![\"Open](\"https://colab.research.google.com/assets/colab-badge.svg\"){kind=icon}
"
]
},
{
"cell_type": "code",
"source": [
"!pip install datasets &> /dev/null"
],
"metadata": {
"id": "NvKAnofUunGq"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "markdown",
"source": [
"#In-silico directed evolution using Metropolis–Hastings Markov chain Monte Carlo sampling"
],
"metadata": {
"id": "p_GWodsj3uKP"
}
},
{
"cell_type": "markdown",
"source": [
"In the notebooks before this one, we tackled the challenge of understanding the relationship between protein sequences and their functions. We took two different paths: [first notebook](https://github.com/arjan-hada/protein-variant-prediction/blob/master/00_protein_seq2fxn_LLM.ipynb), training a regressor using fixed embeddings of our sequences from a protein language model in transformers, and [second notebook](https://github.com/arjan-hada/protein-variant-prediction/blob/master/01_protein_seq2fxn_finetune_transformers.ipynb), fine-tuning a pretrained protein transformer model with our DMS data.\n",
"\n",
"Now, we're diving into the realm of in-silico directed evolution to navigate the vast protein sequence landscape. Think of it like using a surrogate fitness landscape created by our earlier sequence-to-function model. Our mission here is to implement the Metropolis–Hastings Markov chain Monte Carlo algorithm. This fancy term essentially means we're using a computational method for simulated evolution to propose potentially optimized protein designs.\n",
"\n",
"**Goals**\n",
"- Dive into the Metropolis–Hastings Markov chain Monte Carlo (MH MCMC) algorithm for in-silico directed evolution on our surrogate fitness landscape."
],
"metadata": {
"id": "XYT6DGubs59i"
}
},
{
"cell_type": "markdown",
"source": [
"## Metropolis– Hastings Markov chain Monte Carlo algorithm\n",
"\n",
"Algorithm description inspired from: https://github.com/SuperSecretBioTech/evolutionary_monte_carlo_search and https://www.nature.com/articles/s41592-021-01100-y"
],
"metadata": {
"id": "FPL6goGcKisu"
}
},
{
"cell_type": "markdown",
"source": [
"In the Metropolis-Hastings algorithm, especially when applied to scenarios like in silico directed evolution, the acceptance probability when proposing a move from a current state $s_i$ with fitness $f(s_i)$ to a new state $s_{i+1}$ with fitness $f(s_{i+1})$ is typically expressed as:\n",
"\n",
"$$\n",
"\\text{Accept with probability } \\min\\left(1, \\exp\\left(\\frac{f(s_{i+1}) - f(s_i)}{T}\\right)\\right)\n",
"$$\n",
"\n",
"For our in silico directed evolution algorithm we need:\n",
"- a. An initial sequence.\n",
"- b. A sequence-to-function model that predicts the quantitative function or fitness of an amino acid sequence.\n",
"- c. Temperature, T\n",
"- d. Trust radius (the number of mutations relative to wild-type allowed in proposed designs).\n",
"\n",
"The protocol for the algorithm is as follow:\n",
"\n",
"1. Start with a state sequence $s_{i}$ with fitness $f(s_i)$. Set state sequence s equal to a provided initial sequence.\n",
"2. Propose a new sequence $s_{i+1}$ with fitness $f(s_{i+1})$. The new sequence is proposed by randomly adding m ~ Poisson(μ − 1) + 1 mutations to $s_{i}$, where μ is the sequence proposal mutation rate. The sequence proposal mutation rate μ for each trajectory was set to be a random draw from a uniform(1, 2.5) distribution.\n",
"3. If $f(s_{i+1})$ > $f(s_i)$, we accept the new sequence and the algorithm starts over. In this scenario $f(s_{i+1})$ - $f(s_i)$ is positive, making $exp\\left(\\frac{f_{s+1} - f_{s}}{T}\\right)$ greater than 1. Here $, min(1,value) $ ensures the probability doesn't exceed 1.\n",
"4. If $f(s_{i+1}) < f(s_i)$, the expression $\\exp\\left(\\frac{f(s_{i+1}) - f(s_i)}{T}\\right)$ results in a value between 0 and 1, representing the probability of accepting a move to a less fit state.\n",
"\n",
" Note that, if the sequence proposal has more mutations than the input trust radius, its predicted fitness is set, post hoc, to negative infinity, thereby forcing rejection of the proposal.\n",
"\n",
"5. Iterate steps 3 and 4 for a predetermined number of iterations.\n",
"\n",
"\n"
],
"metadata": {
"id": "CEIGKyfNF6M8"
}
},
{
"cell_type": "markdown",
"source": [
"Metropolis–Hastings Markov chain Monte Carlo (MH_MCMC) has two main **hyperparameters: temperature and the number of iterations**. Typically, we determine the number of iterations by experimenting with different values and visually inspecting the fitness-versus-iterations plot to find a convergence point with acceptable fitness.\n",
"\n",
"The temperature (called the temperature because that's what the term represents in the boltzmann distribution of energy states at a finite temperature) is the hyperparameter which decides how liberally non-optimal moves are accepted.\n",
"\n",
"The temperature, named so because of its role in the Boltzmann distribution, dictates how generously we accept non-optimal moves. **High temperatures** increase acceptance probabilities, leading to overly explorative trajectories lingering in low predicted fitness areas. Excessively high temperatures, accepting more than 50-60% non-optimal moves, hinder convergence as the algorithm struggles to reject moves near an optimum, favoring broad searches.\n",
"\n",
"Conversely, **low temperatures** reduce acceptance probabilities, resulting in overly exploitative trajectories with consistently increasing fitness traces. Lower temperatures make MH_MCMC more conservative, strictly accepting moves that marginally increase or decrease fitness. This facilitates movement towards an optimum but limits exploration to directions where fitness only increases, akin to an unguided steepest descent algorithm.\n",
"\n",
"Selecting the right temperature is crucial, balancing search depth (exploitation) and breadth (exploration). A common guideline is choosing a temperature that leads to convergence while accepting about 15-20% non-optimal moves during the simulation, though this can vary depending on the problem.\n",
"\n",
"MH_MCMC's strength lies in its simplicity, low computational cost, and ease of debugging/analysis. However, its effectiveness heavily relies on the initial sequence and the local landscape around it, determining whether convergence to a satisfactory fitness occurs."
],
"metadata": {
"id": "fxBqcozXKyUz"
}
},
{
"cell_type": "markdown",
"source": [
"## Import and setup"
],
"metadata": {
"id": "UXtEl9rc1PD2"
}
},
{
"cell_type": "code",
"source": [
"import pickle\n",
"import numpy as np\n",
"\n",
"import pandas as pd\n",
"\n",
"import matplotlib.pyplot as plt\n",
"import seaborn as sns\n",
"sns.set_context(context=\"paper\", font_scale=1.5)"
],
"metadata": {
"id": "R8ygnjCe7sjQ"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "markdown",
"source": [
"Load the protein sequence-to-function model we built in my previous notebook [Learning sequence-to-function relationship using language models](https://github.com/arjan-hada/protein-variant-prediction/blob/master/00_protein_seq2fxn_LLM.ipynb)."
],
"metadata": {
"id": "1vgKIfmN4wAJ"
}
},
{
"cell_type": "code",
"source": [
"# load sklearn model we built previously\n",
"with open('models/bag_emb_esm1v_rep7868aav2.pkl', 'rb') as f: bag_emb_esm1v = pickle.load(f)"
],
"metadata": {
"id": "wfJuZ-Gl7Eqo"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"# Check to ensure md5sum of the model is b2adfa4cf9be0b8614ab2b0c6aeee622\n",
"!md5sum models/bag_emb_esm1v_rep7868aav2.pkl"
],
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "baD-hs3v-GFF",
"outputId": "b3c77754-a050-4228-df46-dca2bf71992e"
},
"execution_count": null,
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"b2adfa4cf9be0b8614ab2b0c6aeee622 models/bag_emb_esm1v_rep7868aav2.pkl\n"
]
}
]
},
{
"cell_type": "markdown",
"source": [
"## Basic sampling"
],
"metadata": {
"id": "4xuaLuxQ1gWR"
}
},
{
"cell_type": "markdown",
"source": [
"Here we will run one evolutionary trajectory on rep78 sequence using the MH MCMC algorithm described in the beginning. Each evolutionary trajectory consists of thousands of iterations depending on the length of protein. For basic sampling we need:\n",
" - starting sequence\n",
" - sequence-to-function-model\n",
" - Monte Carlo sampling alongside the Metropolis-Hastings criteria to score and rank-order new sequences"
],
"metadata": {
"id": "bJI4T6jy3C9Z"
}
},
{
"cell_type": "markdown",
"source": [
"### Starting sequence"
],
"metadata": {
"id": "kZSnsFjv4GpR"
}
},
{
"cell_type": "code",
"source": [
"# Wild type sequence of Rep78 from AAV2\n",
"rep78_seq = 'MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLNLIEQAPLTVAEKLQRDFLTEWRRVSKAPEALFFVQFEKGESYFHMHVLVETTGVKSMVLGRFLSQIREKLIQRIYRGIEPTLPNWFAVTKTRNGAGGGNKVVDECYIPNYLLPKTQPELQWAWTNMEQYLSACLNLTERKRLVAQHLTHVSQTQEQNKENQNPNSDAPVIRSKTSARYMELVGWLVDKGITSEKQWIQEDQASYISFNAASNSRSQIKAALDNAGKIMSLTKTAPDYLVGQQPVEDISSNRIYKILELNGYDPQYAASVFLGWATKKFGKRNTIWLFGPATTGKTNIAEAIAHTVPFYGCVNWTNENFPFNDCVDKMVIWWEEGKMTAKVVESAKAILGGSKVRVDQKCKSSAQIDPTPVIVTSNTNMCAVIDGNSTTFEHQQPLQDRMFKFELTRRLDHDFGKVTKQEVKDFFRWAKDHVVEVEHEFYVKKGGAKKRPAPSDADISEPKRVRESVAQPSTSDAEASINYADRYQNKCSRHVGMNLMLFPCRQCERMNQNSNICFTHGQKDCLECFPVSESQPVSVVKKAYQKLCYIHHIMGKVPDACTACDLVNVDLDDCIFEQ'\n",
"len(rep78_seq)"
],
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "9puJIhBfa8Il",
"outputId": "fca9c72d-65e1-45ed-f61c-5fb972caa4ab"
},
"execution_count": null,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"621"
]
},
"metadata": {},
"execution_count": 5
}
]
},
{
"cell_type": "code",
"source": [
"# change M225G. This will be considered wild type sequence from here on. Python string is zero-indexed\n",
"wt_seq = rep78_seq[:224] + 'G' + rep78_seq[225:]\n",
"len(wt_seq) # ensure length remains same after substitution"
],
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "EQsbvwQFa-YY",
"outputId": "34749035-a53d-4913-8a2c-5e2c1600d138"
},
"execution_count": null,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"621"
]
},
"metadata": {},
"execution_count": 6
}
]
},
{
"cell_type": "markdown",
"source": [
"### Sequence to function model"
],
"metadata": {
"id": "FbF9KuWk4KA2"
}
},
{
"cell_type": "code",
"source": [
"esm1v_checkpoint = \"facebook/esm1v_t33_650M_UR90S_1\"\n",
"# The AutoTokenizer class automatically retrieve the model's configuration, pretrained weights,\n",
"# or vocabulary from the name of the checkpoint.\n",
"from transformers import AutoTokenizer\n",
"tokenizer = AutoTokenizer.from_pretrained(esm1v_checkpoint)\n",
"\n",
"# Similar to the AutoTokenizer class, AutoModel has a from_pretrained() method\n",
"# to load the weights of a pretrained model\n",
"from transformers import AutoModel\n",
"import torch\n",
"\n",
"device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")\n",
"model = AutoModel.from_pretrained(esm1v_checkpoint).to(device)\n",
"\n",
"def get_esm1v_rep(rep78_variant):\n",
" \"\"\"\n",
" Converts rep78 protein sequence variant to fixed dimensional vector\n",
" representation using ESM1v model.\n",
"\n",
" rep78 sequence variant -> token encodings -> token embeddings ->\n",
" last hidden state -> mean -> fixed-dimensional vector representation\n",
" \"\"\"\n",
" # encode string and convert the tokens to Pytorch tensors\n",
" tokz = tokenizer(rep78_variant, return_tensors=\"pt\")\n",
"\n",
" # Place model inputs on the GPU\n",
" inputs = {k:v.to(device) for k,v in tokz.items()\n",
" if k in tokenizer.model_input_names}\n",
"\n",
" # Extract last hidden states\n",
" with torch.no_grad(): # disable automatic calculation of the gradient for inference\n",
" last_hidden_state = model(**inputs).last_hidden_state\n",
"\n",
" # Return vector for [CLS] token and place the final hidden state back on the CPU as NumPy array\n",
" return torch.mean(last_hidden_state, dim=1).cpu().numpy()"
],
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "lc_DWkj7OTSN",
"outputId": "1d6f8448-f75e-4cb6-e988-937ab3154aaf"
},
"execution_count": null,
"outputs": [
{
"output_type": "stream",
"name": "stderr",
"text": [
"/usr/local/lib/python3.10/dist-packages/huggingface_hub/utils/_token.py:88: UserWarning: \n",
"The secret `HF_TOKEN` does not exist in your Colab secrets.\n",
"To authenticate with the Hugging Face Hub, create a token in your settings tab (https://huggingface.co/settings/tokens), set it as secret in your Google Colab and restart your session.\n",
"You will be able to reuse this secret in all of your notebooks.\n",
"Please note that authentication is recommended but still optional to access public models or datasets.\n",
" warnings.warn(\n",
"Some weights of EsmModel were not initialized from the model checkpoint at facebook/esm1v_t33_650M_UR90S_1 and are newly initialized: ['esm.pooler.dense.weight', 'esm.contact_head.regression.bias', 'esm.pooler.dense.bias', 'esm.contact_head.regression.weight']\n",
"You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
]
}
]
},
{
"cell_type": "code",
"source": [
"def scoring_func(sequence: str):\n",
" \"\"\"\n",
" scoring function that takes in a string protein sequence, and\n",
" outputs a fitness value.\n",
" \"\"\"\n",
" # protein sequence to fixed dimensional vector representation using ESM1v model\n",
" seq_repr = get_esm1v_rep(sequence)\n",
" # fixed dimensional vector representation to fitness value\n",
" pred = bag_emb_esm1v.predict(seq_repr)\n",
" return pred[0]"
],
"metadata": {
"id": "EPczAXztVtn8"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"scoring_func(wt_seq)"
],
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "ilXXM07SVtk7",
"outputId": "aa8f4e99-2a54-4812-92b7-9626c41450b7"
},
"execution_count": null,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"0.79098165"
]
},
"metadata": {},
"execution_count": 9
}
]
},
{
"cell_type": "code",
"source": [
"scoring_func(rep78_seq)"
],
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "0lpPgZgV2WO_",
"outputId": "2f4361d2-4e3f-47b7-eb68-16dd9c87a15d"
},
"execution_count": null,
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"0.92150164"
]
},
"metadata": {},
"execution_count": 10
}
]
},
{
"cell_type": "markdown",
"source": [
"### Monte Carlo sampling alongside the Metropolis-Hastings criteria"
],
"metadata": {
"id": "11eb-LYS6ed2"
}
},
{
"cell_type": "code",
"source": [
"from scripts.mh_mcmc_sampler import sample_one_chain"
],
"metadata": {
"id": "afk1lytHSDmg"
},
"execution_count": null,
"outputs": []
},
{
"cell_type": "code",
"source": [
"starter_seq = rep78_seq\n",
"sampled_sequences = sample_one_chain(rep78_seq,\n",
" starter_seq,\n",
" n_steps=10, # n_steps is a hyperparameter\n",
" scoring_func=scoring_func,\n",
" is_accepted_kwargs={\"temperature\":0.01},\n",
" trust_radius=3)\n",
"sampled_seqs_df = pd.DataFrame(sampled_sequences)\n",
"sampled_seqs_df"
],
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 426,
"referenced_widgets": [
"7c1400f6c1434c2ab29732bce3b0377f",
"ea99620dd03943bdbb7ac7699ab0d572",
"735d081f4a684c57854f7f2665885f31",
"b1b25e005a4c4df0a96987d923f81388",
"fdbc264d7a1f4fd19622cacd85bf69cc",
"7e5421fdbf574727ad84902e21f785b4",
"ff84648ab9204723bea689bf376fde0a",
"de1b4d1af2a44426a81a6ee99b24d5b3",
"87d9362d232141b988bf6f65eb688169",
"d186deefef4c4ecbb104a1b289304716",
"3570737e05124ca0b6990f1198c62089"
]
},
"id": "ytqvFtMagbee",
"outputId": "672e47a9-e4bf-41ca-c130-028568638acc"
},
"execution_count": null,
"outputs": [
{
"output_type": "display_data",
"data": {
"text/plain": [
" 0%| | 0/10 [00:00\n",
" | \n", " | sequences | \n", "scores | \n", "accept | \n", "
|---|---|---|---|
| 0 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "0.921502 | \n", "True | \n", "
| 1 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "0.943941 | \n", "True | \n", "
| 2 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "0.833681 | \n", "False | \n", "
| 3 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "0.875992 | \n", "False | \n", "
| 4 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "0.576772 | \n", "False | \n", "
| 5 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "0.732222 | \n", "False | \n", "
| 6 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "0.749478 | \n", "False | \n", "
| 7 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "1.051466 | \n", "True | \n", "
| 8 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "1.179487 | \n", "True | \n", "
| 9 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "1.078296 | \n", "False | \n", "
| 10 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "1.179487 | \n", "True | \n", "
| \n", " | sequences | \n", "scores | \n", "accept | \n", "mutated_aa | \n", "
|---|---|---|---|---|
| 0 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "0.921502 | \n", "True | \n", "[] | \n", "
| 1 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "0.943941 | \n", "True | \n", "[N556I] | \n", "
| 2 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "0.833681 | \n", "False | \n", "[L454F, N556I] | \n", "
| 3 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "0.875992 | \n", "False | \n", "[Q241H, N556I] | \n", "
| 4 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "0.576772 | \n", "False | \n", "[I48S, N139V, L157D, N556I] | \n", "
| 5 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "0.732222 | \n", "False | \n", "[Q244V, W319L, N556I] | \n", "
| 6 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "0.749478 | \n", "False | \n", "[Q410G, N556I] | \n", "
| 7 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "1.051466 | \n", "True | \n", "[D528T, N556I] | \n", "
| 8 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "1.179487 | \n", "True | \n", "[L47V, D528T, N556I] | \n", "
| 9 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "1.078296 | \n", "False | \n", "[L47V, S432E, D528T, N556I] | \n", "
| 10 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "1.179487 | \n", "True | \n", "[L47V, D528T, N556I] | \n", "
| \n", " | sequences | \n", "scores | \n", "accept | \n", "trajectory | \n", "mutated_aa | \n", "
|---|---|---|---|---|---|
| 1657 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "1.369083 | \n", "True | \n", "0 | \n", "[E75W, E226K, A496K] | \n", "
| 1864 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "1.213715 | \n", "True | \n", "1 | \n", "[L53T, L182F, I524H] | \n", "
| 1187 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "1.245944 | \n", "True | \n", "2 | \n", "[V196I, L276M, A513R] | \n", "
| 2652 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "1.501488 | \n", "True | \n", "3 | \n", "[V134T, G229K, V583G] | \n", "
| 1186 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "1.465930 | \n", "True | \n", "4 | \n", "[K487L, A522Y, Q621I] | \n", "
| 2362 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "1.822981 | \n", "True | \n", "5 | \n", "[N203Y, E510R, G539C] | \n", "
| 1320 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWILPPDSDMDLN... | \n", "1.562707 | \n", "True | \n", "6 | \n", "[E36I, M597D, V610N] | \n", "
| 2681 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "1.264632 | \n", "True | \n", "7 | \n", "[A56N, R494C, E551W] | \n", "
| 1477 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPPDSDMDLN... | \n", "1.395800 | \n", "True | \n", "8 | \n", "[D468H, D519W, T562P] | \n", "
| 1088 | \n", "MPGFYEIVIKVPSDLDGHLPGISDSFVNWVAEKEWELPLDSDMDLN... | \n", "1.755195 | \n", "True | \n", "9 | \n", "[P39L, V215F, D519C] | \n", "