the Multinomial Naive Bayes algorithm). Data used to train model: a subset of CHEMBL_18 data containing pairs of ligand compounds and single-protein targets. Input: compound Output: target and vector
RESOURCES Repository: https://github.com/ersilia-os/eos1vms Source code: https://github.com/chembl/target_predictions Publication: https://chembl.github.io/ligand-based-target-predictions-in/ License: Apache
MODEL TECHNICAL CARD
IDENTIFICATION Model ID: eos1vms Model name: chembl-multitask-descriptor Model description: the model predicts the main target of a small molecule based on ChEMBL data.
GENERAL INFORMATION ML algorithm used to train model: Naive Bayes (specifically the Multinomial Naive Bayes algorithm). Data used to train model: a subset of CHEMBL_18 data containing pairs of ligand compounds and single-protein targets. Input: compound Output: target and vector
RESOURCES Repository: https://github.com/ersilia-os/eos1vms Source code: https://github.com/chembl/target_predictions Publication: https://chembl.github.io/ligand-based-target-predictions-in/ License: Apache
Hi @GemmaTuron
I have written the technical card for a model.
MODEL TECHNICAL CARD
IDENTIFICATION Model ID: eos1vms Model name: chembl-multitask-descriptor Model description: the model predicts the main target of a small molecule based on ChEMBL data.
GENERAL INFORMATION ML algorithm used to train model: Naive Bayes (specifically the Multinomial Naive Bayes algorithm). Data used to train model: a subset of CHEMBL_18 data containing pairs of ligand compounds and single-protein targets. Input: compound Output: target and vector
RESOURCES Repository: https://github.com/ersilia-os/eos1vms Source code: https://github.com/chembl/target_predictions Publication: https://chembl.github.io/ligand-based-target-predictions-in/ License: Apache
Hello @ifeoluwafavour. Nice work
MODEL TECHNICAL CARD
IDENTIFICATION Model ID: eos1vms Model name: chembl-multitask-descriptor Model description: the model predicts the main target of a small molecule based on ChEMBL data.
GENERAL INFORMATION ML algorithm used to train model: Naive Bayes (specifically the Multinomial Naive Bayes algorithm). Data used to train model: a subset of CHEMBL_18 data containing pairs of ligand compounds and single-protein targets. Input: compound Output: target and vector
RESOURCES Repository: https://github.com/ersilia-os/eos1vms Source code: https://github.com/chembl/target_predictions Publication: https://chembl.github.io/ligand-based-target-predictions-in/ License: Apache
Hello @ifeoluwafavour. Nice work
Thank you Elizabeth 😊
MODEL TECHNICAL CARD
IDENTIFICATION Model ID: eos1vms Model name: chembl-multitask-descriptor Model description: the model predicts the main target of a small molecule based on ChEMBL data.
GENERAL INFORMATION ML algorithm used to train model: Naive Bayes (specifically the Multinomial Naive Bayes algorithm). Data used to train model: a subset of CHEMBL_18 data containing pairs of ligand compounds and single-protein targets. Input: compound Output: target and vector
RESOURCES Repository: https://github.com/ersilia-os/eos1vms Source code: https://github.com/chembl/target_predictions Publication: https://chembl.github.io/ligand-based-target-predictions-in/ License: Apache
Great work @ifeoluwafavour
MODEL TECHNICAL CARD
IDENTIFICATION Model ID: eos1vms Model name: chembl-multitask-descriptor Model description: the model predicts the main target of a small molecule based on ChEMBL data.
GENERAL INFORMATION ML algorithm used to train model: Naive Bayes (specifically the Multinomial Naive Bayes algorithm). Data used to train model: a subset of CHEMBL_18 data containing pairs of ligand compounds and single-protein targets. Input: compound Output: target and vector
RESOURCES Repository: https://github.com/ersilia-os/eos1vms Source code: https://github.com/chembl/target_predictions Publication: https://chembl.github.io/ligand-based-target-predictions-in/ License: Apache
Great work @ifeoluwafavour
Thank you @julietugo ☺
Hi @ifeoluwafavour If only, the card is missing the number of molecules in the dataset (sometimes this is not available so don't worry). The rest is good, thanks!
Okay Thank you very much for everything @GemmaTuron Should I close the issue now?
Applicant: @ifeoluwafavour
Welcome to the Ersilia Open Source Initiative. This issue will serve to track all your contributions for the project “Improve the documentation and outreach material of the Ersilia Model Hub”.
Please tick the tasks as you complete them. To make a final application it is not required to have completed all tasks. Only the Initial Steps and Community sections are REQUIRED. The tasks are not ordered from more to less important, they are simply related to different skills. Start where you feel most comfortable. This project can be adapted to the applicants interests, please focus on the type of tasks that you prefer / have better skills / would like to work on as an intern.
Initial steps:
[x] Comment under this issue explaining why are you interested in this project
GitHub documentation:
[ ] Incorporate feedback from the mentor
Writing dissemination material
[ ] Create a template short Newsletter (1 paragraph) to send every month to our community (funders, users, contributors). It should mention metrics (models in the hub, number of users, funding…), thank you etc
Technical skills (required for the tutorial only)
[x] Write a docstring for the ErsiliaModel class. Use the Google Python Style guide. Paste the docstring as a comment below (do not use a PR).
Graphic material
[ ] Incorporate feedback from the mentor
Scientific content
[ ] Search the scientific literature and suggest 3 new models (comment in this issue) that would be relevant to incorporate in the Hub.
Community
[x] If you have feedback from your peers, answer it in this issue.
Other
If you have interest in working on related topics, or have new suggestions, please do the following
[ ] Link in the comments any other PR you have contributed to.
Final application