Back from the archived section: Algorithmic Phenotype Working Group

General
1

Hello All,

As the title specifies, we are bringing back the old Phenotyping WG as a more focused Algorithmic Phenotype WG. The main focus of this WG will be to align efforts on probabilistic phenotyping work done within the OHDSI community. This group is separate, but works in parallel to the Gold Standard Phenotype library WG led by @apotvien

Everybody is welcomed to join, so let’s see a raise of hands to whomever wants to be part of this WG and is interested in attending the monthly calls. Once we have some quorum, we can start planning the class based on people’s schedules.

Don’t be shy, and join the probabilistic side :slight_smile:

2 Likes
2

Hi - I’m relatively new to this topic - can you provide an overview regarding what you mean by “Algorithmic” - does it include various methods and approaches to predicting if a patient belongs to a phenotype? It would help me understand if I can contribute.

Thanks,
MK

3

Thank you. I am interested, Sir :slight_smile:

1 Like
4

Hi,
Thanks for the initiative. I’m relatively new to the OHDSI.
I’m interested to join “the probabilistic side”.

But I am also interested to know more about the “probabilistic phenotype” - is it a vector for each patient that assigns normalized probabilities for each patient to have one particular type of phenotype ?

Thank you !
–Subhajit

5

Count me in please.

6

Sorry for my late response, I had forum notifications turned off. Its using a ML model to define phenotypes for patients, assigning them a probability for a given phenotype. This differs from the traditional rule-based definitions, in which a person has or does not have (binary outcome) a given phenotype.

7

Hello,

Yes, exactly, it is to use ML models that assign probabilities rather than rules that give you a yes or no membership. More details can be found here: https://www.annualreviews.org/doi/10.1146/annurev-biodatasci-080917-013315

8

Since the year is winding down and many of us will be gone for holidays. I will start reaching out in January to kick off the meetings. Thanks for all of you that have shown interest.

10

My apologies for the long delay, stuff has been moving forward in the background but I wanted to have a few things align before fully kicking this off. I will be giving a quick update of this WG on the OHDSI call today, and here is the Doodle poll link for scheduling the first meeting:

https://doodle.com/poll/6d5cyz72enf9arh8

I am shouting out to the people that responded to the previous messages to sign up. @Andrew @rohitv @subhajit06 @mkwong. However, everybody is welcome to join!

11

Thanks for your responses! I will keep this open a few more days to get some other people signed up.

12

I am interested. Thanks

13

Greetings,

Based on the Doodle responses, we have 5 people available for March 27th, Friday. I am sure that availability has probably changed since then. But I will leave it open for anybody wanting a break from the study-a-thon to join in for an hour to the call:

Topic: Algorithmic Phenotype Working Group Kick-off Meeting
Time: Mar 27, 2020 10:00 AM Eastern Time (US and Canada)

Join Zoom Meeting

Meeting ID: 523 482 680

One tap mobile
+13126266799,523482680# US (Chicago)
+19294362866,523482680# US (New York)

Dial by your location
+1 312 626 6799 US (Chicago)
+1 929 436 2866 US (New York)
+1 253 215 8782 US
+1 301 715 8592 US
+1 346 248 7799 US (Houston)
+1 669 900 6833 US (San Jose)
Meeting ID: 523 482 680
Find your local number: Zoom International Dial-in Numbers - Zoom

14

hey, I just saw this, i’ll join the meeting, if that is fine

15

Hello @Juan_Banda ,

Is the meeting on?? It’s now right?

16

Hey @Juan_Banda, just following up on this thread. Is this WG active?

17

Hello, yes. I was waiting for the COVID-19 research activities to slow down a bit to schedule the next call. Will make a post about this on Sunday.