JADBio on Twitter: "Welcome to this month's #Twitterchat!😄Today we're discussing Machine Learning for Cancer! Our first question is coming up shortly!👉Track #AutoML to join the conversation!#Twitterchat #LiveTwitterChat #MachineLearning #Cancer pic.twitter.com/GKLguXO3Vf / Twitter"

Welcome to this month's #Twitterchat!😄Today we're discussing Machine Learning for Cancer! Our first question is coming up shortly!👉Track #AutoML to join the conversation!#Twitterchat #LiveTwitterChat #MachineLearning #Cancer pic.twitter.com/GKLguXO3Vf

Below are each of the questions, along with some of our favorite answers for you to check out, but if you want to see the whole chat, head over to Twitter and follow the #AutoML hashtag!

JADBio on Twitter: "Q1: Is Machine Learning relevant in Cancer research?#AutoML pic.twitter.com/dwtWUbVoqI / Twitter"

Q1: Is Machine Learning relevant in Cancer research?#AutoML pic.twitter.com/dwtWUbVoqI

Vicki 🤖² on Twitter: "In analyzing complex datasets in #cancer biology, we can use machine learning, to look for correlations, biomarkers, genes etc. The advantage of ML is it can do these correlations, faster than humans, and analyze extremely large and complex data sets. #AutoML / Twitter"

In analyzing complex datasets in #cancer biology, we can use machine learning, to look for correlations, biomarkers, genes etc. The advantage of ML is it can do these correlations, faster than humans, and analyze extremely large and complex data sets. #AutoML

Aris Karanikas on Twitter: "It is indispensable! As will all types of bio research, ML can help make sense of the data. Anything that can shorten the time to market for a new drug from the 12-15 years it stands currently, is a benefit not only to pharma and biotech cos but society as a whole! / Twitter"

It is indispensable! As will all types of bio research, ML can help make sense of the data. Anything that can shorten the time to market for a new drug from the 12-15 years it stands currently, is a benefit not only to pharma and biotech cos but society as a whole!

Ioannis Tsamardinos on Twitter: "Definitely! There are so much cancer data that it is impossible for humans alone to make sense of it. Just The Cancer Genome Atlas TCGA, the largest oncological project in the USA, has multi-omics molecular profiles for tens of thousands of cancerous tumors. #automl / Twitter"

Definitely! There are so much cancer data that it is impossible for humans alone to make sense of it. Just The Cancer Genome Atlas TCGA, the largest oncological project in the USA, has multi-omics molecular profiles for tens of thousands of cancerous tumors. #automl

JADBio on Twitter: "Q2: What is the impact of Μachine Learning in Cancer diagnosis and treatment?#AutoML#TwitterChat #LiveTwitterchat pic.twitter.com/igPomLAiPH / Twitter"

Q2: What is the impact of Μachine Learning in Cancer diagnosis and treatment?#AutoML#TwitterChat #LiveTwitterchat pic.twitter.com/igPomLAiPH

Ioannis Tsamardinos on Twitter: "Several groups, including ours, have been creating personalized prediction models for treatment therapy, to optimize medical decisions. I believe in the near future a lot of them will find their way to clinical practice. #automl / Twitter"

Several groups, including ours, have been creating personalized prediction models for treatment therapy, to optimize medical decisions. I believe in the near future a lot of them will find their way to clinical practice. #automl

Ioannis Tsamardinos on Twitter: "There are still many subtypes of cancer we haven't identified. ML is helping understanding all the different types of cancer beyond what is easily identifiable phenotypically or histopathologically. #automl / Twitter"

There are still many subtypes of cancer we haven't identified. ML is helping understanding all the different types of cancer beyond what is easily identifiable phenotypically or histopathologically. #automl

Vicki 🤖² on Twitter: "#Cancer is extremely complicated: Abnormal masses, unpredictable appearance, and shape, etc. That makes cancer diagnosis difficult and obviously treatment even more complicated. But we are collecting massive amounts of data. Only way to find patterns is #ML. #AutoML / Twitter"

Cancer is extremely complicated: Abnormal masses, unpredictable appearance, and shape, etc. That makes cancer diagnosis difficult and obviously treatment even more complicated. But we are collecting massive amounts of data. Only way to find patterns is #ML. #AutoML

JADBio on Twitter: "Q3: Can Machine Learning accelerate knowledge discovery in Cancer diagnosis or treatment?🧐#AutoML#Twitterchat #LiveTwitterchat pic.twitter.com/pcU1IBiruO / Twitter"

Q3: Can Machine Learning accelerate knowledge discovery in Cancer diagnosis or treatment?🧐#AutoML#Twitterchat #LiveTwitterchat pic.twitter.com/pcU1IBiruO

Vicki 🤖² on Twitter: "Prevention, early diagnosis, treatment and cancer management, innovation and collaboration sit at the heart of accelerating oncology research. #AutoML / Twitter"

Prevention, early diagnosis, treatment and cancer management, innovation and collaboration sit at the heart of accelerating oncology research. #AutoML

Vicki 🤖² on Twitter: "Knowledge discovery is finding information at every step of these processes and contributing to the war against #cancer with something new. #autoML / Twitter"

Knowledge discovery is finding information at every step of these processes and contributing to the war against #cancer with something new. #autoML

Vicki 🤖² on Twitter: "That's why IMO #AutoML is a tool that can help analyze #data gathered at every step of this journey, faster and by everyone, regardless of coding expertise. Where there's data there's knowledge to be found. It can help build treatments, new drugs, etc. #AutoML / Twitter"

That's why IMO #AutoML is a tool that can help analyze #data gathered at every step of this journey, faster and by everyone, regardless of coding expertise. Where there's data there's knowledge to be found. It can help build treatments, new drugs, etc. #AutoML

Pavlos Charonyktakis on Twitter: "ML enables knowledge discovery (sth that most of the people do not know due to the hype of DNNs). Other than that ML , in all cases, makes things faster than doing them manually #automl / Twitter"

ML enables knowledge discovery (sth that most of the people do not know due to the hype of DNNs). Other than that ML , in all cases, makes things faster than doing them manually #automl

JADBio on Twitter: "Q4: Do you have Case Studies that are using Μachine Learning and making a difference in Cancer research?#AutoML#TwitterChat #LiveTwitterChat pic.twitter.com/ho49Z6kz2m / Twitter"

Q4: Do you have Case Studies that are using Μachine Learning and making a difference in Cancer research?#AutoML#TwitterChat #LiveTwitterChat pic.twitter.com/ho49Z6kz2m

Vicki 🤖² on Twitter: "Here u go: Multi-omics characterization of left-right colorectal cancer. https://t.co/kytik7H37Z #cancer #colorectal #ML #AutoML / Twitter"

Here u go: Multi-omics characterization of left-right colorectal cancer. https://t.co/kytik7H37Z #cancer #colorectal #ML #AutoML

Vicki 🤖² on Twitter: "or Deciphering the Methylation Landscape in Breast Cancer https://t.co/moqLlGrQJl #autoML / Twitter"

or Deciphering the Methylation Landscape in Breast Cancer https://t.co/moqLlGrQJl #autoML

Ioannis Tsamardinos on Twitter: "Our joint work with the TCGA consortium has identified a set of biomarkers that diagnose the subtype of cancer for 26 different cancer types. I hope it will soon lead to a pancancer diagnostic assay to be soon used in clinical practice. #automl / Twitter"

Our joint work with the TCGA consortium has identified a set of biomarkers that diagnose the subtype of cancer for 26 different cancer types. I hope it will soon lead to a pancancer diagnostic assay to be soon used in clinical practice. #automl

Vicki 🤖² on Twitter: "That's promising. / Twitter"

That's promising.

JADBio on Twitter: "Thank you for joining our #MachineLearning for Cancer Twitter chat! 🥳Keep the conversations going, and let's meet again next month for another insightful #twitterchat! 🤓 #AutoML#TwitterChat #LiveTwitterChat pic.twitter.com/3f2813awdl / Twitter"

Thank you for joining our #MachineLearning for Cancer Twitter chat! 🥳Keep the conversations going, and let's meet again next month for another insightful #twitterchat! 🤓 #AutoML#TwitterChat #LiveTwitterChat pic.twitter.com/3f2813awdl

“Machine Learning for Cancer” #AutoML Twitter Chat Highlights from 14/4/2022 was originally published in AutoML for Life Sciences on Medium, where people are continuing the conversation by highlighting and responding to this story.

😀 Welcome to this month’s #twitterchat!
Today we’re discussing #AutoML for Life Sciences!
Our first question is coming up shortly!
Track #AutoML to join the conversation!

Below are each of the questions, along with some of our favorite answers for you to check out, but if you want to see the whole chat, head over to Twitter and follow the #AutoML hashtag!

JADBio on Twitter: "Q1: What is AutoML for Life Sciences? #AutoML / Twitter"

Q1: What is AutoML for Life Sciences? #AutoML

Aris Karanikas on Twitter: "Research in Life Sciences is a very lengthy and costly process. AutoML can help significantly reduce both, through knowledge discovery and quality predictive models #AutoML / Twitter"

Research in Life Sciences is a very lengthy and costly process. AutoML can help significantly reduce both, through knowledge discovery and quality predictive models #AutoML

Vicki 🤖² on Twitter: "It's #machinelearning for non-machinelearning experts, or coders... or for analysts who want to be more efficient#autoML / Twitter"

It's #machinelearning for non-machinelearning experts, or coders... or for analysts who want to be more efficient#autoML

Pavlos Charonyktakis on Twitter: "AutoML enables life scientists to extract knowledge from their data, without any statistics or coding knowledge #AutoML / Twitter"

AutoML enables life scientists to extract knowledge from their data, without any statistics or coding knowledge #AutoML

Popi Paraschaki on Twitter: "#ML for anyone who wants to speed up the analysis procedure and focus to knowledge and gather insights from his/her data. / Twitter"

ML for anyone who wants to speed up the analysis procedure and focus to knowledge and gather insights from his/her data.

JADBio on Twitter: "Please don't forget to use our hashtag so chat participants can see your tweet.😉 #AutoML / Twitter"

Please don't forget to use our hashtag so chat participants can see your tweet.😉 #AutoML

JADBio on Twitter: "Q2: Why AutoML and not just machine learning? #AutoML / Twitter"

Q2: Why AutoML and not just machine learning? #AutoML

Aris Karanikas on Twitter: "For a number of reasons; Machine Learning has its own challenges, it's not like you follow a recipe, and boom, you get cake! Especially if your data science skills are low, can help you avoid many of the pitfalls (can you say overfitting?) #AutoML / Twitter"

For a number of reasons; Machine Learning has its own challenges, it's not like you follow a recipe, and boom, you get cake! Especially if your data science skills are low, can help you avoid many of the pitfalls (can you say overfitting?) #AutoML

Vicki 🤖² on Twitter: "#autoML can automate the processes, at every stage, from dataset to predictive model. This means less time and cost. Considering the amount of available data that is not analyzed, due to lack of resources, automl can be a great solution for analysts or non-MLexperts / Twitter"

autoML can automate the processes, at every stage, from dataset to predictive model. This means less time and cost. Considering the amount of available data that is not analyzed, due to lack of resources, automl can be a great solution for analysts or non-MLexperts

Pavlos Charonyktakis on Twitter: "Ease and Value at Scale. Streamlining the ML #AutoML / Twitter"

Ease and Value at Scale. Streamlining the ML #AutoML

Popi Paraschaki on Twitter: "Not everyone knows how to #MachineLearning 😊 #AutoML / Twitter"

Not everyone knows how to #MachineLearning 😊 #AutoML

JADBio on Twitter: "Q3: What are the prerequisites for successful AutoML analysis? #AutoML / Twitter"

Q3: What are the prerequisites for successful AutoML analysis? #AutoML

Aris Karanikas on Twitter: "Stating the obvious here perhaps, but yes, data! And a talented researcher with a good idea of what they're solving for. #AutoML can run the analysis and find the correlations for you, but the researcher formulates the problem! / Twitter"

Stating the obvious here perhaps, but yes, data! And a talented researcher with a good idea of what they're solving for. #AutoML can run the analysis and find the correlations for you, but the researcher formulates the problem!

Tech Talks Central on Twitter: "#data! As for all real-world problems that require #ML predictions #AutoML / Twitter"

data! As for all real-world problems that require #ML predictions #AutoML

JADBio on Twitter: "Q4: Is there any published research that used AutoML for its results? #AutoML / Twitter"

Q4: Is there any published research that used AutoML for its results? #AutoML

Vicki 🤖² on Twitter: "Sure! Data-Driven Decision Support for #Autism Diagnosis using Machine Learning https://t.co/9gkCRsOBa8#AutoML / Twitter"

Sure! Data-Driven Decision Support for #Autism Diagnosis using Machine Learning https://t.co/9gkCRsOBa8#AutoML

Vicki 🤖² on Twitter: "or Predicting PD-1/PD-L1 inhibitors treatment on metastatic non-small cell lung #cancer https://t.co/DoJGx9KZCM #AutoML / Twitter"

or Predicting PD-1/PD-L1 inhibitors treatment on metastatic non-small cell lung #cancer https://t.co/DoJGx9KZCM #AutoML

Thank you for joining our #twitterchat about #AutoML & Life Sciences. 🤓
Keep the conversation going, and join us again next month for another enlightening one.
Let’s Meet and Chat 😃
#AutoML

“AutoML and Life Sciences” #AutoML Twitter Chat Highlights from 26/1/2022 was originally published in AutoML for Life Sciences on Medium, where people are continuing the conversation by highlighting and responding to this story.

Get ready to join us on our upcoming ‘Machine Learning for Life Sciences’ #TwitterChat, on Wednesday 23rd February, at 4:00 pm (GMT) | 8:00 am (PT) | 11:00 am (ET)!
Just track #AutoML to join the conversation!
You don’t want to miss this!😉

Below are each of the questions, along with some of our favorite answers for you to check out, but if you want to see the whole chat, head over to Twitter and follow the #AutoML hashtag!

JADBio on Twitter: "Q1:Why is ML important for Life Sciences?#Automl / Twitter"

Q1:Why is ML important for Life Sciences?#Automl

Aris Karanikas on Twitter: "The sheer amount of data that is generated on a daily basis in Life Sciences, plus its growing complexity makes ML a necessity. Traditional methods of extracting knowledge from such data can only go so far. / Twitter"

The sheer amount of data that is generated on a daily basis in Life Sciences, plus its growing complexity makes ML a necessity. Traditional methods of extracting knowledge from such data can only go so far.

Vicki 🤖² on Twitter: "Drug development, clinical trials, diagnostics, and supply chain are some of the areas that can make use of #ML. It has the ability to process and interpret large data sets, but most importantly it can lead to #knowledge discovery, which is crucial for #lifescience. #AutoML / Twitter"

Drug development, clinical trials, diagnostics, and supply chain are some of the areas that can make use of #ML. It has the ability to process and interpret large data sets, but most importantly it can lead to #knowledge discovery, which is crucial for #lifescience. #AutoML

Pavlos Charonyktakis on Twitter: "So much data is produced nowadays in Life Sciences. Through ML, new knowledge can be revealed. #AutoML can scale and make it easy. / Twitter"

So much data is produced nowadays in Life Sciences. Through ML, new knowledge can be revealed. #AutoML can scale and make it easy.

Ioannis Tsamardinos on Twitter: "Standard statistics (e.g. using p-values) examines each measured quantity and its correlation with some disease in isolation. ML considers the quantities in combination. This leads to much better predictions! #Automl / Twitter"

Standard statistics (e.g. using p-values) examines each measured quantity and its correlation with some disease in isolation. ML considers the quantities in combination. This leads to much better predictions! #Automl

Ioannis Tsamardinos on Twitter: "ML using feature selection can also minimize the number of features to predict or diagnose a disease. This leads to drug target identification and cost-effective diagnostics. It is essential for precision medicine. #Automl #PrecisionMedicine / Twitter"

ML using feature selection can also minimize the number of features to predict or diagnose a disease. This leads to drug target identification and cost-effective diagnostics. It is essential for precision medicine. #Automl #PrecisionMedicine

BioLizard on Twitter: "There are many subsectors in the life sciences where we apply ML: diagnosics, drug discovery, -omics data analysis, but also manufacturing, quality assurance and distribution! / Twitter"

There are many subsectors in the life sciences where we apply ML: diagnosics, drug discovery, -omics data analysis, but also manufacturing, quality assurance and distribution!

JADBio on Twitter: "Q2: Are Life Scientists using #MachineLearning & what is the adoption rate?#AutoML / Twitter"

Q2: Are Life Scientists using #MachineLearning & what is the adoption rate?#AutoML

Aris Karanikas on Twitter: "As with every innovation, simplicity drives adoption. ML is challenging, that's why researchers have had to rely on data scientists to get their data analyzed. Making it simple (see #AutoML), will help drive adoption within the researchers themselves, we're already seeing this. / Twitter"

As with every innovation, simplicity drives adoption. ML is challenging, that's why researchers have had to rely on data scientists to get their data analyzed. Making it simple (see #AutoML), will help drive adoption within the researchers themselves, we're already seeing this.

Vicki 🤖² on Twitter: "IMO the #lifesciences industry has only scratched the surface of the potential for #ML to shape its future for the better. Applications of AI in healthcare alone are expected to grow to more than $8 billion USD by 2022, globally. #AutoML #adoption / Twitter"

IMO the #lifesciences industry has only scratched the surface of the potential for #ML to shape its future for the better. Applications of AI in healthcare alone are expected to grow to more than $8 billion USD by 2022, globally. #AutoML #adoption

Ioannis Tsamardinos on Twitter: "no doubt about that ... we are just starting to tap the potential! We have to educate life scientists first and equip them with the right tools ... #Automl #adoption / Twitter"

no doubt about that ... we are just starting to tap the potential! We have to educate life scientists first and equip them with the right tools ... #Automl #adoption

Ioannis Tsamardinos on Twitter: "I think there is an exponential rate of adoption, yet adoption is still quite small and hasn't reached all its full potential. #Automl / Twitter"

I think there is an exponential rate of adoption, yet adoption is still quite small and hasn't reached all its full potential. #Automl

BioLizard on Twitter: "The potential of this #AI technology is definitely recognized by more and more life scientists. Despite this, setting up a solid and useful #ML pipeline that handles the specific features of a complex biological dataset remains challenging for many. / Twitter"

The potential of this #AI technology is definitely recognized by more and more life scientists. Despite this, setting up a solid and useful #ML pipeline that handles the specific features of a complex biological dataset remains challenging for many.

Digital Giraffes on Twitter: "Life/Data Scientists must understand MachineLearning for quality predictions & estimations. This can help machines to take the right decisions & smarter actions in real-time. Machine Learning is transforming how data mining and interpretation work. / Twitter"

Life/Data Scientists must understand MachineLearning for quality predictions & estimations. This can help machines to take the right decisions & smarter actions in real-time. Machine Learning is transforming how data mining and interpretation work.

JADBio on Twitter: "Don't forget to use our hashtag so chat participants can see your tweet.😉 cc @BioLizard_nv @SamDDecker #AutoML / Twitter"

Don't forget to use our hashtag so chat participants can see your tweet.😉 cc @BioLizard_nv @SamDDecker #AutoML

JADBio on Twitter: "Q3: Can you name any case study in which #ML was used to contribute to a life-sciences related problem?#AutoML / Twitter"

Q3: Can you name any case study in which #ML was used to contribute to a life-sciences related problem?#AutoML

Ioannis Tsamardinos on Twitter: "I have worked on dozens of such studies. One of my favorites is the analysis of thymoma cancer from multi-omics data where we find 2 markers that almost perfectly classify to 4 thymoma subtypes. #automl / Twitter"

I have worked on dozens of such studies. One of my favorites is the analysis of thymoma cancer from multi-omics data where we find 2 markers that almost perfectly classify to 4 thymoma subtypes. #automl

Vicki 🤖² on Twitter: "There are hundreds. One that I was skimming thru today: Liquid biopsy in type 2 #diabetes mellitus management: Building Specific Biosignatures via #MachineLearning Learning https://t.co/pelT1MislS #autoML / Twitter"

There are hundreds. One that I was skimming thru today: Liquid biopsy in type 2 #diabetes mellitus management: Building Specific Biosignatures via #MachineLearning Learning https://t.co/pelT1MislS #autoML

JADBio on Twitter: "Q4: Does biomedical or multi-omics data need a different type of machine learning treatment?#AutoML / Twitter"

Q4: Does biomedical or multi-omics data need a different type of machine learning treatment?#AutoML

Aris Karanikas on Twitter: "Absolutely! It's not every day you get a speadsheet with 2 million features for analysis (a colleague lovingly called this the 2Km spreadsheet!). Standard algorithms don't scale up that high to train a model, let alone extract knowledge (feature selection) from it! #AutoML / Twitter"

Absolutely! It's not every day you get a speadsheet with 2 million features for analysis (a colleague lovingly called this the 2Km spreadsheet!). Standard algorithms don't scale up that high to train a model, let alone extract knowledge (feature selection) from it! #AutoML

Ioannis Tsamardinos on Twitter: "Yes, they do! Preprocessing different omics types can be differentiated. But, just dealing with the high dimensionality of the multi-omics datasets requires different approaches altogether. Also, multi-omics datasets are often very low sample, which requires care #automl / Twitter"

Yes, they do! Preprocessing different omics types can be differentiated. But, just dealing with the high dimensionality of the multi-omics datasets requires different approaches altogether. Also, multi-omics datasets are often very low sample, which requires care #automl

BioLizard on Twitter: "Definitely: any given dataset requires a critical assessment of the best approach to apply ML! / Twitter"

Definitely: any given dataset requires a critical assessment of the best approach to apply ML!

Thank you for joining our #AutoML for Life Sciences #twitterchat! 😃
Keep the conversations going, and let’s meet again next month for another insightful #twitterchat! 🤓
#AutoML

<blockquote class=”twitter-tweet”><p lang=”en” dir=”ltr”>There are hundreds. One that I was skimming thru today: Liquid biopsy in type 2 <a href=”https://twitter.com/hashtag/diabetes?src=hash&amp;ref_src=twsrc%5Etfw">#diabetes</a> mellitus management: Building Specific Biosignatures via <a href=”https://twitter.com/hashtag/MachineLearning?src=hash&amp;ref_src=twsrc%5Etfw">#MachineLearning</a> Learning <a href=”https://t.co/pelT1MislS">https://t.co/pelT1MislS</a> <a href=”https://twitter.com/hashtag/autoML?src=hash&amp;ref_src=twsrc%5Etfw">#autoML</a></p>&mdash; Vicki 🤖² (@netWire) <a href=”https://twitter.com/netWire/status/1496520234781532167?ref_src=twsrc%5Etfw">February 23, 2022</a></blockquote> <script async src=”https://platform.twitter.com/widgets.js" charset=”utf-8"></script>

“Machine Learning for Life Sciences” #AutoML Twitter Chat Highlights from 23/2/2022 was originally published in AutoML for Life Sciences on Medium, where people are continuing the conversation by highlighting and responding to this story.

We’ve made some changes to our plans and added a Basic FREE option so that everyone can have access to JADBio’s user-friendly machine learning analysis platform, regardless if they’re working on a single research paper or running thousands of analyses and need a resource-intensive ML engine. There’s a machine learning option for everyone now. For the novice data analyst who wants to start with Machine Learning, up to the data scientist or researcher who doesn’t want to code but prefers to focus on knowledge discovery. Head over to our plans and pick the one that is right for you.

If you have already signed up in the past and would like to reactivate your account with a Basic FREE plan, drop us a line at social@jadbio.com
View our Plans>

NEWS

JADBio’s automated machine learning (ML) platform speeds up BioLizard’s ML model performance evaluation capabilities.

BioLizard & JADBio Announce their Partnership

BioLizard is an agile bioinformatics and AI/ML company globally servicing clients in the life sciences, biotechnology, and pharmaceutical industries. Originating from Belgium and with offices in the Netherlands, Spain, and the US, BioLizard has a strong track record in providing cutting-edge ML technologies to model even the most complex scientific data. Their in-depth modeling and optimization capabilities, and the highly automated and interactive AutoML platform of JADBio, are highly complementary and result in faster and more complete ML services. Read the Announcement>

Hackernoon Startups of the Year

VOTE FOR US!

And take us across the finish line! Hackernoon is organizing the Startup of the Year 2021 awards for each city around the World. JADBio is competing, in Southern Europe, and we’d love to have your vote! Takes only a few seconds and you actually don’t have to log in or have an account.

Thank you for giving us this beautiful season gift 🎁😊! Vote Now>

🌱 ML NEWS

The Hitchhiker’s Guide to Survival Analysis

ReadWrite

Survival analysis is the best thing in the world since sliced bread! However, in most machine learning circles, it’s pretty much synonymous with an “#itscomplicated” relationship status. Read Benedict Timmerman’s, getting started guide, on Survival Analysis to help you through the complex understanding of this amazing type of predictive analysis.

Tip: JADBio has built-in algorithms if you want to run Survival Analysis on your data! Read the full article>

TWITTER

1st Twitter Chat on Knowledge Discovery & ML

In case you missed it here’s a roundup of our first Twitter chat on Knowledge Discovery and Machine Learning. We’ll be doing many more discussions on Twitter and we’d love to have you join. Follow us on Twitter for the next chat or don’t hesitate to open up a discussion using the hashtag #autoML.

Twitter Chat>

A Gift for this Holiday Season!🎄🎁 was originally published in AutoML for Life Sciences on Medium, where people are continuing the conversation by highlighting and responding to this story.

This month, our #AutoML Twitter Chat discussed the topic “Machine Learning and Knowledge Discovery.” Knowledge discovery is an essential process in every new breakthrough, whether that is a new treatment, a new drug, or just understanding a disease.

Below are each of the questions, along with some of our favorite answers for you to check out, but if you want to see the whole chat, head over to Twitter and follow the #AutoML hashtag!

First off, welcome everyone! Follow @WeAreJADBio 😊 and use the #AutoML hashtag if you want to join the chat (yes, you still can!)

JADBio on Twitter: "Q1: In ML, should we only be focusing on creating a model and its predictive performance? #AutoML / Twitter"

Q1: In ML, should we only be focusing on creating a model and its predictive performance? #AutoML

Aris Karanikas on Twitter: "To me, making sense of the data can be even more important, at times, than the predictive model itself. Feature selection is as the pinnacle of that process. #AutoML / Twitter"

To me, making sense of the data can be even more important, at times, than the predictive model itself. Feature selection is as the pinnacle of that process. #AutoML

Vicki Kolovou on Twitter: "There's a lot. Feature selection or insights like understanding what the model represents. #autoML / Twitter"

There's a lot. Feature selection or insights like understanding what the model represents. #autoML

Pavlos Charonyktakis on Twitter: "In some domains, such as molecular biology and life sciences, a researcher is mainly interested in understanding the underlying mechanism. #AutoML / Twitter"

In some domains, such as molecular biology and life sciences, a researcher is mainly interested in understanding the underlying mechanism. #AutoML

JADBio on Twitter: "Q2: What is knowledge discovery in machine learning? #AutoML / Twitter"

Q2: What is knowledge discovery in machine learning? #AutoML

Vicki Kolovou on Twitter: "Newbie in #ML & models, but recently got to c how the loss of the p53 tumour-suppressor, due to mutation of the TP53 gene, actually favours the development of #cancer. This is public knowledge, but I confirmed it in the model I created. That is #knowledgediscovery IMO #autoML / Twitter"

Newbie in #ML & models, but recently got to c how the loss of the p53 tumour-suppressor, due to mutation of the TP53 gene, actually favours the development of #cancer. This is public knowledge, but I confirmed it in the model I created. That is #knowledgediscovery IMO #autoML

Aris Karanikas on Twitter: "We use ML to predict future behavior. Knowledge discovery is the attempt to attribute that future behavior to specific features/quantities. #AutoML / Twitter"

We use ML to predict future behavior. Knowledge discovery is the attempt to attribute that future behavior to specific features/quantities. #AutoML

Pavlos Charonyktakis on Twitter: "need for stopping Machine Learning being considered as a black box, enabling discovery of new knowledge by understanding and interpreting the models. #AutoML / Twitter"

need for stopping Machine Learning being considered as a black box, enabling discovery of new knowledge by understanding and interpreting the models. #AutoML

JADBio on Twitter: "Q3: What is the difference between knowledge discovery and data mining? #AutoML / Twitter"

Q3: What is the difference between knowledge discovery and data mining? #AutoML

Vicki Kolovou on Twitter: "There are several papers out there that put them in the same equation. I presume Data mining also leads to knowledge discovery as can #ML. #autoML https://t.co/Dbr1YYnT8D / Twitter"

There are several papers out there that put them in the same equation. I presume Data mining also leads to knowledge discovery as can #ML. #autoML https://t.co/Dbr1YYnT8D

Vicki Kolovou on Twitter: "I'd ask whether #AutoML has anything to envy from conventional #ML when it comes to knowledge discovery? / Twitter"

I'd ask whether #AutoML has anything to envy from conventional #ML when it comes to knowledge discovery?

Pavlos Charonyktakis on Twitter: "I would not consider these 2 as different when it comes to knowledge discovery. Of course, to me, #AutoML can also automate the process of the knowledge discovery. / Twitter"

I would not consider these 2 as different when it comes to knowledge discovery. Of course, to me, #AutoML can also automate the process of the knowledge discovery.

JADBio on Twitter: "Q4: Can you name any cases where breakthroughs were made from understanding your ML model? #AutoML / Twitter"

Q4: Can you name any cases where breakthroughs were made from understanding your ML model? #AutoML

Vicki Kolovou on Twitter: "Clinically distinguishing between Right (R) vs left (L) sided colorectal cancers (defining molecular differences)... which could lead to new treatments. #ML was used for the predictive model. #AutoMLhttps://t.co/kytik7H37Z / Twitter"

Clinically distinguishing between Right (R) vs left (L) sided colorectal cancers (defining molecular differences)... which could lead to new treatments. #ML was used for the predictive model. #AutoMLhttps://t.co/kytik7H37Z

That’s all for today, everyone! Stay tuned for our next chat 👋 - details are to come soon.

.

“Machine Learning and Knowledge Discovery.” #AutoML Twitter Chat Highlights from 8/12/2021 was originally published in AutoML for Life Sciences on Medium, where people are continuing the conversation by highlighting and responding to this story.

JADBio wins award for Spin-off of the Year (Startup originating from Academia) at an official event organized by Elevate Greece during the 85th Thessaloniki International Fair. 250 startups participated in the competition, and 10 of them were awarded in respective categories. Τhe awards, the brainchild of Christos Dimas, Deputy Minister of Development and Investments, were given out on Saturday, September 11th. The prime minister of Greece, Kyriakos Mitsotakis, was there along with other ministers, including those from Education, Development, Research, and Innovation. The prize was supported with a 20K euro award from the sponsors, including Eurobank, PRAXI Network, Intel-Lex, and Accenture. Read the full Press Release>

RESEARCH

Is it time for Automated Causal Discovery?

Dr. Ioannis Tsamardinos spoke at the International Symposium on Causal Discovery & Machine Learning, during the Hikone Data Science 2021 conference, September 10–11, 2021. The title of his presentation was Towards Automated Causal Discovery and you can now view it on JADBio’s channel over on YouTube.

Watch the lecture>

Our Research Published on Nature.com

It’s a great honour, for the whole team and all the work we’ve put into it, to see our research published on Nature’s Scientific Reports. “Automated machine learning optimizes and accelerates predictive modeling from COVID-19 high throughput datasets” was published in July 2021 and already appeared on social media posts and in Семен Есилевский’s article on Less human labor to save more human lives: AutoML to fight COVID-19

Read the full research article>

🌱 ML NEWS | TOWARDS DATA SCIENCE

Do we really need feature selection in a data analysis pipeline?

Pavlos Charonyktakis CTO at JADBio explains what is Feature Selection and dwells into its advantages and disadvantages.

On why it is important he explains how, feature selection is a common component in supervised machine learning pipelines and is essential when the goal of the analysis is knowledge discovery. Knowledge discovery is really important in some domains such as molecular biology and life sciences, where a researcher is mainly interested in understanding the underlying mechanism of the problem. Selecting a feature subset leads to simplification of models to make them easier to interpret by researchers/users. Finally, an important aspect of feature selection is the cost optimization that a user can achieve by using a model with fewer features. This is especially important if it is very expensive to measure certain features, and each feature is associated with a cost”.

Read the full article >

VIDEO

AI Industry Connect Talks by SKEL

The AI Lab: JADBio

A new series of monthly business talks titled AI Industry Connect was launched in March by SKEL | The Artificial Intelligence Lab and is supported by ahedd Digital Innovation Hub. This initiative aims to bring to the forefront, the possibilities of Artificial Intelligence being put to use by the Industry, in various domains.

Through this series of online talks, wider audiences have the opportunity to find out about the AI innovations of each company in their domain and hear real-life use-cases by professionals. Links between the Industry and the Scientific Community are highlighted, providing the opportunity to find out about the needs of the Industry for applicable scientific results.

The sixth talk in the series of monthly business talks titled AI Industry Connect Talks was held on Wednesday 15 September 2021. This month, we hosted JADBio with the CEO & Founder, Dr Ioannis Tsamardinos as invited speaker.

See talk on SKEL>

Spin-off of the Year 2021 🏆 was originally published in AutoML for Life Sciences on Medium, where people are continuing the conversation by highlighting and responding to this story.

Newsletter #2 — LifeSciences & Machine Learning news

WEBINAR

Discover how you can predict the best quality potato for your next plate of fries! 🍟

Webinar ran successfully on June 24th. See the video recording below, and review the outcome models on JADBio. Watch the Webinar recording below:

We show how a team of researchers applied JADBio’s Automated Machine Learning (AutoML) platform to predict potatoes’ susceptibility to bruising and also its potential for coloration during chip/crisp processing. The aim was to differentiate between potatoes that would be less prone to bruising from those that would more easily bruise during mechanical handling. Another goal was to successfully predict the potatoes’ potential susceptibility to acrylamide formation during chip/crisp processing due to the Maillard reaction.

VIEW POTATO CHIP QUALITY ANALYSIS>
VIEW POTATO BLACKSPOT BRUISING ANALYSIS>

NEWS

Multi-omics characterization of
left-right colorectal cancer

Researchers used JADBio to classify left and right Colorectal Cancer, based on the fold change in values between tumor and adjacent normal, with promising results. The researchers, John Marshall, Takayuki Yoshino et al., published their paper, on Multi-omics Characterization of Left-Right Colorectal Cancer, in the Journal of Clinical Oncology and presented their work at ASCO 2021 annual Meeting.

Future Data Analyst Will Be Customising AutoML Tools

Datatechvibe writes about how AutoML needs educated users. In an exclusive interview with Dr. Ioannis Tsamardinos, CEO and Co-founder at JADBio, they discuss how AutoML is changing the face of ML-based solutions. How AutoML can provide significant value to life sciences and why Healthcare companies often struggle with unlocking value from their own data.
Read the interview>

🌱 BIGGER NEWS

What is Survival Analysis?

The field of statistical analysis that applies specific methodologies to explore the time it takes for an event to happen is the bare-bones definition of Survival Analysis. The amount of time it takes before a predetermined event takes place is also known as time-to-event analysis. By “survival” in this context, we refer to what remains free of a particular outcome over time.

JADBio performs feature selection with hundreds of thousands of markers and image features, which filters out not only the irrelevant ones but also the redundant markers. Read the specific example on the problem of predicting the survival time after surgery of low-grade glioma patients from miRNA profiles measured in tumor biopsies. Read the full article >

CASE STUDY

Predicting PD-1/PD-L1 inhibitors treatment on metastatic non-small cell lung cancer

The researchers used JADBio for binary classification modelling for the prediction of the probability of a single individual to achieve DS (PR or SD vs. PD) with ICIs as second-line treatment. The feature classification of the parameters used as input in JADBio is demonstrated in S1 Table. The tool applied the following modelling algorithms: support vector machines (SVM) with full polynomial and Gaussian kernels [25], random forests [26], ridge logistic regression [27], and decision trees [28]. The performance metric we chose over the several ones available at JADBio, is the AUC. In most cases, the result of an analysis will be a complex model, incomprehensible to the human user. To aid in that regard, JADBio additionally outputs the best interpretable model. In their work, they report the performance estimation of the best-performing model. Read the Case Study>

Head over to all our Case Studies 🤓>

✨ KNOWLEDGE

JADBio enhances security by adding 2-factor authentication. Now, our users can increase the protection of their accounts by enabling two-factor authentication using the free Google authenticator app.

🧬It’s not all about Humans🥔 was originally published in AutoML for Life Sciences on Medium, where people are continuing the conversation by highlighting and responding to this story.

NEWS

The future of medical treatment lies where AutoML & life scientists meet

by Benedict Timmerman

AutoML allows life scientists to be life scientists, doing what they do best, analyzing the final information, and making educated connections between data and disease. Read the full article on ITProPortal>

JADBio on InsideBigData

Tooting our own horn here… JADBio was featured on InsideBigData, by their editorial team. They write how “JADBio is an AI startup company working with BioMed data. This remarkable team, headed by Prof. Ioannis Tsamardinos, has created an automated machine learning (AutoML) platform designed for life scientists. No Coding. No Statistics. No Math. No Problem … just add data.” Read the full article on Inside BigData>

BIGGER NEWS…

JADBio adds image classification

JADBio’s API interface now has image analysis capabilities. Using pre-trained Deep Neural Networks for image analysis, we enable researchers, clinicians, bioinformaticians, biologists, and biomedical staff to use their medical image datasets to train and deploy machine learning models with the addition of feature construction optimization without writing any code. Read the full anouncement>

CASE STUDY

Automated Machine Learning Optimizes & Accelerates COVID-19 Predictive Modeling

Standard machine learning analysis of proteomic and metabolomic data from COVID-19 patients produced biosignatures that contain large numbers of predictors, hampering their clinical application. Moreover, their performance often drops significantly when validated in independent groups, which is expected as sample numbers are often inevitably low. By applying automated Machine Learning, we attempt to improve modeling and deliver models/signatures that can be readily available for diagnostic assays to aid the fight against the pandemic. Read the Case Study>

Head over to all our Case Studies 🤓

KNOWLEDGE 👨‍🎓

Did you know you can SHARE your analysis results with viewers outside the JADBio platform? Check out an example from Diagnosing Alzheimer patients from circulating miRNA biomarkers

👨‍🏫 WEBINAR: 5/27/2021

Automated Machine Learning for Bioinformaticians: the JADBio platform

In this webinar series, Aris Karanikas, Business Development at JADBio, and Giorgos Papoutsoglou, Product Manager at JADBio, will demonstrate some advanced capabilities of AutoML using the JADBio platform and real case studies. He will explain how you can build your data-driven models and utilize a large set of visualizations and reports to gain insights from your data. Join the Webinar on Zoom >

🌱 We could all use a fresh start 🚀 was originally published in AutoML for Life Sciences on Medium, where people are continuing the conversation by highlighting and responding to this story.

If you look up Wikipedia, you’ll read Automated machine learning (AutoML) is the process of automating the process of applying machine learning to real-world problems. AutoML covers the complete pipeline from the raw dataset to the deployable machine learning model. AutoML was proposed as an artificial intelligence-based solution to the ever-growing challenge of applying machine learning. The high degree of automation in AutoML allows non-experts to make use of machine learning models and techniques without requiring becoming an expert in the field first.

Basically, it hits the nail on the head but if we were Wikipedia editors we would make a slight change where it says “to the ever-growing challenge of applying machine learning”. The actual challenge in today’s world is the ever-growing amount of data. We’ve read that biomedical data doubles every year, but although we don’t have credible statistics of the actual overall size or rate, we do know the amount of collected data is huge. Machine learning has the tools to tackle all sorts of data and amounts, but there’s not enough data scientists or analysts to go around or enough hours in the day for them to tackle these huge amounts of data produced daily. While this may sound less significant if we’re analyzing demographics to predict election outcomes — the US 2020 Election polls proved it wasn’t mundane — it’s a lot more crucial when it comes to health data, and unlocking valuable information that can lead to cures for serious diseases. So how do we go about scaling and streamlining machine learning?

Until a few years ago the field of Machine Learning was limited to academic research and to a few select data scientists who knew how to program. This has changed significantly for three main reasons:

• The cost to store, access, and analyze Big Data has fallen significantly, thereby opening the potential for its use
• Cutting-edge technological advances including Artificial Intelligence applications and Cloud Computing has given broader access to many organizations
• Several industries have embraced technological solutions like sensors (IoT) and other smart solutions, almost eliminating downtime of assets. This means more data and an increased need for actionable insights. In the meantime, while technology has led to cost reduction it also has yielded a hunger for real-time insights.

Basically, it’s a combination of technological innovations and economic factors that has shifted machine learning from an apparatus in the academic’s lab to a power-tool on the industry floor.

Why AutoML?

Machine learning for a data scientist involves preparing data, training models, manually tuning for hyperparameters and model compression, and iterating through thousands of models which take days or weeks to do manually. AutoML aims to fully automate the machine learning process end-to-end, democratizing machine learning for non-experts and drastically increasing the productivity of expert analysts. Its goals are to completely automate the application of machine learning, statistical modeling, data mining, pattern recognition, and all advanced data analytics techniques. So while providing an efficient productivity tool it strives to reach credible results shielding against statistical methodological errors, and even surpassing manual expert analysis performance, for e.g. by using meta-level learning.

AutoML focuses on three targets:

• Accelerate human productivity while cutting costs
• Democratize machine learning for all irrespective of the level of expertise
• Improve replicability of analyses, sharing of results, and facilitate collaborative analyses

Let’s Clarify What Makes a Machine Learning Solution Truly Automated

The minimal requirements of an AutoML platform are the ability to return a. a predictive model that can be applied to new data, and b. an estimate of the predictive performance of that model, given a data source, e.g. a 2-dimensional matrix (tabular data). Thus, DIY tools that allow you to graphically construct the analysis pipeline (e.g. Microsoft’s Azure ML) are not considered as AutoML platforms. Open-source libraries and services like Scikit Learn, Weka, and Keras require coding knowledge by the user and thus are not AutoML according to the meaning suggested here. AutoML services usually include a user interface and while striving to make machine learning-friendly to coders, they also want anybody with a computer to be able to use them, typically offering a much wider range of functionalities.

Algorithmically, AutoML encompasses techniques regarding hyper-parameter optimization (HPO), algorithm selection (CASH), automatic synthesis of analysis pipelines, performance estimation, and meta-level learning, to name a few. In addition, an AutoML system could not only automate the modeling process, but also the steps that come before and after. Pre-analysis steps include data integration, data preprocessing, data cleaning, and data engineering (feature construction). Post-analysis steps include interpretation, explanation, and visualization of the analysis process and the output model, model production, model monitoring, and model updating. The ideal AutoML system should only require the human to specify the data source(s), their semantics, and the goal of the analysis to create and maintain a model into production indefinitely.

One of the first AutoML solutions was the Gene Expression Model Selector(GEMS) introduced as early as 2004 and since then several academic and commercial solutions have appeared. Comparatively evaluating them is not easy. You need lots of datasets with different characteristics, extensive computational time, and in the end finding comparative ways to challenge them fairly. Researchers Iordanis Xanthopoulos, Ioannis Tsamardinos, Vassilis Christophides, and SAP’s Eric Simon and Alejandro Salinger, argue, in Putting the Human Back in the AutoML Loop, that while “AutoML strives to take the human expert out of the ML loop, unfortunately, it seems the majority of AutoML surveys and evaluations also take the human user out of the loop, focusing solely on predictive performance and ignoring the user experience for the most part”. They endeavored into a qualitative evaluation of different AutoML platforms including Auger.ai, BigML, Darwin, H2O’s Driverless AI, RapidMiner, IBMs Watson, and JADBio, which was launched in November 2019 by Professor I. Tsamardinos focused on the analysis of molecular biological data (small-sample, high-dimensional) and an emphasis on feature selection.

As the authors of the above paper acknowledge “…statistical estimations are particularly challenging with low samples; even more so with high dimensional data. Is performance overestimated, standard deviations underestimated, probabilities of individual predictions uncalibrated, feature importance’s accurate, or multiple feature subsets returned not statistically equivalent? Which AutoML services return reliable results one can trust, and which ones are actually misleading the user and potentially harmful? In the case of medical applications, overestimating performance or confidence in a prediction (uncalibrated predicted probabilities) is dangerous and could impact human health, while in business applications it may have significant monetary costs”. While arguing that significant experimentation is needed to test all the different services they foresee a future for AutoML and predict that “…within a few years, most of data analysis will involve the use of an AutoML service or library; scripting as a means to manual ML analysis will gradually become obsolete or pass to the next level, where it is customizing and invoking AutoML functionalities”.

While prediction reliability is crucial, users also care about the interpretability of results. They need to understand patterns in their data, visualize, and interpret them. They also need to be able to examine the analysis process and ensure its correctness or optimality. As Dr. Ioannis Tsamardinos, CEO of JADBio says “AutoML should automate, not obfuscate. The analysis process should be transparent, verifiable, and customizable by the user”.

Can AutoML replace Data Scientists?

Did python reduce the need for code developers? No. On the contrary, there is an ever-growing need for people who can code in python. Once a task is automated, humans move to other tasks, that yet to be automated. And there never seems to be a shortage of interesting things to do. AutoML frees data scientists to focus on formulating the analysis problem in different ways, exploring more options, interpreting results, and applying ML on data there was no time for.

As Tommy Blanchard, data science manager at Klaviyo, says on Towards Data Science “I’ve had people ask me if I’m worried about my job security as a data scientist. No, I am not. I can’t wait until these tools are there and open source so I can just type “import machinelearn” and just have it do the stupid hyperparameter optimization and I can get on with the hard part of the job”. Obviously, he understandably argues there’s more to be expected from these automation tools. We wouldn’t expect less.

AutoML is here to expedite Data Scientists’ work, permit them to attempt things, and focus on what matters, improved insights.

A Few Words About JADBio AutoML

JADBio is the culmination of 20 years of machine learning research and know-how in biomedical data analysis, several years of development, and hundreds of thousands of lines of computer code. JADBio accepts a 2D data matrix where the rows correspond to samples (e.g., molecular profiles) and the columns to features (variables, quantities, attributes). One of the features should be selected as the outcome of interest to model and learn to predict. The outcome could be a binary quantity, discrete quantity, continuous quantity, or a time to an event of interest, leading to binary classification, multi-class classification, regression, and time-to-event analysis, respectively. JADBio automatically searches in the space of combinations of algorithms for all steps of the analysis, namely preprocessing, imputation of missing values, feature selection, and modeling, and their corresponding hyper-parameter values (tuning parameters). It thus tries thousands of analysis pipelines, called configurations, to identify the best one (i.e., performs HPO) to produce the final model. It returns (a) a final model, (b) estimates of its out-of-sample (i.e., on new samples) predictive performance, and © the selected features after removing the irrelevant and redundant ones. JADBio performs what is called multiple feature selection returning multiple selected feature subsets that lead to equally predictive models. It also returns numerous other useful pieces of information such as features’ importance (i.e., the added value of each feature in the final performance), an explanation of each feature’s role in the model (e.g., to allow one to understand it is a risk factor or a protective factor), an indication of the samples that could be mislabeled, results on each configuration and algorithm tried, and the best approximation achieved with a humanly interpretable model (e.g., linear). The final model could be downloaded in executable form, applied to an external validation set, or run manually by feeding in the observed value of the selected features.

See It In Action [CASE STUDY]

Thymoma Case Study

This case study was from a collaboration with the AWG on cancer subtype classification. Primary tumor biopsies from 117 thymoma patients were profiled by The Cancer Genome Atlas (TCGA) for copy number variation, gene expression, methylation levels, microRNA and genomic mutations. 4 subclasses were defined in a previous study based on all 27796 molecular measurements combined, using a multi-omics clustering approach. The 2-dimensional 117×27797 data matrix with the measurements (rows corresponding to samples from individual patients, columns corresponding to features and the outcome) was uploaded to JADBio (Figure 1, left) as a CSV file. Subsequently, the column with the outcome (thymoma subclass) was selected (Figure 1, right) and the analysis begins. That was all a user was required to do. Once the analysis began, JADBio’s AI system searched the space of possible models to identify the optimal one and estimate its performance. For this analysis, JADBio trained 20890 models within 41 minutes using 16 CPU cores. The winning model turned out a Random Forest ensemble of 100 Decision Trees after feature selection with the LASSO algorithm.

JADBio presented multiple visuals, graphs, and reports. A few examples are included in Figure 2. Figure 2(a) shows the out-of-sample estimated ROC curve along with confidence intervals. The AUC averaged over all classes for the best predictive model is 0.976 (C.I. 0.931–1.000). Feature selection for the winning model returned 25 features. However, JADBio also reported the best approximate model that is humanly interpretable. The best interpretable model can distinguish the four thymoma subclasses based on just two molecular features with a slightly lower 0.946 AUC. There are multiple signatures reported, i.e., combinations of two features that lead to an equally predictive model (multiple feature selection). Some are shown in Figure 2(b). The first one reported is the expression values of the gene CD3E and the miRNA miR-498, respectively. miR-498 is the marker most associated (pairwise) to the outcome. However, CD3E ranks only 190 in terms of pairwise association with the outcome! In other words, if one performs standard differential expression analysis, they will have to select 189 other markers before reaching CD3E. In contrast, JADBio’s feature selection algorithms recognize and filter out the redundant features. Figure 2(c) shows the individual importance for each of the two features of the reference signature measured as the drop in relative performance when a single feature is removed from the model.

Learn more through implementation by taking advantage of its 14-day free trial.

JADBio‘s easy-to-use interface allows biologists, bioinformaticians, clinicians, and non-expert analysts to perform sophisticated analyses with the click of a button. It produces multiple visuals, graphs, and reports, in order to provide intuition, understanding, and support decision making. Novel statistical methods avoid overfitting and overestimation of performance even for low sample sizes. It performs feature selection -biosignature identification- by removing irrelevant, but also redundant features(markers) for prediction. It has been validated on hundreds of public datasets, producing novel scientific results.

Further reading:

Putting the Human back In the Loop, EDBT-ICDT-WS 2020
Just Add Data: Automated Predictive Modeling and BioSignature Discovery, Biorxiv 2020
Automated Mortality Prediction in Critically-ill Patients with Thrombosis using Machine Learning, BIBE, 2020
Accurate Blood-Based Diagnostic Biosignatures for Alzheimer’s Disease via Automated Machine Learning, PMC, 2020

What is AutoML? was originally published in AutoML for Life Sciences on Medium, where people are continuing the conversation by highlighting and responding to this story.

SBV-IMPROVER organized the Metagenomics Diagnosis for IBD Machine Learning Challenge (MEDIC) aimed to investigate the diagnostic potential of metagenomics data to classify patients with Inflammatory Bowel Disease (IBD) and non-IBD subjects. The participants attempted to classify Ulcerative Colitis (UC) and Crohn’s Disease (CD) subjects with data obtained from non-invasive clinical samples. The challenge came with a prize pool of $12,000.

The aim was to find the best classification algorithm that can be used in diagnosing Inflammatory Bowel Disease with data obtained from non-invasive clinical samples. The basic question addressed by the participants was: Can you predict IBD status using metagenomics data?

More specifically, MEDIC aimed to verify that shotgun metagenomics sequencing data is sufficiently informative to allow for accurate classification of human subjects as:

IBD vs. non-IBD
UC vs. non-IBD
CD vs. non-IBD
UC vs. CD

With the analysis of the predictions submitted in the challenge, the goal was to answer the following scientific questions:

Which predictive computational approaches are the most accurate across the four 2-class problems described above?
What do the most discriminative metagenomic features tell us?
Are they rather based on taxonomy, functions/pathways, and/or other types, e.g., k-mers?
Are they distinct between UC vs non-IBD and CD vs non-IBD or do they show commonalities?

The Machine Learning Sub-Challenge

There were two sub-challenges and JADBio participated in the second sub-challenge:

In the second sub-challenge (“MEDIC PROCESSED”), participants were provided with pre-calculated taxonomic and pathway abundances matrices derived from the raw data. This allowed data scientists with no access to metagenomics analysis pipelines to solve the Challenge, as well as to compare the performance of classification methods beyond the role of pre-processing steps. The organizers provided participants with shotgun metagenomics sequencing data as raw and processed data for predictor model training and testing.

The Results

JADBio’s automated machine learning findings were ranked in 4th place out of 13 participants. Although ranking might not be the ultimate goal compared to knowledge discovery our Senior Data Scientist at JADBio, Konstantinos Paraschakis, has a few thoughts on the total number of submissions and actual outcomes. He notes that “…this[4th place] doesn’t quite tell the whole story, in my opinion”. He argues that “There was a prize of 2,000 USD for each one of the top three teams in each sub-challenge and every participant could participate with multiple submissions. So, there were teams, obviously motivated by the prize, that participated with several submissions. And it was indeed those teams that made it to the top three. The winner team of the second sub-challenge, for example, submitted five prediction sets, the second winner eight, and the third winner… hold your breath… 32!!! So, if one team had sent 1000 submissions with completely randomly generated predictions, they would have most probably won the challenge with one of them”.

For those of you who are familiar with the JADBio algorithms, this is exactly what BBC-CV(Bootstrap Bias Correction) is trying to fix in the estimation of the best model’s performance. Below is a table of all participants of the second sub-challenge, along with the number of submissions by each one, their best rank, as well as their average rank. JADBio is the best performing team among teams with only one submission and has the best average ranking as well.

To view full rankings go to the official website here.

How Multiple Submissions May Be Distorting Real Outcomes in Machine Learning Challenges was originally published in AutoML for Life Sciences on Medium, where people are continuing the conversation by highlighting and responding to this story.