From Data Mining to Pleiotropic Effects, Environmental Interactions, and Phenomic Predictions of Natural Genetic Variants in Sorghum and Maize
Ravi Mural, Research Asst. Professor; Center for Plant Science Innovation UNL
Author
10/02/2023
Added
40
Plays
Description
Crop yield hinges on genotype variation, environmental responses, and genotype x environment interactions. Understanding links between plant traits and yield across diverse environments is crucial for predicting performance. Leveraging plant quantitative genetics, we compile data from a maize super-panel, 1,118 accessions, with 100+ traits and 17M genetic markers using community association populations. This offers insights into genetic correlations, pleiotropy, and the genetic control of genotype-environment interplay. Similar approaches in other crops hold potential for new discoveries to enhance breeding and commercial gains.
Searchable Transcript
Toggle between list and paragraph view.
- [00:00:00.750]The following presentation
- [00:00:02.220]is part of the Agronomy
- [00:00:03.570]and Horticulture seminar series
- [00:00:05.790]at the University of Nebraska Lincoln.
- [00:00:08.850]Good afternoon everybody.
- [00:00:10.470]Welcome to the agronomy and hoticulture seminar series
- [00:00:15.510]and my name is Camilla Karina Rodriguez.
- [00:00:18.630]I'm a masters student in this department
- [00:00:21.150]under Dr. Chris Proctor.
- [00:00:22.980]And today I am pleased to host Dr. Ravi Mural,
- [00:00:26.970]an assistant professor of quantitative
- [00:00:29.220]genetics and genomics
- [00:00:30.660]at South Dakota State University.
- [00:00:33.450]Dr. Mural received his bachelor degree in agriculture
- [00:00:37.620]from Agriculture University Dharwad India
- [00:00:40.770]and his masters degree in
- [00:00:42.090]plant breeding and genetics
- [00:00:43.500]from Agriculture University Bangalore, India.
- [00:00:47.130]After completing his masters degree,
- [00:00:49.020]he moved to university of Arkansas
- [00:00:51.660]where he earned another masters degree
- [00:00:53.730]in applied bioscience.
- [00:00:55.410]He then pursued his PhD
- [00:00:57.420]in agronomy and horticulture
- [00:00:59.370]with a specialization in plant breeding and genetics.
- [00:01:03.270]During his PhD, he had trained
- [00:01:05.581]in the soybean breeding program
- [00:01:08.490]at the University of Nebraska Lincoln,
- [00:01:10.830]working under the supervision of Dr. George Grief.
- [00:01:15.030]After completing his PhD,
- [00:01:16.590]he joined Professor Shanabel's group as a postdoc,
- [00:01:19.950]followed by a role as research assistant professor
- [00:01:22.710]for about a year and a half.
- [00:01:24.360]And recently in his current position today,
- [00:01:27.690]he will be discussing a bit about his academic journey
- [00:01:30.870]as well as his research during his time
- [00:01:33.210]with Professor Shanabel,
- [00:01:34.950]which he will talk about the title
- [00:01:37.920]and who is here in person.
- [00:01:40.140]After the presentation,
- [00:01:41.355]just raise your hand for questions
- [00:01:43.320]and I will give you the microphone.
- [00:01:45.360]And for those online,
- [00:01:46.770]just put your question in the chat.
- [00:01:48.630]Thank you.
- [00:01:50.220]Thank you Camilla.
- [00:01:55.860]Thank you.
- [00:01:57.240]Thank you for giving me this opportunity
- [00:01:58.710]to come and share about like, you know,
- [00:02:01.170]some research program and things that I did earlier.
- [00:02:04.410]So as you see, my research title is from Data Mining
- [00:02:07.440]to Pleiotropic Effects, Environment Interactions
- [00:02:10.410]and Phenomic Predictions
- [00:02:11.670]of Natural Genetic Variants in Sorghum and Maize.
- [00:02:14.070]That's a lot of things that I'm packing within like,
- [00:02:15.777]you know, small presentation.
- [00:02:18.521]But before like I started,
- [00:02:20.700]I prepared everything only with respect to research.
- [00:02:23.550]But recently some students came to me and they were like,
- [00:02:25.770]yeah, it will be great if we can talk about your journey.
- [00:02:29.310]So then like I read some of these slides and put some extra
- [00:02:32.550]stuff about myself.
- [00:02:34.020]So I'm gonna talk about those.
- [00:02:37.320]So I have divided my presentation
- [00:02:39.360]in four different parts,
- [00:02:40.920]mainly academic journey,
- [00:02:42.300]then research that I have done here,
- [00:02:45.270]then future research work and acknowledgement
- [00:02:48.450]and let's start with academic journey.
- [00:02:52.830]So I was born and brought up
- [00:02:54.510]in Bharat that is former name is
- [00:02:57.090]India, which is being changed now.
- [00:02:59.010]And I was born in north of India
- [00:03:01.380]and then I moved to various
- [00:03:02.400]places because my dad was
- [00:03:03.690]in Indian army and every three
- [00:03:05.070]years or four years we would move
- [00:03:06.330]to a new place and most of
- [00:03:08.760]my education was in army school,
- [00:03:10.590]which were like, you know,
- [00:03:11.850]the elite school in India.
- [00:03:13.890]And that really helped me a lot.
- [00:03:15.390]I think initial years.
- [00:03:16.920]But then like now after his retirement,
- [00:03:18.450]we moved back to his native where like, you know,
- [00:03:20.610]I did my bachelor's in agriculture
- [00:03:22.410]from Dharwad Agriculture University.
- [00:03:24.720]And after my bachelor's I moved to Bangalore Agriculture
- [00:03:28.076]University where I did my master's
- [00:03:30.990]in genetics and plant breeding.
- [00:03:32.310]I worked with Dr. Sheleshja Hitalmani,
- [00:03:34.980]she was a rice breeder and a finger millet breeder.
- [00:03:38.100]And I also worked a little bit
- [00:03:39.540]on chili breeding program
- [00:03:40.860]like you know, after my completion of my master's.
- [00:03:44.580]But after that I moved to US
- [00:03:46.620]and I joined University of
- [00:03:48.810]Arkansas at Little Rock where I work mainly on host pathogen
- [00:03:51.810]interaction and mainly focusing
- [00:03:53.880]on ubiquitination pathway,
- [00:03:56.229]how it orchestrated different like you know,
- [00:03:58.380]diseases and plants.
- [00:04:00.810]Post that I moved to University of Nebraska Lincoln,
- [00:04:04.500]which has been my home
- [00:04:05.430]for a long time now because I did my
- [00:04:08.250]PhD here with Dr. Graf.
- [00:04:10.830]And so I've been breeding program
- [00:04:12.480]and I learned a great deal
- [00:04:13.770]of things to work in field and like, you know,
- [00:04:16.710]learn soybean breeding and crossing
- [00:04:18.960]and managing
- [00:04:20.370]big field evaluations basically.
- [00:04:23.160]And post that after completion of my PhD,
- [00:04:27.690]I learned a lot of these things
- [00:04:28.860]before I worked in molecular
- [00:04:30.090]biology and then like in field evaluation,
- [00:04:31.830]but I was still missing like
- [00:04:32.910]I want to learn this big data analytics.
- [00:04:34.560]Everywhere you hear about it, it's all like,
- [00:04:36.237]you know, big data analytics and all.
- [00:04:38.580]So at that time I got some position
- [00:04:40.290]in other university and they like, you know,
- [00:04:43.509]offered me that position three times saying like, no,
- [00:04:46.170]you're gonna increase your salary.
- [00:04:47.370]But at that time my mind was very clear,
- [00:04:49.710]I want to learn couple of things
- [00:04:51.300]that Dr. Shanabel's is doing.
- [00:04:52.890]I went and I bugged him a lot
- [00:04:54.240]I guess at that time.
- [00:04:56.220]And we had a very good
- [00:04:58.020]interaction first time I met
- [00:04:59.430]because he asked me very genuine questions,
- [00:05:02.610]what's your idea?
- [00:05:03.450]What do you want to do for long term?
- [00:05:05.580]And that's when like, you know,
- [00:05:07.050]I explained him why I want
- [00:05:08.520]to get into his lab and what I
- [00:05:09.530]want to do, a small idea,
- [00:05:11.340]although that I want to combine
- [00:05:13.110]multiple phenotypes and see like, you know, (indistinct)
- [00:05:14.997]and he was generous enough
- [00:05:16.080]to let me work on those things.
- [00:05:17.691]And the whole time with James, like you know,
- [00:05:20.640]I worked as a postdoc with him
- [00:05:22.290]for about a year and a half
- [00:05:24.030]and then he gave me position
- [00:05:26.070]of research assistant professor.
- [00:05:28.080]So after that he moved to Google
- [00:05:32.820]for a short period of time to
- [00:05:33.990]work and I got a very good chance
- [00:05:35.640]of managing the big lab.
- [00:05:37.080]He has like in a huge lab,
- [00:05:38.130]you all know 15 to 20 people and that was really
- [00:05:40.860]transformative for me.
- [00:05:43.110]So many people like you know,
- [00:05:44.280]handling everyone coming
- [00:05:45.390]from different cultural background and like, you know,
- [00:05:47.370]talking to them and understanding how they think managing
- [00:05:50.100]all that initial one or two weeks
- [00:05:52.410]were really tough for me
- [00:05:53.430]because James gave like, you know,
- [00:05:55.200]this big responsibility
- [00:05:56.310]and he was gone and then all of a sudden,
- [00:05:57.720]okay I had to write these
- [00:06:00.030]like you know, proposal reports.
- [00:06:02.010]I had no clue what to do at that time.
- [00:06:03.870]And then like, and I thought eh, it's gonna be fine.
- [00:06:05.910]But all of a sudden I got bombarded with five,
- [00:06:07.860]six different things.
- [00:06:08.693]Okay, you had to write a report
- [00:06:10.200]and like, you know,
- [00:06:11.033]as soon as possible but business
- [00:06:12.660]people were really helpful
- [00:06:14.370]and collaborators were really helpful.
- [00:06:16.140]I still remember Dr. Yufanga,
- [00:06:18.000]he called me and he was like, okay,
- [00:06:19.530]let's have a zoom call.
- [00:06:20.640]I didn't even ask for it and we got into call,
- [00:06:23.220]he explained me what to write,
- [00:06:24.390]how to write, what to do,
- [00:06:25.470]and that was like, you know,
- [00:06:26.370]really great at that time.
- [00:06:27.450]And I started like, you know,
- [00:06:28.590]getting going but even before I joined James lab,
- [00:06:32.580]I went to him somewhere in October,
- [00:06:34.620]I think it was mid or end
- [00:06:36.750]of October and when like our first
- [00:06:38.820]conversation, everything was fine and he was like, okay,
- [00:06:41.640]start attend my lab meetings.
- [00:06:43.980]I went there and I saw how he conducts lab meeting,
- [00:06:46.170]how he was like in a very amicable, very,
- [00:06:48.360]very friendly with students and how he was explaining
- [00:06:50.880]everything I was, ah, this life also exists.
- [00:06:53.220]So I was like nope, I want to like, you know,
- [00:06:54.840]really work with him.
- [00:06:56.280]And that's where like everything
- [00:06:58.050]started and I had learned
- [00:07:01.740]like you know, a great deal of things.
- [00:07:03.360]That's what I can say.
- [00:07:05.970]One thing is James is really
- [00:07:07.950]good at explaining very minute
- [00:07:09.600]details about your research.
- [00:07:10.800]Some things that you cannot
- [00:07:11.940]get in your books like you know
- [00:07:13.650]from book and it cannot be taught in a classroom setup
- [00:07:16.260]I would say that's why if ever
- [00:07:18.000]you get a chance to work with
- [00:07:19.137]him for a small project, also do collaborate,
- [00:07:21.510]do work with him, you will learn like a lot.
- [00:07:24.870]And with that, let's get into some
- [00:07:28.560]of the things in the research.
- [00:07:31.740]So by now at University of Nebraska Lincoln,
- [00:07:34.140]I'm already like you know,
- [00:07:34.973]trained in molecular aspects,
- [00:07:36.240]quantitative genetics aspect
- [00:07:37.470]and crop breeding aspects
- [00:07:39.150]and managing the big lab.
- [00:07:40.860]Then I got this opportunity
- [00:07:41.970]at South Dakota State University
- [00:07:44.040]where I joined recently as a assistant professor.
- [00:07:48.630]It's been like, you know, two weeks I started.
- [00:07:51.090]So I'm still learning a lot of things.
- [00:07:53.550]Alright, so getting into research presentation.
- [00:07:57.090]First let's talk about
- [00:07:57.990]what are community association panels.
- [00:08:00.600]So these are common genetic
- [00:08:01.740]variants which can be used to
- [00:08:03.210]illustrate biological networks,
- [00:08:04.890]contributing to complex traits.
- [00:08:07.650]And these can be assembled once
- [00:08:09.360]and can be distributed to various researchers
- [00:08:11.880]and they can evaluate them for their like, you know,
- [00:08:13.560]trait of interest or whatever they want to do.
- [00:08:16.530]And this can also be genotype ones and once they are
- [00:08:19.350]genotype, all this market information can be also like,
- [00:08:21.447]you know, repeatedly used.
- [00:08:24.300]But there are a lot of challenges
- [00:08:25.470]because researchers work
- [00:08:27.173]with one or few traits of their interest.
- [00:08:29.640]But a plant is made up of many traits, not just one trait.
- [00:08:32.550]And it's impossible for a single
- [00:08:33.960]research group to screen
- [00:08:34.950]everything because it's time consuming,
- [00:08:37.410]it's very costly and it requires a lot of resources.
- [00:08:41.340]Also, how many times, like you know,
- [00:08:42.720]you have so many data you have
- [00:08:44.100]collected that sitting in your shelves
- [00:08:47.400]and you're thinking what can I do with that,
- [00:08:50.280]with a given data set and you have spent so much of time,
- [00:08:53.755]cost and like, you know, efforts,
- [00:08:56.130]so you want to do something with data,
- [00:08:58.320]but saying that you don't like, you know,
- [00:09:00.690]want to use the data just because you have
- [00:09:03.570]that, you have to first understand
- [00:09:05.190]if the data makes any sense and then like, you know,
- [00:09:07.800]make sure the data is good
- [00:09:09.660]and just don't force your data into your models.
- [00:09:12.900]Like, you know, try to come up with the hypothesis,
- [00:09:15.120]figure out like you know,
- [00:09:15.953]if it can be fit into your models,
- [00:09:18.060]if it can make a good story out of it and then like,
- [00:09:20.157]you know, try to use those data.
- [00:09:22.914]With that whole concept
- [00:09:24.210]in my mind, what I first initially
- [00:09:25.980]started doing was we started with
- [00:09:27.810]couple of traits like you know,
- [00:09:29.160]seed related trait.
- [00:09:30.060]I wanted to check clear trophy
- [00:09:31.470]and understand how they
- [00:09:32.400]modulate like you know, seed composition.
- [00:09:34.650]But then we started, okay there are related, right,
- [00:09:36.570]there are some more related right?
- [00:09:37.740]And then we decided okay,
- [00:09:38.670]we'll put all the data
- [00:09:39.690]that's being published on sorghum
- [00:09:40.950]association panel across like you know,
- [00:09:43.620]the US or whatever papers are published.
- [00:09:46.410]I was able to find out about 234 phenotypes which were
- [00:09:51.120]distributed across like, you know, 20 different
- [00:09:55.380]environments across US.
- [00:09:56.850]And all these are coming
- [00:09:58.500]from over 16 to out of 30, 35
- [00:10:00.990]papers and coming from eight research group.
- [00:10:04.020]All these data were again divided
- [00:10:06.150]into economic biochemical
- [00:10:07.320]disease, reproductive roots, seed
- [00:10:09.270]and vegetative traits.
- [00:10:10.230]Most of them are biochemical
- [00:10:11.760]traits and many are like,
- [00:10:12.593]you know, very good vegetative traits.
- [00:10:15.660]And because these all are single environment traits,
- [00:10:18.690]either blobs like in know we don't have raw data.
- [00:10:21.570]So that is one thing
- [00:10:22.500]I think that researchers should do is
- [00:10:24.870]along with their paper,
- [00:10:25.710]they should always publish
- [00:10:26.610]their raw data also because other
- [00:10:28.140]people can reuse them in like in various ways.
- [00:10:30.810]So because I had just one single data point,
- [00:10:33.480]I couldn't calculate the broad
- [00:10:34.800]sensory sensibility but I could
- [00:10:36.060]calculate marker based narrow sensibility.
- [00:10:38.340]So usually like you know the for reproductive traits it's
- [00:10:40.650]really high and then like you know,
- [00:10:42.450]followed by seed and vegetative trait.
- [00:10:47.190]So this was not a easy task I'm telling
- [00:10:49.557]in very simple like in a way, okay,
- [00:10:51.690]I collected all these things
- [00:10:52.860]but imagine every single paper
- [00:10:54.630]they have their own way of like, you know,
- [00:10:56.100]writing these genotypic IDs.
- [00:10:57.990]So some are writing pi underscore
- [00:11:00.090]xxx number, pi.xxx number,
- [00:11:02.760]PI space xx number or like you know,
- [00:11:05.070]some different names.
- [00:11:05.970]Some are just giving numbers
- [00:11:07.140]and they're alternate names for many PIs.
- [00:11:09.570]So it took me forever.
- [00:11:10.440]I wrote like you know,
- [00:11:11.273]lines and lines of scripts,
- [00:11:12.510]R, Python but nothing worked out at that time.
- [00:11:14.910]So finally I had to sit there
- [00:11:16.140]and physically fix each and
- [00:11:17.375]everything and get all the data.
- [00:11:21.000]Well that was something about sorghum
- [00:11:23.010]and then like I did something similar
- [00:11:24.750]with the maize dataset where I had about
- [00:11:26.610]162 phenotypes and coming
- [00:11:28.920]from 33 different environments
- [00:11:30.480]and seven different groups,
- [00:11:31.530]16 papers and six research groups.
- [00:11:34.200]So maize has a huge panel Wisconsin
- [00:11:37.380]diversity panel which has
- [00:11:38.460]about 900 genotypes of that,
- [00:11:41.160]about 202 are like you know,
- [00:11:44.430]overlapping with all three
- [00:11:46.140]major panels that's being used.
- [00:11:47.490]That is Wisconsin Diversity SAM and MAP panel.
- [00:11:51.060]So both and map are essentially subset of Wisconsin
- [00:11:54.360]diversity panel with some lines
- [00:11:55.801]which are specific to these particular panels.
- [00:11:59.700]So that's why I decided
- [00:12:00.780]to get all the lines from here
- [00:12:02.610]and all the publications that screens
- [00:12:05.250]this line for various traits.
- [00:12:07.140]And again these traits were divided
- [00:12:08.520]into multiple phenotypic groups.
- [00:12:13.170]But that now I have this huge dataset
- [00:12:14.934]in both sorghum and corn.
- [00:12:16.800]What can I do with that?
- [00:12:18.300]What can be done with this dataset?
- [00:12:20.220]So these are powerful datasets
- [00:12:22.440]for a couple of reasons.
- [00:12:24.600]One, we know like you know
- [00:12:26.790]from the classical genetic perspective,
- [00:12:28.800]often genes are pleiotropic
- [00:12:30.690]and then like in a mutation
- [00:12:31.800]in one gene can affect multiple traits
- [00:12:34.830]and including like you know,
- [00:12:35.970]some are expected traits
- [00:12:37.200]or some are unexpected associations.
- [00:12:39.690]So with this dataset,
- [00:12:40.740]now I can test with the continuer
- [00:12:42.780]genetics perspective.
- [00:12:43.980]Are there any pleiotropic like you know,
- [00:12:46.200]effects over there?
- [00:12:48.120]Another question I can ask
- [00:12:49.470]is often some of the genes like you know,
- [00:12:52.860]traits for example say flowering time
- [00:12:55.290]that's measured in a lot of environments,
- [00:12:56.880]they might have different effect
- [00:12:58.050]in different environments.
- [00:12:58.980]What are those effects?
- [00:12:59.940]Why are those effects?
- [00:13:01.110]I can test that.
- [00:13:02.730]And from there I can also go
- [00:13:04.080]to cross environment prediction
- [00:13:05.760]using genomic prediction,
- [00:13:06.900]phenomic prediction because now we
- [00:13:08.280]have this dataset which is
- [00:13:09.450]multi environment, multi trait dataset.
- [00:13:11.610]You can use these to do so many like
- [00:13:14.520]you illustrating all these things.
- [00:13:17.790]So with that, let's talk about
- [00:13:19.980]first like you know, pleiotropy.
- [00:13:23.190]Once again getting back to this sorghum dataset,
- [00:13:25.320]all these 234 phenotypes,
- [00:13:27.390]what I did was I used the newly developed marker set
- [00:13:31.710]and which was about 400 markers
- [00:13:34.710]in that from Chen Young who was
- [00:13:36.210]a former student in James lab.
- [00:13:39.450]Used that marker set
- [00:13:40.470]and ran GWAS in all those phenotypes.
- [00:13:43.650]And then like you know,
- [00:13:44.483]I put everything together to find out okay,
- [00:13:46.260]if we are getting any pleiotropic region,
- [00:13:48.960]well we didn't have like, you know,
- [00:13:50.580]many pleiotropy headset at that time.
- [00:13:52.140]There was like, you know,
- [00:13:52.973]one on chromosome two with seven traits
- [00:13:55.320]and one on chromosome six with six traits
- [00:13:58.470]and another on chromosome
- [00:13:59.580]nine which was associated with dwarfing gene.
- [00:14:03.360]Yeah, over there with 13 traits.
- [00:14:05.899]But we didn't have enough hits.
- [00:14:08.880]There were unique hits which are like, you know,
- [00:14:10.320]56 number with composed to genomic interval.
- [00:14:14.370]It was only 31 genomic interval
- [00:14:16.140]and there were about 25 single trait peaks
- [00:14:18.480]and only three with two trait peaks
- [00:14:20.400]and three with more than three trait peaks.
- [00:14:23.070]So not enough pleiotropy that
- [00:14:25.020]we could dissect.
- [00:14:25.853]So what we decided is to use
- [00:14:27.900]another very robust statistical
- [00:14:30.600]method that's called MashR multi-variate adapt shrinkage.
- [00:14:34.740]So using that method,
- [00:14:35.880]I'm not gonna get in details of that,
- [00:14:37.380]otherwise it will take over like in a whole time.
- [00:14:40.230]But multi-trait analysis
- [00:14:42.210]was able to recover several known
- [00:14:44.250]pleiotropic features for large effects segregating
- [00:14:46.830]in the population
- [00:14:48.172]and we were also able to detect various
- [00:14:52.620]previously known pleiotropic
- [00:14:53.760]regions as well as some
- [00:14:55.530]of the effects which were not known.
- [00:14:57.390]Mainly like you know,
- [00:14:58.223]some aboveground traits
- [00:14:59.160]and some below ground traits which
- [00:15:00.390]were pleiotropic and that was really interesting.
- [00:15:03.540]And mainly these three dwarfing genes,
- [00:15:05.190]they were also pleiotropic for various traits.
- [00:15:09.480]So because of time I don't think
- [00:15:11.280]I can get into each and
- [00:15:12.300]everything, but I'm gonna talk
- [00:15:13.410]about two examples, quick examples.
- [00:15:15.990]One is on dw2.
- [00:15:18.360]So this marker was associated
- [00:15:22.389]with plant height and which is like you know,
- [00:15:25.200]dw2 peak and was also associated
- [00:15:27.360]with various matrix of root size,
- [00:15:30.090]root area, panicle length plant,
- [00:15:31.740]surface area and seed weight.
- [00:15:34.560]The effects of dw2 on panicle length
- [00:15:36.570]is already known and as
- [00:15:38.040]well as seed weight and leaf area
- [00:15:40.020]have been previously published,
- [00:15:41.340]but it was not known
- [00:15:42.330]to be associated with the size or area
- [00:15:44.420]of the various root traits.
- [00:15:48.030]So because of this robust analysis,
- [00:15:50.220]we are able to detect that association.
- [00:15:54.240]So the apparent impact of dw2
- [00:15:55.890]on root phenotypes suggests
- [00:15:57.450]that the genes may play equivalent
- [00:15:58.955]role in determining size
- [00:16:00.480]of the aboveground traits as well
- [00:16:02.280]as the below ground trait.
- [00:16:04.650]Similar to that, there is another genome
- [00:16:06.960]on chromosome six which was
- [00:16:09.030]associated with increase
- [00:16:09.863]in the seed oil and protein as well
- [00:16:12.000]as micronutrients.
- [00:16:13.110]But the same allele
- [00:16:13.943]is also associated with the decrease of
- [00:16:15.611]panicle volume and its solidity.
- [00:16:18.390]So which means you'll either
- [00:16:20.250]get more number of quality
- [00:16:21.690]grains or you'll get less number of grains,
- [00:16:26.130]less number of grains with
- [00:16:27.240]more quality traits within it
- [00:16:29.130]or like more number of grains
- [00:16:30.360]with less quality traits.
- [00:16:32.070]So it's just a trait off between like you know,
- [00:16:33.660]different traits what you're getting.
- [00:16:37.080]So with that, this paper is already
- [00:16:39.600]published in genetics and yes we also
- [00:16:41.760]made it to the cover of the genetics at that time.
- [00:16:45.180]So there are many other like you know,
- [00:16:46.680]examples over there,
- [00:16:47.610]which if you are interested
- [00:16:49.470]you can always go and look into that.
- [00:16:52.320]With that then I moved
- [00:16:53.370]to maize association panel with all
- [00:16:55.900]these 162 phenotypes.
- [00:16:58.320]So when I started this after like you know,
- [00:17:01.650]sorghum paper was done and me
- [00:17:04.050]and James were discussing
- [00:17:05.040]about that, I was like yeah,
- [00:17:05.873]yeah I have such experience
- [00:17:07.200]right now with the previous paper.
- [00:17:08.490]Within three months I'm gonna give you like you know,
- [00:17:10.080]all the analysis, all the results.
- [00:17:12.390]He just laughed at me
- [00:17:13.223]and I was like ah, sure, sure,
- [00:17:14.965]we'll see about that.
- [00:17:16.500]So then I go there and like and I start analyzing,
- [00:17:19.260]six months are gone,
- [00:17:20.160]I'm still pulling my hair
- [00:17:21.150]to put the dataset together.
- [00:17:22.950]So that is the problem
- [00:17:24.870]because there is no set convention
- [00:17:27.510]how the dataset should
- [00:17:28.343]be prepared and made public.
- [00:17:30.060]So that's why every single researcher they are like,
- [00:17:32.157]you know, putting dataset in the way they want,
- [00:17:35.481]because many like you know,
- [00:17:37.440]are thinking in direction
- [00:17:38.520]of one phenotype or associated phenotype,
- [00:17:40.440]not a plant as a whole
- [00:17:41.700]and like in all the phenotypes as a
- [00:17:43.110]whole, if that is done
- [00:17:44.430]that would be like you know, really good.
- [00:17:45.690]But at the same time I really want
- [00:17:46.836]to mention the sorghum
- [00:17:48.720]and maize community has been really amazing.
- [00:17:50.850]They have like, you know,
- [00:17:51.683]made most of the data public
- [00:17:52.740]but if you see other crops
- [00:17:54.330]it's tough to find that
- [00:17:55.710]because I tried to work with other
- [00:17:56.970]crops and like you know,
- [00:17:58.020]the data was not available as such.
- [00:18:00.330]So I'm really glad like you know,
- [00:18:02.190]I worked on maize and sorghum initially,
- [00:18:03.960]otherwise I might be bit depressed about
- [00:18:06.150]the not getting the data.
- [00:18:08.550]But anyways coming to this,
- [00:18:10.500]so this time what I did
- [00:18:11.580]is we used about 1014 individuals
- [00:18:15.480]with 18 million markers,
- [00:18:17.310]which was on the version four.
- [00:18:19.230]And then instead of running
- [00:18:20.940]just straight away mlm model GWAS,
- [00:18:23.280]I did use farm CPU but a step
- [00:18:25.863]ahead of farm CPU do I use
- [00:18:27.780]something called resampling model
- [00:18:29.580]inclusion probability.
- [00:18:31.200]So what it does is say like you know you have a phenotype
- [00:18:34.050]number of genotype thousand lines with phenotypic value,
- [00:18:36.810]you take subset of the lines about 80% of lines or 85% of
- [00:18:40.950]line or you run a GWAS farm CPU with that because
- [00:18:44.280]farm CPU algorithm is that way,
- [00:18:45.780]it gives you one single like know significant snip.
- [00:18:48.660]So you take random 80% of lines
- [00:18:50.520]you run with the same trait
- [00:18:52.140]then again you take another random 80%, you rerun that,
- [00:18:55.200]you repeat that a hundred times
- [00:18:56.610]and then you count how many
- [00:18:57.780]times the same snip has come as significant.
- [00:19:00.630]So if a snip is coming significant again and again,
- [00:19:03.240]which means it's not a chance factor,
- [00:19:04.890]it's actually like you know, a meaningful snip.
- [00:19:07.697]So that's why like you know,
- [00:19:09.930]this is very robust analysis
- [00:19:11.790]and imagine doing that with all
- [00:19:13.530]162 phenotypes running like you know,
- [00:19:15.990]so many times a hundred times
- [00:19:17.550]I think like you know HCC
- [00:19:19.496]people would have been really
- [00:19:20.329]mad at me like you know taking
- [00:19:21.162]over all the resources
- [00:19:21.995]for long enough but we have really
- [00:19:24.480]great resource here at UNL
- [00:19:25.590]with HCC which really helped
- [00:19:27.017]run this analysis and like make this work possible.
- [00:19:32.220]Alright, so I ran all these things
- [00:19:33.990]and then like you put them
- [00:19:35.010]together but there are so many hits you find,
- [00:19:37.560]how do you study them because they're like, you know,
- [00:19:39.660]more than 2154 significant
- [00:19:42.960]snips across like, you know,
- [00:19:45.720]categories across eight categories
- [00:19:47.130]with more than five mlp.
- [00:19:48.390]That is the cutoff we chose to say like you know,
- [00:19:50.430]anything about that is significant.
- [00:19:52.740]So what we did after
- [00:19:53.670]that is combine all these peaks based on the LD.
- [00:19:57.060]If anything is above five LD and within one kb,
- [00:19:59.332]one MB interval,
- [00:20:00.720]then that belongs to one single peak we call peaks.
- [00:20:03.240]After doing that, we were left with 1,466 peaks.
- [00:20:07.110]Out of that there are many pleiotropic
- [00:20:08.462]peaks within the trait categories.
- [00:20:10.950]So which is like, you know,
- [00:20:12.540]very much possible and it's very common
- [00:20:14.670]because if you go to any given trait,
- [00:20:16.380]say agronomic traits,
- [00:20:17.700]they are always related
- [00:20:18.660]to each other within biochemical
- [00:20:19.950]they're related to each other.
- [00:20:21.630]But then also there are 26 pleiotropic peaks which were
- [00:20:24.540]associated across the, you know, groups.
- [00:20:28.020]So those I'm showing here in this upset plot.
- [00:20:30.930]So out of that there were
- [00:20:32.580]at least three peaks which are
- [00:20:33.840]associated with three different trait categories.
- [00:20:38.130]And that's really interesting
- [00:20:39.900]because there are times
- [00:20:41.580]when there is something in the seed
- [00:20:43.410]is associated with the root,
- [00:20:45.330]something in the seed is associated with some like,
- [00:20:47.457]you know, plant morphological traits
- [00:20:49.920]and people have not found it
- [00:20:51.510]because they had never combined them
- [00:20:52.710]and studied sometimes like you know,
- [00:20:53.880]only given set of traits
- [00:20:55.020]have been studied and then here I'm
- [00:20:57.750]showing all the like you know, traits,
- [00:21:00.690]GWAS hits and the number of traits
- [00:21:02.190]that I analyzed for that.
- [00:21:04.110]So how many like you know, GWAS hits are there,
- [00:21:05.910]how many traits we analyzed?
- [00:21:07.710]So it was a very good numbers that we got.
- [00:21:11.250]Again, considering time,
- [00:21:12.900]I'm gonna talk about two examples, once again.
- [00:21:15.480]There are many like you know,
- [00:21:16.380]really cool examples but let's talk about two.
- [00:21:19.950]I really like this hit
- [00:21:20.831]which is associated with starch
- [00:21:22.980]content and oil content in corn.
- [00:21:26.250]So this data was dataset
- [00:21:28.080]was taken from some other study,
- [00:21:30.120]they also used farm CPU
- [00:21:31.830]but they didn't find this hit.
- [00:21:33.150]But we are finding this hit
- [00:21:34.470]because of the robust like
- [00:21:35.760]analysis and the new marker set that we are using.
- [00:21:38.310]So that's why we are able to recover that hit,
- [00:21:41.190]which is really interesting.
- [00:21:42.750]And also here you can see the like you know,
- [00:21:44.640]major allele associated with reduction
- [00:21:46.500]in oil or increase in starch content.
- [00:21:50.790]One more thing is DGAT is
- [00:21:54.750]this gene on like you know
- [00:21:56.550]on this particular hit is associated with,
- [00:21:59.280]well the gene within that particular location is DGAT.
- [00:22:03.180]So this DGAT gene is formally like it's already known to be
- [00:22:06.990]associated in formation of triacylglycerol.
- [00:22:09.990]So which are basically
- [00:22:11.340]the oil bodies like you know your oil
- [00:22:13.976]content which modulates your oil content.
- [00:22:17.550]So these are already known.
- [00:22:19.170]But one more interesting
- [00:22:20.220]thing is the DGAT is also known
- [00:22:22.680]to be associated with the many abiotic,
- [00:22:25.800]yeah abiotic stress for example
- [00:22:27.660]chilling tolerance and drought.
- [00:22:29.640]But here you can see there
- [00:22:30.690]is a hit for the southern rust
- [00:22:32.010]which is very close
- [00:22:32.843]to it within like you know an mb which
- [00:22:34.260]means like it could be also associated with some
- [00:22:36.330]of the disease traits like other environmental factors.
- [00:22:40.050]So it again like can proves
- [00:22:42.333]some like in a hypothesis we
- [00:22:43.860]have, which can be tested later at some point.
- [00:22:48.480]And then there is another example that is liguleless4,
- [00:22:51.150]it is a KNOX11 kind of gene
- [00:22:53.490]and this gene has been reported
- [00:22:54.720]to be a dominant mutation
- [00:22:56.430]because it's a dominant mutation
- [00:22:57.810]that's why it is bit tough
- [00:22:59.100]to study in the field
- [00:23:00.060]and describe like you know,
- [00:23:00.930]what it is doing actually in the plant,
- [00:23:03.180]but it is known to alter the sheath blade
- [00:23:06.060]and boundaries in the maize where you have this ligule
- [00:23:09.090]and oracle and the absence
- [00:23:10.459]of that results in the fluted boundaries of the leave
- [00:23:15.210]for the like, you know,
- [00:23:16.043]because of absence of ligule and oracle.
- [00:23:18.330]So earlier there was no association reported
- [00:23:20.250]between the leaf number, root area, root
- [00:23:22.260]width or regulation of flowering
- [00:23:23.940]time with this particular loci.
- [00:23:26.160]But because of this analysis like you know,
- [00:23:27.840]we could see these are associated
- [00:23:29.100]with all these traits as well.
- [00:23:31.770]But now when we think about
- [00:23:32.850]these traits and all these different traits,
- [00:23:35.080]there are two different ways
- [00:23:36.420]these things can be happening.
- [00:23:38.100]So because of mutation in this gene,
- [00:23:41.910]the plant is flowering early
- [00:23:44.130]and because it is flowering early,
- [00:23:45.660]there are lesser number of leaves
- [00:23:46.890]because it doesn't have enough time
- [00:23:48.570]to like, you know, mature,
- [00:23:49.410]it has to mature very fast
- [00:23:51.060]and the root system is smaller
- [00:23:52.380]and there is limited development
- [00:23:53.580]in the plant or it can
- [00:23:54.750]happen the other way around also
- [00:23:56.400]because there are smaller
- [00:23:57.390]root system, that's why plant
- [00:23:58.842]is taking more time to like, you know,
- [00:24:01.860]it induces stress in the plant
- [00:24:03.660]and that's why it flowers early.
- [00:24:05.070]So there are like, you know,
- [00:24:05.903]many different ways you can see these things,
- [00:24:07.620]what's happening over there
- [00:24:11.430]and this paper was then like you know published in Giga
- [00:24:13.680]Science and like you know all the results
- [00:24:16.410]and interpretation, everything is available over there.
- [00:24:21.000]One more thing that I quickly want
- [00:24:22.232]to touch upon is this one
- [00:24:24.060]like you know quick paper that we published in 2020,
- [00:24:26.250]we went to field collect some data
- [00:24:28.050]and tissue sample and we saw like,
- [00:24:29.577]and there was very severe southern rush,
- [00:24:32.070]me being like you a breeder
- [00:24:33.300]and then I had a very great
- [00:24:34.680]colleague at that time,
- [00:24:36.427]Goncaioson and he was a plant pathologist.
- [00:24:38.730]We started discussing same thing
- [00:24:40.080]and I was like you know,
- [00:24:40.913]we should screen the whole field
- [00:24:42.960]and it was very short
- [00:24:45.270]period of time we started checking
- [00:24:46.530]like you know all the weather condition.
- [00:24:47.880]Next day it was supposed
- [00:24:48.750]to rain and we had to finish it as soon as possible.
- [00:24:51.690]So in the photo, myself
- [00:24:53.760]Gonciao and John, who is here,
- [00:24:57.120]we three went there and screened the whole panel like you
- [00:24:59.253]know that day itself plus
- [00:25:00.990]the issue being it was COVID time
- [00:25:02.730]and we cannot like, you know, work together.
- [00:25:04.110]That's the reason we are just funnily
- [00:25:06.330]far away from like in each other.
- [00:25:08.940]But it was really good.
- [00:25:10.320]This paper gave like,
- [00:25:11.153]you know there are a lot
- [00:25:12.360]of interpretations and nice hits in there.
- [00:25:14.310]I think Sebil like you know
- [00:25:15.750]I spoke with her and she's also
- [00:25:17.310]looking into these things right now.
- [00:25:19.860]So that's something interesting.
- [00:25:21.240]Sometimes you just have to like, you know,
- [00:25:23.130]turn the disaster into opportunity the way you want.
- [00:25:28.020]Alright, so that's about pleiotropy.
- [00:25:30.450]I found all this pleiotropic
- [00:25:31.650]region which are like you know
- [00:25:32.790]over there and you can address so many stories.
- [00:25:36.660]Let's jump into GXE.
- [00:25:39.630]So now we have this dataset which is like you know,
- [00:25:43.020]multi environment, multi trait dataset.
- [00:25:44.730]We can use that for detecting like other linear,
- [00:25:47.750]non-linear plasticity or GXE in maize.
- [00:25:51.090]So there are three approaches to dissect GXE.
- [00:25:54.810]So we had a very good colleague
- [00:25:56.670]and every time he would talk
- [00:25:57.940]about GXE, Diego Hurricane,
- [00:26:00.180]everyone knows about him
- [00:26:01.770]and he would always tell in his
- [00:26:03.300]presentation like no I took it from there, ignore it,
- [00:26:06.030]control it or exploit it.
- [00:26:08.730]And by practice we know
- [00:26:10.260]we cannot ignore GXE because it's
- [00:26:12.600]always there and it's great
- [00:26:14.940]to control it but it's extremely tough.
- [00:26:17.670]Best is to model it but it's not that easy to model it
- [00:26:20.910]because it's a complex thing.
- [00:26:22.440]So here like you know there are say like,
- [00:26:24.565]you know there are three environments
- [00:26:26.100]I'm showing here in this graph.
- [00:26:28.620]There are three environments they have like you know,
- [00:26:30.840]different means and equal linear
- [00:26:32.190]plasticity and equal non-linear plasticity.
- [00:26:35.970]And then like you know these
- [00:26:37.170]two lines are showing equal
- [00:26:39.115]means and equal linear plasticity and unequal linear
- [00:26:43.740]plasticity and equal non-linear plasticity.
- [00:26:46.920]And these two are showing equal means but equal linear
- [00:26:50.160]plasticity and unequal linear plasticity.
- [00:26:52.472]When I say linear plasticity
- [00:26:54.150]it is how like you know given
- [00:26:56.880]phenotype is in multiple environments,
- [00:26:59.040]say you have like three environments
- [00:27:00.510]if you draw a line your
- [00:27:01.440]linear regression how like
- [00:27:02.273]you know it's but non-linear
- [00:27:04.230]plasticity is which a difference between your a linear
- [00:27:06.450]plasticity and the mean
- [00:27:07.470]and how it changes across the environment.
- [00:27:11.010]So for this what we did is I used all the previously
- [00:27:14.190]published flowering time data from maize
- [00:27:16.560]and then like you know we had data
- [00:27:18.240]on Wisconsin diversity panel of about 800
- [00:27:20.197]genotypes from four different environment.
- [00:27:22.680]Nebraska 2020 Michigan 2020,
- [00:27:25.230]then Nebraska 21 and Michigan 21.
- [00:27:28.260]All these were grown in randomized
- [00:27:29.700]complete block design
- [00:27:31.200]and then like you know with
- [00:27:33.597]two replications and from all these
- [00:27:36.390]four locations we collected data
- [00:27:38.070]on anthesis silking and difference between anthesis
- [00:27:41.367]and silking as a subtraction
- [00:27:43.500]of anthesis minus silking
- [00:27:44.670]and their respective growing degree days.
- [00:27:46.560]The growing degree days were calculated using
- [00:27:48.312]all the temperature variables from
- [00:27:50.820]the nearby weather station.
- [00:27:55.380]And at the same time we
- [00:27:56.940]were working on the genotypic data set
- [00:27:58.560]because if you remember
- [00:27:59.700]from my previous work I used
- [00:28:01.140]the version four of the marker
- [00:28:02.670]set but then like you know the
- [00:28:04.290]version five came up at that time.
- [00:28:06.030]Then from our lab what
- [00:28:07.170]we did was we sequenced some of the
- [00:28:08.863]lines which didn't have re-sequencing data and there were
- [00:28:11.430]some lines for which re-sequencing data was available.
- [00:28:14.100]We combined all of them
- [00:28:14.933]and again called new set of snips
- [00:28:17.100]on the whole population.
- [00:28:20.250]So now we had like in about 366 million variants
- [00:28:23.850]with 46 million high confidence variants out
- [00:28:26.310]of which like an after filtering
- [00:28:27.510]and everything I was still left with
- [00:28:28.620]26 million high quality markers to run my GWAS.
- [00:28:32.820]So all the previously published flowering time
- [00:28:35.970]and individual environment flowering time data,
- [00:28:38.340]I took those, I ran GWAS on all
- [00:28:40.560]of them same RMIP model
- [00:28:42.420]that I explained earlier
- [00:28:44.070]and after that I saw these heads.
- [00:28:46.590]One important thing is MADS69
- [00:28:49.710]we detected MADS69 jet CNN eight and Rap2.7,
- [00:28:54.480]these are earlier known
- [00:28:55.440]to be associated with flowering time.
- [00:28:57.450]So it's good that we are detecting
- [00:28:59.130]something which is known
- [00:28:59.963]but we are also detecting
- [00:29:00.990]something which is not known.
- [00:29:02.580]Not only that we are also
- [00:29:04.590]detecting some of the hits which
- [00:29:06.150]are associated with multiple traits,
- [00:29:08.580]these flowering traits
- [00:29:09.750]in different environments while we are
- [00:29:11.640]detecting some hits which
- [00:29:12.780]are specific to given environment
- [00:29:14.100]but not in other environments.
- [00:29:15.990]What does that mean?
- [00:29:17.160]Which means these genes,
- [00:29:18.840]some of these genes are associated with the
- [00:29:21.367]adaptation or the plasticity
- [00:29:23.400]of these traits across like you know,
- [00:29:25.260]environments or there could be other genes also involved
- [00:29:28.140]which could be associated
- [00:29:29.280]with plasticity of these traits.
- [00:29:30.780]That's how like you know these plants are expressing
- [00:29:32.820]different and you're getting
- [00:29:33.653]different hits and different environments.
- [00:29:37.110]So with that hypothesis I was like okay,
- [00:29:39.510]what should we do?
- [00:29:40.343]How do we detect if there is any plasticity?
- [00:29:43.080]So what we decided was I decided to run something called
- [00:29:46.740]patient fin level consent regression analysis here,
- [00:29:51.120]this is like in a total outline
- [00:29:52.053]of what I'm doing for this
- [00:29:53.355]particular project.
- [00:29:54.480]I'm just used to draw those things.
- [00:29:56.010]It helps me to be within
- [00:29:57.562]that like you know space and don't go outside.
- [00:30:00.840]But I use this model which is Finlay Wilkinson model
- [00:30:03.960]well-known GXE model where YIZ
- [00:30:07.140]phenotype of the i'th accession
- [00:30:08.760]in the j'th environment,
- [00:30:10.050]mu is your general mean
- [00:30:11.520]and G your main effect of the i'th
- [00:30:13.318]accession and both together
- [00:30:15.120]it gives you intercept value
- [00:30:16.830]which is nothing but the general performance of accession in
- [00:30:19.200]the average environment.
- [00:30:21.540]Then beta gives your slope
- [00:30:23.100]that is your linear plasticity
- [00:30:24.660]or the change of the expected
- [00:30:26.040]performance of the i'th accession
- [00:30:28.140]per unit change in the environment
- [00:30:30.386]and then your E is your
- [00:30:32.760]effect of the environment
- [00:30:33.870]and this is your error term which
- [00:30:35.670]is nothing but it consists of your non-linear plasticity.
- [00:30:39.360]So I took, I analyzed that and then like you
- [00:30:41.343]know we got this intercept value,
- [00:30:43.860]beta value and like you know non-linear
- [00:30:45.930]plasticity beta is linear plasticity.
- [00:30:48.390]I again ran GWAS on all those
- [00:30:49.845]individual traits from these
- [00:30:52.710]four environments which are combined.
- [00:30:54.810]I found a lot of hits over there,
- [00:30:56.340]different kind of hits.
- [00:30:57.630]So what I did is in the background,
- [00:30:59.220]if you see all the gray colored
- [00:31:01.565]ones are the ones I got from
- [00:31:03.780]single environment hits and colored ones are
- [00:31:06.480]from the multiple environments combined
- [00:31:08.190]and this fin level kitchen analysis.
- [00:31:10.944]So obviously there are hits
- [00:31:12.810]which are overlapping them
- [00:31:14.670]and then if you see I got about
- [00:31:17.250]120 conventional GWA sets from
- [00:31:19.320]single environment
- [00:31:20.760]and there are at least 23 which are
- [00:31:22.710]associated with the plasticity
- [00:31:24.720]and there are 12 which are
- [00:31:25.710]overlapping between conventional
- [00:31:27.000]and main effect and about like you know,
- [00:31:29.430]22 if you calculate
- [00:31:30.750]which are associated 23 which are
- [00:31:32.280]associated with the main effect.
- [00:31:34.053]But the most important thing
- [00:31:35.790]is there are at least 23 hits
- [00:31:37.740]which are associated only with plasticity,
- [00:31:40.140]not with the main effect.
- [00:31:41.910]So this clearly proves
- [00:31:43.800]there are genes which are associated
- [00:31:45.150]with adaptation or plasticity of these lines.
- [00:31:49.110]I also calculated narrow sensibility
- [00:31:51.210]of these effects may affect
- [00:31:53.269]plasticity and the non-linear plasticity they were like
- [00:31:56.917]in a bit, okay, non-linear plasticity,
- [00:31:59.340]not much and we didn't even
- [00:32:01.260]get any hits for non-linear
- [00:32:02.550]plasticity but otherwise like
- [00:32:04.230]it was normal because it's
- [00:32:05.970]combination of all foreign get single value,
- [00:32:07.722]I cannot do broad sense.
- [00:32:09.090]So I did like in a marker based narrow sense over there.
- [00:32:13.650]Then all these hits that I found,
- [00:32:16.020]what I did was I tried
- [00:32:17.370]to compare them with all the known
- [00:32:19.020]flowering time genes.
- [00:32:19.860]So there are about 32 flowering
- [00:32:21.780]time genes which are known
- [00:32:23.610]and out of that I found hits
- [00:32:26.070]which are associated with known
- [00:32:27.660]flowering time genes.
- [00:32:28.500]But if you see all of them are only
- [00:32:30.420]the main effect or the
- [00:32:31.770]single environment GWAS.
- [00:32:33.480]So I'm not getting any overlap
- [00:32:35.010]with the plasticity with this.
- [00:32:36.900]So that made me think what's happening.
- [00:32:38.460]So maybe there are genes which are uncharacterized earlier
- [00:32:40.527]for flowering time
- [00:32:41.490]and are associated with plasticity.
- [00:32:43.650]So, but I wanted to dig more
- [00:32:46.200]into that and at the same time
- [00:32:48.360]in our lab we were collecting
- [00:32:49.511]this tissue sample for RNA-seq
- [00:32:53.130]and within a time period of 1.5 hours I think we collected
- [00:32:57.930]RNA like tissue sample for performing RNA from almost like
- [00:33:02.190]seven or 50 genotypes.
- [00:33:03.990]And if you guys know biology RNA changes so much because
- [00:33:07.260]of so many variation across the time.
- [00:33:09.510]So that's why we had to device
- [00:33:10.517]some way like you know we can
- [00:33:12.090]collect everything in a short period of time.
- [00:33:14.700]Jon Turkus who's technician in our lab,
- [00:33:16.560]he devised like he made all in-house like equipments
- [00:33:20.250]and everything and such a great setup
- [00:33:22.230]we could collect all the
- [00:33:23.310]data within no time,
- [00:33:24.360]all like you know tissue samples.
- [00:33:25.800]We go there like you know like a machine and that's it.
- [00:33:28.020]Within one and a half hour
- [00:33:28.920]we had all tissue sample and all
- [00:33:31.257]the data is analyzed by Dr. Torres who is sitting here
- [00:33:35.130]and he also has tested,
- [00:33:36.188]there is no variation
- [00:33:37.710]in the transcripts when you like,
- [00:33:39.837]you know, look across the timeframe it's not like you know,
- [00:33:43.260]horrendous or something like it's really a good dataset.
- [00:33:45.960]So that can be used.
- [00:33:47.250]I'm not gonna get in depth
- [00:33:48.450]of all the dataset
- [00:33:49.980]because he's working on that
- [00:33:51.450]and someday like you know give him a like
- [00:33:54.117]you know he's gonna talk about that
- [00:33:55.980]and that paper is almost
- [00:33:57.660]ready and we are gonna like, you know, publish that soon.
- [00:34:00.270]But anyways, using that data I ran
- [00:34:02.790]like you know TWAS with all the
- [00:34:05.910]patient win level consent analysis
- [00:34:07.800]that I got for the plasticity hits
- [00:34:09.930]as well as the main effects.
- [00:34:11.550]Most of them were associated main effect but I also found
- [00:34:14.490]this pebp12 that is phosphate ethanol amide binding
- [00:34:19.320]protein which is also associated with the plasticity
- [00:34:23.697]and this was earlier known to be associated with plasticity
- [00:34:26.730]which means our like you know,
- [00:34:28.290]whole methodology and analysis is working good
- [00:34:31.770]because we are finding something
- [00:34:32.970]which is associated and we are also
- [00:34:34.290]finding something which others have not found.
- [00:34:36.540]So which can be like heavily
- [00:34:38.070]relied on and we can like, you know,
- [00:34:39.450]characterize that in the future
- [00:34:40.740]and study those further.
- [00:34:45.090]With that, I would like to conclude
- [00:34:46.500]about this GXE paper because
- [00:34:48.060]this is still underwriting
- [00:34:49.170]and I'm going to publish it pretty soon.
- [00:34:52.440]So resampling based GWAS using
- [00:34:54.240]about 26 million markers gave
- [00:34:55.770]me about 136 unique peaks and about 273 market rate
- [00:35:00.330]association at a suggested cutoff of 0.5 and 47 peaks
- [00:35:04.770]with 95 markers with a high confidence
- [00:35:07.200]cutoff of 10 and then out
- [00:35:09.330]of that about 22 were solely associated
- [00:35:11.820]with main effect and about 24 peaks
- [00:35:14.190]are associated with the linear plasticity.
- [00:35:17.100]I think that should be 23 so I remember it is 23.
- [00:35:21.090]So 23 are associated with linear plasticity
- [00:35:23.910]and so we also compared with set
- [00:35:27.510]of the known flowering time genes
- [00:35:29.160]and we see that the main effect in single environment GWAS,
- [00:35:32.310]they do have overlap with the known flowering
- [00:35:35.040]time genes but not the plasticity,
- [00:35:36.810]which means high confidence
- [00:35:39.060]and large effect peaks
- [00:35:39.930]associated with linear plasticity such as that currently
- [00:35:42.630]uncharacterized genes play key role in determining
- [00:35:45.390]the response of maize two new environments
- [00:35:48.240]and the results is also
- [00:35:49.500]backed up by TWAS as I just now showed.
- [00:35:55.080]With that as I said like you know this dataset,
- [00:35:57.060]I used it for pleiotropy dissection
- [00:35:58.890]and then like GXE we
- [00:36:00.990]can use it for cross environment prediction as well.
- [00:36:04.290]So when we talk about cross environment prediction,
- [00:36:06.870]there are many things we can test
- [00:36:10.290]GXE effects or like you
- [00:36:11.793]know in the tested genotypes
- [00:36:13.320]in the tested environment
- [00:36:14.280]which is very easy but when you start testing your untested
- [00:36:17.460]genotypes in the untested environment it's less easy.
- [00:36:21.180]But when you want to test
- [00:36:22.350]your untested genotypes
- [00:36:23.550]in the untested environment it's extremely hard.
- [00:36:26.370]So there are various ways you can do it.
- [00:36:29.100]So say like now I now have multiple environments,
- [00:36:32.130]multiple dataset from multiple
- [00:36:34.321]environments and we can like, you know,
- [00:36:37.260]put into our all these phenotypes into our either crop
- [00:36:39.960]growth model or the kinship based like you know,
- [00:36:43.470]prediction model and we can perform prediction
- [00:36:45.630]in the untested environment
- [00:36:47.076]having like all the environmental
- [00:36:48.580]variables say soil
- [00:36:49.913]and like you know weather variables,
- [00:36:52.080]rainfall and everything
- [00:36:53.040]and we can still predict to certain
- [00:36:54.630]extent but there is still another
- [00:36:56.329]way like you know using
- [00:36:59.030]phenomic prediction.
- [00:36:59.910]So let's talk about a little bit about phenomic prediction.
- [00:37:02.820]So here like you know
- [00:37:03.822]this is a permutation based estimation
- [00:37:06.060]of feature importance
- [00:37:07.260]in a supervised machine learning model
- [00:37:09.930]that is random forest model
- [00:37:11.640]and trained to predict 2020 eel
- [00:37:13.530]data from non eel non year related trait.
- [00:37:17.190]So here I'm showing all the,
- [00:37:20.670]so here I'm showing relation
- [00:37:21.810]between eel and the flowering
- [00:37:23.190]time in Nebraska 2020.
- [00:37:25.020]So you can see nothing is like, you know,
- [00:37:27.450]less is not good or more
- [00:37:28.800]is not good but there is a sweet
- [00:37:30.000]spot over here where like you have higher yield
- [00:37:31.954]and at that point like you know
- [00:37:33.870]with the flowering time.
- [00:37:36.270]So what I did is I used
- [00:37:37.855]all the other phenotypes that we
- [00:37:41.100]collected from that particular year
- [00:37:43.830]and tried to predict eel
- [00:37:45.060]using that in a random forest model
- [00:37:47.700]and we could see that this
- [00:37:48.720]two pollen and plant height
- [00:37:50.850]and branches per tassel they
- [00:37:52.860]show up like you know,
- [00:37:53.693]highest ability to predict Detroit,
- [00:37:57.750]this is within a environment,
- [00:37:59.430]how about cross environment?
- [00:38:00.720]So now what I did is I also took data from Michigan 2020.
- [00:38:05.160]Now you can see the sweet spot has shifted a little bit.
- [00:38:08.250]So I again fed this thing
- [00:38:11.520]in like you know all the traits
- [00:38:12.780]measured in Nebraska and you had to like,
- [00:38:14.967]you know pay attention only the Nebraska traits but I'm
- [00:38:17.730]predicting Michigan eel as well as the Nebraska eel
- [00:38:20.970]which is in like you know different
- [00:38:22.170]colors here if you see still
- [00:38:23.940]flowering time and branches per tassel as well as plant
- [00:38:27.540]height are the major ones for predicting eel.
- [00:38:29.640]And what does it mean?
- [00:38:31.560]It means like in at some point
- [00:38:33.180]of time you can just fly a drone,
- [00:38:34.950]collect data on these things
- [00:38:36.060]and before you can even harvest
- [00:38:37.260]you can predict which lines are gonna be better at your like
- [00:38:39.442](indistinct) and it is also possible
- [00:38:42.450]someday you don't even have to grow your plants
- [00:38:44.310]in different environments in Michigan
- [00:38:46.590]and you can do that,
- [00:38:47.580]you can extend this to further like you know by adding
- [00:38:49.770]environmental variables
- [00:38:50.670]and everything to increase your
- [00:38:52.290]prediction but okay,
- [00:38:53.730]phenomic prediction we did
- [00:38:55.590]but everyone is working on genomic
- [00:38:57.330]prediction so how about comparison within that?
- [00:38:59.700]So one of the grad students
- [00:39:00.780]in lab, Hongyu Jin, he's been working working on like,
- [00:39:04.107]you know comparison both like you know phenomic prediction
- [00:39:07.680]and genomic prediction
- [00:39:09.120]and cross environment ell predictions
- [00:39:12.390]and if you see for the phenomic prediction we had R squared
- [00:39:14.700]of 0.2 while like you know for the genomic
- [00:39:18.030]prediction while the phenomic
- [00:39:19.470]prediction is about 0.33
- [00:39:21.630]and then he also tested like you
- [00:39:23.010]know compared the P-value,
- [00:39:24.420]the cross validation with 20 x randomization still phenomic
- [00:39:28.380]prediction is doing better than genomic production.
- [00:39:31.950]So these are like in a lot of different research
- [00:39:35.190]and everything that I worked
- [00:39:36.034]with James when I joined James lab
- [00:39:38.550]but when I initially joined James lab he put me in a
- [00:39:41.610]collaborative project with other people that's like,
- [00:39:44.790]you know there is a huge group
- [00:39:47.010]in Utah and they have been working
- [00:39:49.830]in developing nano gap based sensors
- [00:39:51.960]for detecting green leaf volatile detection.
- [00:39:54.420]So what happens whenever insects,
- [00:39:56.040]they attack plants,
- [00:39:57.360]they release certain kind of volatile compounds.
- [00:39:59.850]So what if, if you may make a sensor
- [00:40:01.560]and deploy them in the field
- [00:40:02.790]and they can detect this volatile compounds
- [00:40:04.890]and send a prompt on your phone,
- [00:40:06.300]hey you have like thousands
- [00:40:07.830]of acres and in this particular
- [00:40:09.150]area there is increase
- [00:40:10.680]in the insect population you can go
- [00:40:12.090]and just pray over there instead of like taking spray
- [00:40:14.790]throughout the field.
- [00:40:16.110]So this was really interesting
- [00:40:17.280]for me because I started
- [00:40:18.387]and I went to James lab with
- [00:40:20.070]that time where like I was really
- [00:40:22.637]scared of even sending emails
- [00:40:24.570]or something and straight away
- [00:40:25.620]he put me in collaboration
- [00:40:26.670]where I had to like contact them
- [00:40:28.170]and do everything a hundred percent.
- [00:40:30.210]So initially every time I would write
- [00:40:31.800]emails and I would go
- [00:40:32.633]to James, is this okay, is this okay?
- [00:40:34.140]Two times I did third time he was like, you know,
- [00:40:36.163]it's fine go ahead and do it.
- [00:40:37.620]If something goes wrong
- [00:40:38.520]I'm there to have your back.
- [00:40:39.960]Sometimes you need that, that day
- [00:40:41.700]and like you know till now
- [00:40:42.900]like any collaborator anything,
- [00:40:44.700]I'm like I'm handling multiple things
- [00:40:46.287]and it just makes a difference you know.
- [00:40:49.650]So again coming to this,
- [00:40:54.540]so I used to screen like you know
- [00:40:55.920]all their first used to
- [00:40:57.960]collect all the gases in the tedler bag,
- [00:41:01.290]do some like you know,
- [00:41:02.580]physical damage to plants
- [00:41:03.750]and then go to field and from
- [00:41:05.070]there and then test all their equipments
- [00:41:06.930]that they're creating all these electronic
- [00:41:09.015]engineers and then like you know,
- [00:41:11.520]send those to them
- [00:41:12.360]and they would test like different kind
- [00:41:14.970]of volatile compounds within that.
- [00:41:17.850]It was really interesting
- [00:41:19.230]because all the time we have been
- [00:41:21.287]talking to only people who have worked with plant but
- [00:41:24.210]imagine someone who doesn't even know
- [00:41:25.890]like you know what a plant is,
- [00:41:27.600]how like you know it actually
- [00:41:28.680]works because first question I
- [00:41:30.240]was asked was what is the difference
- [00:41:31.950]between sorghum and maize?
- [00:41:33.210]How do you differentiate them?
- [00:41:34.770]And I sat there and like no
- [00:41:36.870]that was a very tough question
- [00:41:38.010]for me because we all assume we know everything and everyone
- [00:41:40.920]would know that but it took me time
- [00:41:42.417]and I got to them with
- [00:41:43.980]lot of slides and everything
- [00:41:45.180]how everything is different,
- [00:41:46.140]different growth stages, why they are like,
- [00:41:47.617]you know, different things.
- [00:41:49.103]They really enjoyed it.
- [00:41:50.640]They came over here,
- [00:41:51.473]they worked in heat like you know,
- [00:41:52.680]first time in their life
- [00:41:53.670]and they're like, this is terrible.
- [00:41:54.840]But they enjoyed it at the same time.
- [00:41:56.880]I remember one point of time these students were there
- [00:42:00.690]and it was really hot like about almost like reaching 200
- [00:42:05.100]and they had to screen it
- [00:42:06.270]because they had to go back like you
- [00:42:07.233]know after screening there equipments, James
- [00:42:10.020]came there with some of the like you know,
- [00:42:11.430]mangoes and everything and those students like,
- [00:42:13.257]and you could see the happiness on on their face.
- [00:42:16.140]They just jumped out there and now we can work,
- [00:42:18.270]this is the best mango we ate.
- [00:42:19.860]So sometimes those kind of like, you know,
- [00:42:21.210]small support and good gestures are always, but anyways,
- [00:42:25.590]getting back to this sensor is already
- [00:42:28.560]in the beta test stage and we have
- [00:42:30.720]written like you know,
- [00:42:31.553]a couple of articles
- [00:42:32.490]on that which is already published multiple articles
- [00:42:35.310]and I know like you know there are more
- [00:42:36.810]articles that's coming out soon from this research.
- [00:42:42.607]With that a little bit like you
- [00:42:45.540]I think I have five more minutes
- [00:42:46.890]before I finish this so
- [00:42:47.970]I'll try to cover up within five minutes I'll talk a little
- [00:42:50.826]bit about what I'm doing now.
- [00:42:53.100]So I recently joined South Dakota State University
- [00:42:56.310]in the department of Economic horticulture
- [00:42:57.990]and plant science.
- [00:42:59.310]So this is Dr. Wright who is the department head,
- [00:43:01.950]very vibrant guy and like you know,
- [00:43:05.280]he is so enthusiastic about seeing like you know,
- [00:43:07.500]how science works and what's happening, whatnot.
- [00:43:10.320]He asked so many questions,
- [00:43:11.269]great guy to work with
- [00:43:13.020]and he's from Midwest so which is like, you know,
- [00:43:14.600]it makes really easy for me right
- [00:43:16.530]now to communicate because
- [00:43:18.720]Midwest has been my home
- [00:43:19.800]for a really, really long time.
- [00:43:21.124]I mean like you know Nebraska and Iowa.
- [00:43:23.640]So there are like lot of multidisciplinary departments
- [00:43:28.080]within the department, like minor departments.
- [00:43:30.690]First one the nation
- [00:43:31.650]to start bachelor's degree in precision agriculture.
- [00:43:34.560]There are about like you know 45 faculty,
- [00:43:36.630]hundreds of grad student
- [00:43:38.730]and there are about nine research
- [00:43:40.080]forms which are spread across
- [00:43:40.986]South Dakota and they also
- [00:43:43.380]have a centralized sequencing
- [00:43:44.490]and molecular biology facility
- [00:43:46.290]and also metabolomic facility right now.
- [00:43:49.110]And they have a new building
- [00:43:50.700]which is Raven Precision
- [00:43:52.920]Center with about 130K thousand
- [00:43:55.380]flow space and a newly renovated Berg Hall.
- [00:43:58.800]That's where I am stationed at
- [00:44:00.600]and within the like you know,
- [00:44:01.620]main campus with 17K square feet of the wet chemistry lab
- [00:44:06.210]and about 13K of the office space and everything.
- [00:44:09.450]And they also have a seed lab which is hosting right now
- [00:44:11.940]like you know, winter and spring weed breeders,
- [00:44:13.890]oats and their other breeding
- [00:44:15.210]programs that's going on over there.
- [00:44:18.390]With that I'm looking for PhD students,
- [00:44:20.880]that's my lab logo that created, I had that idea,
- [00:44:24.900]I wrote it down and Hendrik really helped me to put that
- [00:44:28.290]into like, you know, the Inkscape.
- [00:44:32.310]So my lab will be centrally focused
- [00:44:34.380]on doing like in a lot
- [00:44:35.451]of research for production
- [00:44:37.440]of crop launch for the commercial
- [00:44:38.850]utility using various omics approaches
- [00:44:42.600]and then main emphasis will be
- [00:44:44.070]on the quantitative genetics,
- [00:44:45.060]genomics and transcriptomics
- [00:44:46.740]of sorghum and corn and a bit
- [00:44:49.410]of breeding and sorghum.
- [00:44:50.910]But I also intend to like,
- [00:44:52.047]you know, extend my research work to legumes
- [00:44:54.900]and millet crops
- [00:44:56.430]and I'll make sure that like you know,
- [00:44:57.647]lab is committed to establishing a vibrant research
- [00:45:00.370]environment and provide extension training to students
- [00:45:03.180]and give the right opportunity to everyone.
- [00:45:06.570]So one other thing
- [00:45:07.530]that I'll be doing there with some
- [00:45:08.670]of the colleagues is developing
- [00:45:11.010]and setting up this agri data hub
- [00:45:13.020]or I don't know what I'm going to call it right now,
- [00:45:15.480]but yeah it will be basically center for advanced data
- [00:45:18.210]analysis for agriculture.
- [00:45:19.920]So which will help breeders
- [00:45:21.090]and quantitative geneticists over there to advance
- [00:45:22.980]like their analytics pipeline
- [00:45:24.600]and get everything turned
- [00:45:25.448]around like in a faster manner
- [00:45:27.900]so that they can make their
- [00:45:29.820]breeding like decisions as soon as possible.
- [00:45:33.930]With that I would like
- [00:45:35.160]to first thank Schnable Lab for like you
- [00:45:39.600]know hosting me there.
- [00:45:40.620]I was there as a postdoc,
- [00:45:42.120]one and a half year research assistant professor.
- [00:45:44.340]I learned so much.
- [00:45:45.660]Totally different world that I saw over there.
- [00:45:48.210]I would like to thank Dr. Yufanga
- [00:45:50.280]who initially told me like you know,
- [00:45:51.690]how to manage couple of things.
- [00:45:53.600]It was really helpful.
- [00:45:56.190]Dr. Rebecca Austin, every time on weekend,
- [00:45:58.200]whenever I go to office she
- [00:45:59.580]would be there and she would
- [00:46:00.600]always talk to me and motivate me, you know,
- [00:46:03.030]you need to do this, you need to do that.
- [00:46:03.863]And like, you know, telling stories,
- [00:46:05.910]which was really good.
- [00:46:07.200]Dr Jinlang Yang with whom like, you know,
- [00:46:09.210]we have been collaborating
- [00:46:10.170]in a couple of projects and I
- [00:46:11.550]have learned a lot and Andrew Benson who has been like
- [00:46:15.780]really supportive and always
- [00:46:16.656]whenever he talked to me he is
- [00:46:17.940]like, you know, you'll do great.
- [00:46:18.900]You need to go for this, you need to go for that.
- [00:46:20.700]For pushing me obviously, Tom Clemente,
- [00:46:23.400]I want to take his name for sure
- [00:46:24.960]because harsh words but real meaningful words.
- [00:46:28.860]He always like you know,
- [00:46:29.790]says something but it has meaning
- [00:46:31.440]if you really sit and like, you know,
- [00:46:32.640]think about it and he really
- [00:46:34.110]cares about your like in the future
- [00:46:35.580]and he gives very good advice.
- [00:46:38.100]This is where you can think like, you know,
- [00:46:40.230]experience really matters.
- [00:46:42.240]Like he's a good example
- [00:46:43.380]of that, Dr. Kahun for unwavering support
- [00:46:46.590]and he has been like such a great personality.
- [00:46:49.650]You always enjoy talking to him.
- [00:46:51.630]Dr. Martha sitting here like great leadership,
- [00:46:55.830]it's really difficult.
- [00:46:57.660]Imagine what she does like
- [00:46:58.860]in managing such huge group and people,
- [00:47:01.708]but she has been really great
- [00:47:03.210]support and obviously everyone
- [00:47:05.700]in the, oops something happened.
- [00:47:08.280]But yeah, obviously everyone in the Schnable lab,
- [00:47:12.360]mainly Jon because last time when James was not here,
- [00:47:16.140]we both were managing the whole lab and doing like,
- [00:47:18.903]you know, a lot of stuff
- [00:47:20.739]and whole field work, lab, everything.
- [00:47:22.213]It was bit of pressure
- [00:47:23.490]but he really held up with me
- [00:47:25.860]and like you know did
- [00:47:26.790]because sometimes I think I can push a
- [00:47:28.590]little bit to get things done.
- [00:47:30.570]I would really like to thank Vla, Vla has been great help
- [00:47:33.360]because whenever I was like,
- [00:47:35.460]you know when I was managing this it's not easy task.
- [00:47:39.180]I had to rant out sometime
- [00:47:40.350]and he was the person every time
- [00:47:41.808]I would go and like you know say he'll just listen,
- [00:47:44.610]not giving any advice, not doing anything,
- [00:47:46.350]just listen, you'll be fine.
- [00:47:47.520]That's it.
- [00:47:48.353]Which is required sometimes I think.
- [00:47:49.920]So I'm sorry if I like, you know,
- [00:47:51.690]really bothered you so much
- [00:47:54.000]but yeah that was great.
- [00:47:55.650]I think it's okay.
- [00:47:56.520]I think I'm already there.
- [00:47:57.835]Yeah.
- [00:48:06.210]I want to say one more line though.
- [00:48:08.130]I really want to thank Dr. Schnabel.
- [00:48:10.170]If he was not there I wouldn't be standing here today.
- [00:48:12.600]I'm sure about that.
- [00:48:14.160]And I always talked to our previous lab members,
- [00:48:16.710]whenever we discuss, we always discuss this.
- [00:48:19.050]James, I never told you this, but yes,
- [00:48:21.780]if James was not there we wouldn't
- [00:48:23.100]be where we are and we have been like,
- [00:48:24.267]you know, trained really well.
- [00:48:25.980]It's a great mentorship.
- [00:48:27.300]I would suggest every student to really like, you know,
- [00:48:29.370]learn and work with him at least
- [00:48:30.720]a little bit to get that experience.
- [00:48:33.900]Thank you.
- [00:48:34.733]Thanks a lot.
- [00:48:36.399](audience applauding)
- [00:48:46.380]Thank you so much for the wonderful presentation.
- [00:48:49.800]Does somebody has any question?
- [00:48:55.920]Yep.
- [00:48:59.820]Thank you and thanks for explaining
- [00:49:01.620]your talk to someone who's not
- [00:49:02.543]in your field and it made a lot
- [00:49:04.140]of sense so you should be
- [00:49:04.973]proud of yourself for that.
- [00:49:05.940]Thank you.
- [00:49:07.200]But so I saw that selfishly you
- [00:49:09.870]had two lipid synthesis genes,
- [00:49:11.550]the DGAT and then also the PE binding proteins and I think
- [00:49:15.900]the PE binding protein was for your RNA analysis work.
- [00:49:19.140]Yes. Yeah.
- [00:49:19.973]So because lipid metabolism genes fluctuate, diurnally,
- [00:49:24.060]do you have any plans if you plan
- [00:49:25.860]to follow these up you
- [00:49:27.090]or the lab to sample across like
- [00:49:29.310]a diurnal scale to see like
- [00:49:30.367]did you just capture something
- [00:49:32.040]that happened to be highly
- [00:49:33.450]transcribed at that time?
- [00:49:35.130]So currently like you know,
- [00:49:37.140]I don't because I'm working
- [00:49:38.970]a lot on the population level
- [00:49:40.290]but at some point when I have
- [00:49:41.820]a molecular lab and everything set up,
- [00:49:43.500]I would like to look into some
- [00:49:44.850]of these peaks and things.
- [00:49:47.490]It does make sense especially DGAT gene.
- [00:49:49.530]I'm always interested
- [00:49:50.460]in that because my PhD revolved around
- [00:49:52.770]DGAT a lot and it has bothered me
- [00:49:55.500]as well as it has made my life also.
- [00:49:57.420]So, and then like you know also
- [00:49:59.658]the interesting fact it is so
- [00:50:01.260]much closely associated with southern rust
- [00:50:03.270]like you know it is just within like you a couple
- [00:50:05.220]of base pairs and thousand base pairs.
- [00:50:07.830]So that makes it interesting and like you see papers,
- [00:50:10.830]it's always cold tolerance or drought and everything.
- [00:50:14.070]So this will be something to really dig in, look into.
- [00:50:17.550]Thank you.
- [00:50:18.420]Thank you.
- [00:50:22.890]Somebody else has another question.
- [00:50:33.180]Congrats for everything.
- [00:50:34.560]Thank you.
- [00:50:36.120]So scientific question.
- [00:50:38.010]So when you talk about pleiotropy
- [00:50:40.110]and also have you think
- [00:50:41.940]about this a really pleiotropy gene
- [00:50:45.420]or is because of a linkage that by chance that region
- [00:50:50.340]or is not by chance maybe
- [00:50:52.184]the primary trait being selected,
- [00:50:56.880]which likely have a large selective
- [00:51:01.890]bottleneck or a linkage block?
- [00:51:06.000]The a lot of other genes actually
- [00:51:08.550]fall in the L block linkage,
- [00:51:10.716]dis-equilibrium block,
- [00:51:13.770]therefore that can link
- [00:51:16.290]with associated with multiple other trees?
- [00:51:19.890]That's a great question because
- [00:51:21.960]it's always there whether it is pleiotropy,
- [00:51:23.597]real pleiotropy or linkage
- [00:51:25.560]for many traits as such.
- [00:51:27.540]But the thing is because there are a couple of traits which
- [00:51:30.060]are already being reported to be pleiotropic,
- [00:51:32.580]that's why I would like know,
- [00:51:33.450]rely more on like other traits
- [00:51:36.330]also which showed pleiotroph
- [00:51:37.680]with them as well as like, you know,
- [00:51:39.270]some other hits that we found
- [00:51:40.890]because say when you go
- [00:51:42.180]to sorghum, I did this (indistinct)
- [00:51:44.010]which is designed to detect like you know pleiotropy
- [00:51:46.320]because it considers LD and everything
- [00:51:47.940]within the genome so that like
- [00:51:49.770]know really hurts.
- [00:51:50.603]So I think like those are real geographic regions.
- [00:52:12.720]Thanks for a great seminar.
- [00:52:14.160]Two questions.
- [00:52:15.090]First one, you identify many genes
- [00:52:17.356]affecting flowering in maize.
- [00:52:20.670]However plant breeders usually think flowering time
- [00:52:24.100]is relatably easily to manipulate,
- [00:52:27.510]which suggests relatively
- [00:52:29.490]few genes make a difference.
- [00:52:31.650]How do you explain the easy
- [00:52:34.620]of changing flowering time with
- [00:52:36.780]the number of genes you found?
- [00:52:40.260]That's a good question.
- [00:52:41.610]So when you develop new models and new things,
- [00:52:43.620]you had to start somewhere.
- [00:52:44.970]So that's where like we started with flowering time.
- [00:52:46.980]So all these will be extended
- [00:52:48.300]to other important phenotypes as well.
- [00:52:51.510]One thing is flowering time
- [00:52:52.710]is associated with most of the traits.
- [00:52:54.600]So say e like you know,
- [00:52:55.920]important agro-important traits.
- [00:52:57.990]So that's why I started
- [00:52:58.920]with flowering time and then like you
- [00:53:00.720]know, we'll look into eel and other stuff as well.
- [00:53:04.620]And the second question is,
- [00:53:06.990]tassel branching affects pollen production.
- [00:53:10.230]Why do you think tassel branching affects yield?
- [00:53:13.950]What is the biological explanation.
- [00:53:17.130]Again, that's a good point.
- [00:53:18.300]Which we have been like, you know,
- [00:53:19.380]thinking about and right now like, you know,
- [00:53:22.320]I really don't know why exactly and as mentioned
- [00:53:25.680]in the question itself,
- [00:53:26.640]it's associated with like, you know,
- [00:53:27.780]more Poland and everything.
- [00:53:29.400]So that's why like you know the crossing,
- [00:53:32.400]like going and producing more seeds
- [00:53:34.140]because you need Poland
- [00:53:34.973]to produce seeds and all.
- [00:53:36.090]So that's why you're getting like,
- [00:53:36.923]you know, better eel.
- [00:53:38.340]That could be one reason.
- [00:53:39.750]But plant height and flowering
- [00:53:40.964]time I can clearly define but
- [00:53:43.470]tassel branching, yes.
- [00:53:44.580]That could be only explanation right now.
- [00:53:48.450]Thank you.
- [00:53:52.380]We have time for one more question.
- [00:53:55.530]Yep.
- [00:53:59.760]Sorry, this is interesting 'cause this is not,
- [00:54:01.170]I never hear about this stuff.
- [00:54:03.210]So I'm curious too,
- [00:54:04.688]you mentioned that you mostly
- [00:54:06.120]do this with sorghum and maize
- [00:54:09.510]and you talked in your future directions that you plan
- [00:54:12.390]to work with other crops, legumes
- [00:54:14.010]and other things.
- [00:54:14.843]Do you plan to maybe build
- [00:54:17.040]the dataset that you said does
- [00:54:18.360]not exist for other crops
- [00:54:20.310]that are not sorghum and maize?
- [00:54:22.716]Or is that like a you know,
- [00:54:24.905]a pipe dream that is never available.
- [00:54:26.760]No, that's a really,
- [00:54:29.460]that's something I want to do.
- [00:54:30.870]But at the same time, as I said, again,
- [00:54:33.840]not stepping on anybody's toe or something,
- [00:54:36.270]but sorghum and maize community has been really great.
- [00:54:39.690]They make a lot of dataset available.
- [00:54:41.250]But if you go to other crop, for example, soybean,
- [00:54:44.100]I wanted to start to this thing with soybean.
- [00:54:46.230]At one point I was working with Diego,
- [00:54:47.820]but there are a lot of dataset that's not available.
- [00:54:50.460]It's not like, you know,
- [00:54:51.293]handy every time researchers like,
- [00:54:53.040]okay on either like, you know,
- [00:54:55.230]request for this data or like,
- [00:54:56.640]it's never available.
- [00:54:57.510]So we don't know what's happening over there.
- [00:54:59.220]There are many, like, you know,
- [00:55:00.053]breeding programs and everything.
- [00:55:00.886]They have set of data and all,
- [00:55:02.820]it will never be made available
- [00:55:03.687]and I don't know what's gonna
- [00:55:04.890]happen with that.
- [00:55:05.723]Once you get those things
- [00:55:06.556]that those data then only you can make.
- [00:55:08.910]I think like in a again,
- [00:55:10.470]not trying to step on anything,
- [00:55:11.970]but if every data was available,
- [00:55:14.460]I think research was,
- [00:55:15.510]would have been a bit ahead of where we are today.
- [00:55:18.540]Because there are so much
- [00:55:19.500]that you can understand from like,
- [00:55:20.577]you know, previously available data.
- [00:55:22.200]There's so many things you can infer.
- [00:55:24.690]So yes, I want to do it in other crops, but again,
- [00:55:27.750]it depends on like availability of that,
- [00:55:30.000]which is really difficult.
- [00:55:32.490]I wish everyone made it public, like you know,
- [00:55:34.350]made their data public because
- [00:55:35.287]there are so many things that can be done.
- [00:55:38.850]There are many data sets which just never published.
- [00:55:41.040]It's just gone.
- [00:55:43.230]So I think there should be a rule that every dataset,
- [00:55:45.330]raw data process data has to be made public.
- [00:55:49.710]All the funding agencies should do that.
- [00:55:50.700]Like and as soon as you collect data,
- [00:55:52.380]first thing is made it public.
- [00:55:54.780]I don't know.
- [00:55:55.613]I'm not in that powerful position
- [00:55:57.030]to say things, but yeah,
- [00:55:59.310]that's my personal feeling.
- [00:56:02.460]Thank you.
- [00:56:06.960]So thank you again Dr. Mural
- [00:56:09.900]and our speaker next week will be Jackson Stencil,
- [00:56:14.160]founder and CEO Sten Channel Verigation Lincoln,
- [00:56:18.870]Nebraska presenting about scaling on-farm research
- [00:56:22.240]in a image-based fertigation
- [00:56:24.870]with customer driving development.
- [00:56:27.810]So everybody is more than welcome
- [00:56:29.654]to join and thank you for everybody as well.
- [00:56:34.050]Thanks a lot.
- [00:56:35.210](audience applauding)
The screen size you are trying to search captions on is too small!
You can always jump over to MediaHub and check it out there.
Log in to post comments
Embed
Copy the following code into your page
HTML
<div style="padding-top: 56.25%; overflow: hidden; position:relative; -webkit-box-flex: 1; flex-grow: 1;"> <iframe style="bottom: 0; left: 0; position: absolute; right: 0; top: 0; border: 0; height: 100%; width: 100%;" src="https://mediahub.unl.edu/media/21396?format=iframe&autoplay=0" title="Video Player: From Data Mining to Pleiotropic Effects, Environmental Interactions, and Phenomic Predictions of Natural Genetic Variants in Sorghum and Maize" allowfullscreen ></iframe> </div>
Comments
0 Comments