{ "topic": { "id": 401532, "title": "Welcome to the CAFA Challenge", "authorName": "Iddo Friedberg", "commentCount": 17, "votes": 31, "postDate": "2023-04-13T17:25:22.093000" }, "comments": [ { "id": 2225958, "authorName": "Predrag Radivojac", "votes": 11, "postDate": "2023-04-18T14:49:20.403000", "content": "

Hello, I am Pedja. I am a Professor of Computer Science at Northeastern. We are interested in machine learning, computational biology and precision health -- we develop methods for protein function prediction, for understanding the impact of genomic variation, for predicting risk of complex phenotypes, etc. and we often work with data in a biased semi-supervised setting. We are also very interested in evaluation of machine learning algorithms, which can be complex in some cases. Iddo and I jointly started CAFA more than a decade ago. I look forward to seeing if the Kaggle community can move the needle on this important problem. Good luck everyone!

" }, { "id": 2224716, "authorName": "Iddo Friedberg", "votes": 10, "postDate": "2023-04-17T15:33:01.637000", "content": "

I'm Iddo Friedberg and I am an Associate Professor at the department of Veterinary Microbiology & Preventive Medicine at Iowa State University. I am interested in large scale analyses of protein sequence, structure and function connections, genome evolution, metagenome analysis and computational protein function prediction. I have been co-leading CAFA since 2010 with Pedja, and I am excited to team up with Kaggle and welcoming Kagglers to this challenge.

\n

In my spare time (ha!) I enjoy long distance running, 1970’s cop shows, spaghetti westerns, and embarrassing my kids.

" }, { "id": 2226097, "authorName": "Parnal Joshi", "votes": 7, "postDate": "2023-04-18T16:37:20.887000", "content": "

Hi, I'm Parnal. I'm a PhD candidate in the Bioinformatics and Computational Biology program at Iowa State University. I am working as a Graduate Research Assistant with Iddo Friedberg at the Department of Veterinary Microbiology and Preventive Medicine. My research interests include investigating biases in protein function annotation databases, analyzing protein structures, and developing bioinformatics tools and pipelines. I am currently working on applying ontologies and text-mining to identify biases in scientific publications.

" }, { "id": 2225579, "authorName": "Damiano Piovesan", "votes": 7, "postDate": "2023-04-18T09:11:13.550000", "content": "

My name is Damiano Piovesan, and I am currently Associate Professor of Bioinformatics at the Department of Biomedical Sciences at the University of Padova. As a former CAFA developer, I have recently transitioned into a more organizational role, where I now maintain the CAFA evaluation software used in Kaggle. My research interests lie in the function and biophysics of Intrinsically Disordered Proteins (IDPs), and I am proud to serve as co-leader of both the DisProt and MobiDB IDP databases. Additionally, I have been co-organizing the Critical Assessment of protein Intrinsic Disorder (CAID) since 2018, and I am co-lead of the ELIXIR IDP Community since 2020.
\nI enjoy playing chess, particularly blitz games, and I'm not much of a swimmer. I'm a pizza lover, who isn't?!

" }, { "id": 2225968, "authorName": "Clara De Paolis", "votes": 8, "postDate": "2023-04-18T14:58:55.687000", "content": "

Hi, I’m Clara, a PhD candidate at Northeastern University's Khoury College of Computer Science working with Professor Predrag Radivojac. My research interests are in data bias in machine learning; structured prediction and evaluation; geometric deep learning; and representation and generation of structured data. I am currently working on methods for biological sequence (proteins & RNA) representation and also on detection and quantification of bias in datasets.

" }, { "id": 2228858, "authorName": "MT", "votes": 3, "postDate": "2023-04-20T21:52:27.277000", "content": "

Hi guys, can you please check this id in test A0A1D6E0S8? It appears two times. Thanks!

\n

ID: A0A1D6E0S8
\nSeq('MPSRSPACRPRGRNRRSAADAVARPLALALILVSTLPRAAHSQDLALPPVQPRG…SFC')

\n

ID: A0A1D6E0S8
\nSeq('MPSRSPACRPRGRNRRSAADAVARPLALALILVSTLPRAAHSQDLALPPVQPRG…SFC')

" }, { "id": 2382718, "authorName": "Fernando Rodrigues", "votes": 2, "postDate": "2023-08-10T01:24:58.343000", "content": "

\"Hello Kagglers… This is the second Kaggle competition I'm participating in, and with great enthusiasm, I've just completed my submission. I'm eager to see how this competition unfolds and to continue my learning journey here on Kaggle.\"

" }, { "id": 2266320, "authorName": "Tilii", "votes": 2, "postDate": "2023-05-19T22:43:18.900000", "content": "

@iddux1 @predragradivojac

\n

I have already sent this through CAFA contact website, but there was no response.

\n

On CAFA website it says \"CAFA 5 is not open for registration and submission on Kaggle\" while I think the intent was to say NOW instead of not.

\n

Could you re-upload the original CAFA4 files? For example, this file is missing:

\n

https://biofunctionprediction.org/cafa-targets/cafa4ontologies.zip

" }, { "id": 2281432, "authorName": "Tilii", "votes": 2, "postDate": "2023-05-30T22:04:13.980000", "content": "

@iddux1 @predragradivojac Could I please get an answer here? It appears that \"CAFA 5 is not open for registration and submission on Kaggle\" has been fixed, so I assume someone read my post or received the message.

" }, { "id": 2227200, "authorName": "Julie Fan", "votes": 0, "postDate": "2023-04-19T15:09:29.760000", "content": "

I am Julie and I'd like to get practiced. Thanks.

" }, { "id": 2389679, "authorName": "SHU_Mos_Z", "votes": 0, "postDate": "2023-08-14T07:28:38.577000", "content": "

I am Mos_Z and I'd like to get practiced. Thanks.

" }, { "id": 2250759, "authorName": "elisacarli", "votes": 0, "postDate": "2023-05-08T18:59:34.587000", "content": "

Hello, I've tried to implement T5 embedding, but when I run it, I get the following error: RuntimeError during embedding for Q93ZJ3 (L=199). Try lowering batch size. If single sequence processing does not work, you need more vRAM to process your protein.
\nRuntimeError during embedding for P05745 (L=199). Try lowering batch size. If single sequence processing does not work, you need more vRAM to process your protein.

\n

batch size = 5

" }, { "id": 2281774, "authorName": "Abdullah", "votes": 0, "postDate": "2023-05-31T05:49:02.710000", "content": "

you can directly use this from here /kaggle/input/t5embeds.
\nWhich is already uploaded by @sergeifironov .

" }, { "id": 2229417, "authorName": "Ankit_Jangid9", "votes": 0, "postDate": "2023-04-21T10:54:04.040000", "content": "

I don't have biology background , can anybody explain me what parameter We are try to predict ?
\nit is \"GO:0008152\" from protein sequence or \"nucleobase-containing compound catabolic process\"

" }, { "id": 2281777, "authorName": "Abdullah", "votes": 0, "postDate": "2023-05-31T05:51:14.400000", "content": "

To process the sequences for training, you can start from pre trained embeddings from famous languages models trained on protein sequences (ProtBERT, EMS2, T5). You can also represent GO terms in each proteins as a sparse matrix and do clustering methods (KNN).

" }, { "id": 2264964, "authorName": "", "votes": 0, "postDate": "2023-05-18T22:23:20.363000", "content": "" }, { "id": 2232383, "authorName": "", "votes": 0, "postDate": "2023-04-24T09:45:52.277000", "content": "" } ], "index": { "id": "401532", "title": "Welcome to the CAFA Challenge", "authorName": "", "commentCount": "17", "votes": "31", "postDate": "2023-04-13 17:25:22.093000" }, "competition": "cafa-5-protein-function-prediction" }